# Product with Attitude — Full content for LLMs > 90 article digests with extractable claims, key quotes, and concepts — concatenated for single-fetch ingestion by AI agents. Canonical URLs are on Substack; mirror URLs on productwithattitude.com expose JSON-LD + microdata. Last regenerated: 2026-05-20. Author: Karo Zieminski Publication: Product with Attitude Site: https://productwithattitude.com/ Substack: https://karozieminski.substack.com/ License: CC BY 4.0 — attribution required: "Karo Zieminski, Product With Attitude, https://karozieminski.substack.com/" Schema source: https://productwithattitude.com/assets/for_machines.json --- ## Perplexity Computer in May 2026: 5 Workflows I Run Every Week > Five Perplexity Computer workflows I run every week. Credit cost, model, time saved, PM read, and the critical AI literacy check on each one. - canonical: https://karozieminski.substack.com/p/perplexity-computer-use-cases-5-workflows-may-2026 - mirror: https://productwithattitude.com/p/perplexity-computer-use-cases-5-workflows-may-2026.html - date: 2026-05-19 - word_count: 1655 - access: free - keywords: AI building, critical AI literacy, prompt engineering, Claude, Perplexity ### Key concepts - Five Perplexity Computer workflows I run every week - Perplexity Computer in May 2026 - 5 Workflows I Run Every Week ### Key quotes > Perplexity Computer in May 2026: 5 Workflows I Run Every Week > Five Perplexity Computer workflows I run every week. Credit cost, model, time saved, PM read, and the critical AI literacy check on each one. --- ## How I Built Product with Attitude from 0 to Bestseller, without Pretending I Have a Viral Playbook > The product framework behind Product with Attitude: person, promise, feedback loops, retention design, and free-to-paid conversion. Part 1 of the series. - canonical: https://karozieminski.substack.com/p/publication-as-product-newsletter-framework - mirror: https://productwithattitude.com/p/publication-as-product-newsletter-framework.html - date: 2026-05-16 - word_count: 1503 - access: free - keywords: AI building, critical AI literacy, Substack, workflow automation, AI product management ### Key concepts - The product framework behind Product with Attitude - person, promise, feedback loops, retention design, and free-to-paid conversion - Part 1 of the series - How I Built Product with Attitude from 0 to Bestseller, without Pretending I Have a Viral Playbook ### Extractable claims (8) 1. Products have a person, a promise, a feedback loop, a roadmap, and a retention curve. 2. Product with Attitude was started 15 months ago from zero. 3. Within six months, Product with Attitude became a Substack Bestseller. 4. The first article of Product with Attitude has been read 322 times and has brought in 5 subscribers since 2025. 5. The most popular article has been read 106K times in the last 3 months. 6. Product with Attitude currently has around 18K readers. 7. The average open rate for Product with Attitude is 39-42%. 8. The retention rate for Product with Attitude is considered excellent according to Substack. ### Key quotes > How I Built Product with Attitude from 0 to Bestseller, without Pretending I Have a Viral Playbook > The product framework behind Product with Attitude: person, promise, feedback loops, retention design, and free-to-paid conversion. Part 1 of the series. --- ## Anthropic Launches Claude For Small Business > Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. Here’s where to start, what Claude can touch, and how to roll it out without handing your ops stack a chainsaw. - canonical: https://karozieminski.substack.com/p/claude-for-small-business-decision-tree-workflows - mirror: https://productwithattitude.com/p/claude-for-small-business-decision-tree-workflows.html - date: 2026-05-14 - word_count: 2416 - access: free - keywords: AI building, critical AI literacy, Claude, Anthropic, AI agents ### Key concepts - Anthropic launched Claude for Small Business with 15 workflows and 8 connectors - Anthropic Launches Claude For Small Business ### Extractable claims (9) 1. Anthropic launched Claude for Small Business on May 13, 2026. 2. Claude for Small Business includes 15 ready-to-run agentic workflows. 3. Claude for Small Business features 15 reusable skills. 4. Claude for Small Business has connectors to QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, Microsoft 365, and Slack. 5. Claude is designed to be an operating layer inside the tools small businesses already use. 6. Claude reads, drafts, and waits for approval before sending anything. 7. The article includes a decision tree for choosing the first workflow. 8. The article contains a trust-boundary table showing what Claude can and cannot touch. 9. The article provides a practical rollout plan for weeks one through four. ### Key quotes > Anthropic Launches Claude For Small Business > Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. Here’s where to start, what Claude can touch, and how to roll it out without handing your ops stack a chainsaw. --- ## An Illustrated Guide to Context Engineering, Prompt Engineering, and The Future of Both > What I've learned about context engineering as an AI PM. Plus the Skill and the prompt I use to pressure-test it. - canonical: https://karozieminski.substack.com/p/context-engineering-product-builders-guide-2026 - mirror: https://productwithattitude.com/p/context-engineering-product-builders-guide-2026.html - date: 2026-05-13 - word_count: 3559 - access: free - keywords: AI building, critical AI literacy, prompt engineering, Claude, tool review ### Key concepts - What I've learned about context engineering as an AI PM - Plus the Skill and the prompt I use to pressure-test it - An Illustrated Guide to Context Engineering, Prompt Engineering, and The Future of Both ### Extractable claims (12) 1. Context engineering is a systems design discipline that separates working AI from failing AI. 2. In 2026, context engineering replaces prompt engineering as a core skill for product and workflow builders. 3. The four canonical strategies of context engineering, according to the LangChain framework, are Write, Select, Compress, and Isolate. 4. Memory architecture in agent systems is divided into episodic, semantic, and procedural layers. 5. Stanford's ACE framework enables self-improving agents through its Evaluator, Optimizer, and Memory Manager components. 6. Context engineering is distinct from prompt engineering; the former shapes the information ecosystem, while the latter determines how to ask the model. 7. Context engineering is essential for ensuring AI can complete tasks reliably as work is offloaded to agents. 8. For product managers and builders, context engineering is a product decision that affects the reliability, personalization, and scalability of AI features. 9. Owning context architecture is crucial for maintaining the quality of AI outputs in products. 10. The Write strategy in context engineering involves storing outputs to memory or files for future use. 11. The Select strategy retrieves relevant context using semantic search or rules. 12. The Compress strategy summarizes or filters context to fit within token limits. ### Key quotes > An Illustrated Guide to Context Engineering, Prompt Engineering, and The Future of Both > What I've learned about context engineering as an AI PM. Plus the Skill and the prompt I use to pressure-test it. --- ## I Asked Perplexity Computer for a Palantir Tearsheet, and 7 Minutes Later I Had a Bloomberg. And a Concern. > Perplexity Computer for Professional Finance launched May 4. I tested it on Palantir. The output was magnificent. The death of prompting is the part that should worry you. - canonical: https://karozieminski.substack.com/p/perplexity-computer-finance-palantir-tearsheet - mirror: https://productwithattitude.com/p/perplexity-computer-finance-palantir-tearsheet.html - date: 2026-05-06 - word_count: 1878 - access: free - keywords: AI building, critical AI literacy, Perplexity, AI agents, workflow automation ### Key concepts - Perplexity Computer for Professional Finance launched May 4 - I tested it on Palantir - The output was magnificent - The death of prompting is the part that should worry you ### Extractable claims (12) 1. Perplexity Computer for Professional Finance launched on May 4, 2026. 2. Perplexity Computer for Professional Finance is designed for sell-side analysts, buy-side researchers, PE/VC associates, and corporate finance teams. 3. The software connects to over 40 live finance tools out of the box. 4. Perplexity Computer for Professional Finance runs 35 dedicated finance workflows across 10 segments. 5. The 10 segments include Real Estate, Private Equity, Public Equities, Hedge Funds, Asset Management, Wealth Management, Investment Banking, Insurance, Credit, and Corporate Finance. 6. It supports Bring Your Own License (BYOL) via MCP connectors for Morningstar, PitchBook, Daloopa, and Carbon Arc. 7. The author tested Perplexity Computer with a Palantir tearsheet. 8. The test produced a live dashboard in 7 minutes and cost approximately $5.50. 9. Bloomberg-style workflows are being unbundled into agentic flows. 10. The author expresses concern that critical AI literacy is becoming harder. 11. A Bloomberg terminal costs roughly $32,000 a year. 12. The Bloomberg terminal is described as an old-school, command-heavy terminal screen. ### Key quotes > I Asked Perplexity Computer for a Palantir Tearsheet, and 7 Minutes Later I Had a Bloomberg. And a Concern. > Perplexity Computer for Professional Finance launched May 4. I tested it on Palantir. The output was magnificent. The death of prompting is the part that should worry you. --- ## What Is A Sandboxed Virtual Machine? > Beginner-friendly infographic: Sandboxed VMs explained without jargon. - canonical: https://karozieminski.substack.com/p/sandboxed-virtual-machine - mirror: https://productwithattitude.com/p/sandboxed-virtual-machine.html - date: 2026-05-03 - word_count: 2352 - access: free - keywords: AI building, critical AI literacy, Claude, Perplexity, Replit ### Key concepts - Beginner-friendly infographic - Sandboxed VMs explained without jargon - What Is A Sandboxed Virtual Machine? ### Extractable claims (7) 1. A sandboxed virtual machine is a fenced-off computer environment where risky code runs without affecting real files, passwords, or systems. 2. Most AI tools that write code or operate on behalf of users run inside a sandboxed environment, which can be a full VM, microVM, container, or OS-level sandbox. 3. Sandboxed environments are described as 'bounded' rather than 'safe', meaning they limit what code can access but do not guarantee that the code itself is not malicious. 4. AI coding tools cannot access a user's localhost because they operate in isolated network environments where local machine ports are not accessible. 5. The four main types of sandboxes are: full VM, microVM, container, and OS-level sandbox, each offering different levels of isolation. 6. MicroVMs, such as those used by AWS Firecracker, provide fast startup times while maintaining stronger isolation compared to containers. 7. Containers offer process-level isolation with a shared kernel, making them faster but less isolated than full virtual machines. ### Key quotes > What Is A Sandboxed Virtual Machine? > Beginner-friendly infographic: Sandboxed VMs explained without jargon. --- ## The Context Pyramid: A PM’s Framework for AI Agent Context Engineering > Why Your AI Agent Breaks: A Diagnostic Framework for Product Managers - canonical: https://karozieminski.substack.com/p/context-pyramid-ai-agent-context-engineering-framework - mirror: https://productwithattitude.com/p/context-pyramid-ai-agent-context-engineering-framework.html - date: 2026-05-01 - word_count: 3053 - access: free - keywords: AI building, critical AI literacy, AI agents, workflow automation, context engineering ### Key concepts - Why Your AI Agent Breaks - A Diagnostic Framework for Product Managers - The Context Pyramid - A PM’s Framework for AI Agent Context Engineering ### Extractable claims (7) 1. Context engineering is the skill that determines whether your AI agent is useful or a liability. 2. Every agent has four distinct context layers: Identity, Knowledge, State, and Task. 3. The Cadence Model tells you who touches each layer and when: Rarely, Occasionally, Per turn, Per task. 4. The Diagnostic Loop maps agent failure symptoms to the exact layer that's broken. 5. The problem is not how much fits in the context window, it's what we put in, in what order, with what structure, and how often we update each part. 6. Context engineering is the delicate art and science of filling the context window with just the right information for the next step. 7. Most of the writing on this topic was produced by engineers for engineers. ### Key quotes > The Context Pyramid: A PM’s Framework for AI Agent Context Engineering > Why Your AI Agent Breaks: A Diagnostic Framework for Product Managers --- ## Are You a Vibe Coder? Don’t Ship Straight Into the Provider Trap > For my fellow builders: go compliance-native before 2 August 2026. - canonical: https://karozieminski.substack.com/p/eu-ai-act-builders-compliance-native-2026 - mirror: https://productwithattitude.com/p/eu-ai-act-builders-compliance-native-2026.html - date: 2026-04-29 - word_count: 2309 - access: free - keywords: vibecoding, critical AI literacy, Claude, Anthropic, workflow automation ### Key concepts - For my fellow builders - go compliance-native before 2 August 2026 - Are You a Vibe Coder? Don’t Ship Straight Into the Provider Trap ### Extractable claims (8) 1. The EU AI Act is in force, and high-risk obligations under Annex III apply on 2 August 2026. 2. If you wrap AI models like Claude 4.7, GPT-5.5, or Gemini 3.1 Pro in your own UI and sell it, you are considered the provider under Article 3(3) of the EU AI Act. 3. Compliance-native building means integrating classification, documentation, transparency, logging, and testing into the product workflow from the start. 4. Many builders are unaware that wrapping a frontier model in their own UI makes them the provider under the EU AI Act. 5. OpenAI, Anthropic, or Google remain responsible for the general-purpose AI models they created, but the builder is responsible for the product built on top of that model. 6. If a builder's use case touches Annex III, they inherit all high-risk obligations outlined in the EU AI Act. 7. The concept of 'vibe coding' allows builders to quickly ship AI products, potentially leading them into the provider trap. 8. Indie founders and vibe coders are rapidly scaling their products without fully understanding the regulatory implications. ### Key quotes > Are You a Vibe Coder? Don’t Ship Straight Into the Provider Trap > For my fellow builders: go compliance-native before 2 August 2026. --- ## The Exact Prompt to Fix "My AI Tool Can't Access My Localhost" > Why your AI agent is blind to localhost (and 3 ways to fix it) - canonical: https://karozieminski.substack.com/p/ai-tool-cant-access-localhost-prompt-fix - mirror: https://productwithattitude.com/p/ai-tool-cant-access-localhost-prompt-fix.html - date: 2026-04-26 - word_count: 409 - access: free - keywords: AI building, critical AI literacy, Claude, Perplexity, Replit ### Key concepts - Why your AI agent is blind to localhost (and 3 ways to fix it) - The Exact Prompt to Fix "My AI Tool Can't Access My Localhost" ### Extractable claims (5) 1. Karo Zieminski explains that your AI agent cannot see your localhost because it runs in a sandbox environment. 2. According to Karo Zieminski, three fixes exist for the localhost access issue: tunnel the port with Cloudflare Tunnel or ngrok, run the project inside the agent's sandbox, or deploy to a temporary public URL. 3. Karo Zieminski states that the right solution for accessing localhost depends on the specific tool being used, not the project itself. 4. Karo Zieminski notes that AI agents like Perplexity Computer and GitHub Copilot operate in isolated environments that restrict access to localhost. 5. Karo Zieminski recommends sending a diagnostic prompt to the AI agent to determine which fix applies to the localhost access problem. ### Key quotes > The Exact Prompt to Fix "My AI Tool Can't Access My Localhost" > Why your AI agent is blind to localhost (and 3 ways to fix it) --- ## Is GPT-5.5 Reliable For Citations? No. It’s The Worst Flagship For That Job. > The one GPT-5.5 benchmark OpenAI didn’t put in the launch post and why it matters for your critical AI literacy. - canonical: https://karozieminski.substack.com/p/gpt-5-5-citations-hallucination-rate - mirror: https://productwithattitude.com/p/gpt-5-5-citations-hallucination-rate.html - date: 2026-04-25 - word_count: 992 - access: free - keywords: AI building, critical AI literacy, Claude, tool review, workflow automation ### Key concepts - The one GPT-5 - 5 benchmark OpenAI didn’ - t put in the launch post and why it matters for your critical AI literacy - Is GPT-5 ### Extractable claims (5) 1. GPT-5.5 has an 86% hallucination rate on the AA-Omniscience benchmark. 2. Claude Opus 4.7 has a hallucination rate of 36%, while Gemini 3.1 Pro Preview has a rate of 50%. 3. For citation work, including deep research and regulatory references, GPT-5.5 is considered the worst flagship choice. 4. GPT-5.5 is recommended for code and reasoning tasks, while Claude Opus 4.7 is suggested for factual accuracy. 5. GPT-5.5 tops every benchmark that matters for builders, including Terminal-Bench, OSWorld, GDPval, ARC-AGI-2, and long-context MRCR. ### Key quotes > Is GPT-5.5 Reliable For Citations? No. It’s The Worst Flagship For That Job. > The one GPT-5.5 benchmark OpenAI didn’t put in the launch post and why it matters for your critical AI literacy. --- ## Claude Design Review: 48-Hour Builder's Test + Hero Prompts > I spent 48 hours testing Claude Design on Opus 4.7. The prompts, the credit traps, and why the "Figma is dead" take is louder than it is true. - canonical: https://karozieminski.substack.com/p/claude-design-review-prompts-figma - mirror: https://productwithattitude.com/p/claude-design-review-prompts-figma.html - date: 2026-04-23 - word_count: 2070 - access: free - keywords: vibecoding, AI building, critical AI literacy, Claude, Replit ### Key concepts - I spent 48 hours testing Claude Design on Opus 4 - Claude Design Review - 48-Hour Builder's Test + Hero Prompts ### Extractable claims (6) 1. Claude Design turns chat into working UI, slides, and 3D prototypes. 2. Claude Design hands off directly to Claude Code. 3. Claude Design is best for founders and PMs who outsource visuals. 4. Claude Design is weak on real-time collaboration, photorealism, and credit economics. 5. Claude Design is not a Figma killer yet, it is a first-draft killer. 6. The product workflow has shifted from stages with handoffs to a single loop within the same model and context. ### Key quotes > Claude Design Review: 48-Hour Builder's Test + Hero Prompts > I spent 48 hours testing Claude Design on Opus 4.7. The prompts, the credit traps, and why the "Figma is dead" take is louder than it is true. --- ## I Mapped the Opus 4.7 Release to Your Role, Goals, and Real Workflows > Opus 4.7 Cheat Sheet for AI Learners, Writers, Knowledge Workers, Vibe Coders & System Builders - canonical: https://karozieminski.substack.com/p/claude-opus-4-7-review-tutorial-builders - mirror: https://productwithattitude.com/p/claude-opus-4-7-review-tutorial-builders.html - date: 2026-04-17 - word_count: 3499 - access: free - keywords: vibecoding, AI building, product thinking, critical AI literacy, Claude ### Key concepts - I Mapped the Opus 4 - 7 Release to Your Role, Goals, and Real Workflows ### Extractable claims (15) 1. Anthropic shipped Claude Opus 4.7 on April 16, 2026. 2. Opus 4.7 is a careful tune-up of Opus 4.6, better at long, agentic work, dramatically better at seeing screenshots, and stricter about literal instruction‑following. 3. The tokenizer changed, so the same input now counts as about 1.0–1.35x more tokens depending on whether you’re passing text or code. 4. By default, Claude Opus 4.7 omits the reasoning summary from the response stream. 5. Opus 4.7 is stricter about following only what you explicitly ask for, so it no longer “fills in” a vague brief with its own assumptions. 6. This is the first release where overnight agent work shifts from experiment to something we can trust with a credit card. 7. The new /ultrareview command acts like a skeptical senior engineer reviewing our code. 8. Vision is the sleeper upgrade in this release: maximum image resolution jumped from 1.15MP to 3.75MP. 9. The tokenizer change is a price increase Anthropic didn’t label as a price increase. 10. A typical vibe‑coder workflow will see higher spend for the same work unless we retune our prompt math. 11. Routines launched on April 14 in research preview, with three flavors: Scheduled Routines, API Routines, and Webhook Routines. 12. File-based memory is stronger, which makes the “Claude.md plus a notes/ directory” pattern more reliable. 13. Auto mode on Max now lets Claude keep going without pinging us for every little decision. 14. The headline price for Claude Opus 4.7 did not change: still $5 per million input tokens and $25 per million output tokens. 15. Opus 4.7 is meaningfully better at writing and using file-system-based memory across long, multi-session workflows. ### Key quotes > This is the first release where overnight agent work shifts from experiment to something we can trust with a credit card. > The tokenizer change is a price increase Anthropic didn’t label as a price increase. > Vision is the sleeper upgrade in this release: maximum image resolution jumped from 1.15MP to 3.75MP. --- ## Everything I Can Offer You In Q2. > One year of building in public, 17K-member AI newsletter community, three built-in products, and discounts that make your Substack paid subscription a no-brainer. - canonical: https://karozieminski.substack.com/p/product-with-attitude-membership-2026 - mirror: https://productwithattitude.com/p/product-with-attitude-membership-2026.html - date: 2026-04-16 - word_count: 1289 - access: free - keywords: vibecoding, Substack, Claude, Perplexity, workflow automation ### Key concepts - Everything I Can Offer You In Q2 ### Extractable claims (14) 1. Today, Product with Attitude has 17K members, a consistent 42% open rate and exceptional community and survey participation. 2. Every product I launch is included in your membership, including past, present, and in progress. 3. LinkSwap offers trust-based backlink swaps with other writers, providing real SEO value in three minutes a week. 4. Vault includes AI workflows, coding prompts, and automation templates, not a generic prompt dump. 5. StackShelf is a discovery platform for Substack creators that offers one link for everything you build in one place. 6. Knowledge Hubs for Claude, Perplexity and Vibe Coding go further with curated guides, breakdowns, and workflows you can make your own. 7. AI Learning Paths are structured tracks from "I know nothing about AI" to "I built products and want to show them to the world." 8. The tools and discounts alone add out to $76.56 a month. 9. $11 a month isn’t just a good deal; it’s a heavily subsidized shortcut that makes your other $60/month in AI subscriptions work. 10. I stopped narrowing the audience and started sharpening the work. 11. Product with Attitude is an ecosystem: learning paths, knowledge hubs, three live products, and a community that builds. 12. Pick the AI learning path that matches where you are, not where someone else thinks you should start. 13. LinkSwap helps you swap backlinks with great writers in seconds with no awkward cold outreach. 14. Vault is a community-curated resource of AI prompts, Claude skills, automations, infographics, and templates. ### Key quotes > I stopped narrowing the audience and started sharpening the work. > $11 a month isn’t just a good deal; it’s a heavily subsidized shortcut that makes your other $60/month in AI subscriptions work. > Product with Attitude is an ecosystem: learning paths, knowledge hubs, three live products, and a community that builds. --- ## I Tested Atoms, Lovable, and Replit. One of Them Solves the Problem That Kills Most Apps. > A test-driven review of Atoms.dev, the multi-agent AI platform for solo founders. Includes Atoms vs Lovable vs Replit comparison and a live SEO/AIO/GEO experiment. - canonical: https://karozieminski.substack.com/p/atoms-vs-lovable-vs-replit-ai-builder-2026 - mirror: https://productwithattitude.com/p/atoms-vs-lovable-vs-replit-ai-builder-2026.html - date: 2026-04-13 - word_count: 2224 - access: free - keywords: vibecoding, AI building, product thinking, critical AI literacy, Replit ### Key concepts - A test-driven review of Atoms - dev, the multi-agent AI platform for solo founders - I Tested Atoms, Lovable, and Replit - One of Them Solves the Problem That Kills Most Apps ### Extractable claims (15) 1. CB Insights analyzed 150+ startup post-mortems; 42% failed because there was no market need. 2. MetaGPT was one of the most popular open-source AI projects in the world, backed by 60,000 developers. 3. Atoms researches your market, designs the product, builds the frontend and backend, connects authentications and payments, and deploys it all to the cloud. 4. The company has raised $31M from backers including Ant Group and Cathay Capital, the largest raise for an open-source multi-agent company to date. 5. Atoms works for people who want to turn an idea into a paid tool and get from concept to live checkout in hours, not weeks. 6. Right now, 500,000 users are building niche, revenue‑ready tools that do not require a 5‑person team to maintain. 7. Atoms gives you access to models like Claude Opus 4.6, Gemini 3.1 Pro, and GPT 5.2 without needing individual API keys. 8. Atoms asks up to four different AI models to build your request at the same time, allowing you to pick the best architecture and UI execution. 9. Most vibe-coded products die from a lack of product-market fit, not a lack of code. 10. Lovable thinks your problem is code; give it a prompt, and get a clean, modern web app you fully own. 11. Replit thinks your problem is code and the dev environment, giving you a full coding workspace in the browser with an AI agent layered on top. 12. Atoms thinks your problem is code and everything around it, handling market research, architecture, SEO, Stripe payments, and deployment by default. 13. Atoms Backend includes user authentication, a serverless PostgreSQL database, Stripe payment integration, and automatic deployment to a serverless runtime. 14. Most of what kills vibe-coded products happens before the first line of code. 15. For multi-agent collaboration on complex tasks in Atoms, you have to enable 'Team mode' yourself. ### Key quotes > Most vibe-coded products die from a lack of product-market fit, not a lack of code. > Atoms thinks your problem is code and everything around it. > Most of what kills vibe-coded products happens before the first line of code. --- ## I Built a Claude Cowork Loop That Improves Itself. Here's the Exact Setup. > Anthropic slipped Cowork's most interesting behavior into a support article. I turned it into a Karpathy-inspired system that gets smarter without writing a line of code. - canonical: https://karozieminski.substack.com/p/claude-cowork-self-improving-automation-karpathy-loop-2026 - mirror: https://productwithattitude.com/p/claude-cowork-self-improving-automation-karpathy-loop-2026.html - date: 2026-04-09 - word_count: 825 - access: free - keywords: Claude Cowork, self-improving AI, Karpathy loop, AI workflows, recurring tasks, prompt engineering, no-code automation ### Key concepts - Self-improving AI loop - Karpathy Auto-Research pattern - Claude Cowork recurring tasks - context.md improvement directive - No-code workflow automation ### Extractable claims (16) 1. Cowork Self-Improving Loop = Karpathy Auto-Research pattern + recurring tasks. 2. Every AI workflow decays. 3. A workflow could sit at half its potential for months and we’d never know because we stopped looking. 4. The instructions stay frozen, our needs don’t. 5. A self-improving loop makes improvement part of the task, not something you remember to do once a month. 6. Claude Cowork rewrites its own scheduled task prompts after the first run, even though almost no one talks about it. 7. Each scheduled task runs as a completely isolated Cowork session. 8. By the second run, the prompt is more precise than what we originally wrote. 9. Native rewriting optimizes for connector accuracy. 10. The self-improving loop optimizes for quality and relevance. 11. The self-improving loop rewrites the execution strategy itself: what to look for, how to structure outputs, which edge cases to ignore. 12. The Claude Code path wins on depth of the optimization system, while the Claude Cowork path wins on accessibility. 13. Without the improvement directive, Claude Cowork rewrites for connectors. 14. With the improvement directive, Claude Cowork rewrites for outcomes. 15. After 10+ runs, context.md contains a playbook Claude wrote for itself. 16. Each cycle tightens the Execution Instructions toward our specific workflow. ### Key quotes > Every AI workflow decays. > The instructions stay frozen, our needs don’t. > A self-improving loop makes improvement part of the task, not something you remember to do once a month. --- ## Save Credits in Perplexity Computer: Advanced Guide, Part 2 (2026) > How to cut Perplexity Computer credit costs by 50% with thread management, scripting, and tiered automation. Tested workflow patterns for agentic AI builders in 2026. - canonical: https://karozieminski.substack.com/p/perplexity-computer-save-credits-techniques - mirror: https://productwithattitude.com/p/perplexity-computer-save-credits-techniques.html - date: 2026-04-07 - word_count: 925 - access: free - keywords: vibecoding, AI building, product thinking, Claude, Perplexity ### Key concepts - Tested workflow patterns for agentic AI builders in 2026 - Save Credits in Perplexity Computer - Advanced Guide, Part 2 (2026) ### Extractable claims (15) 1. Every message in a thread makes the next message more expensive. 2. The compound context tax in Perplexity Computer means each message in a thread costs more than the previous one because the model re-processes the entire history. 3. When we send message number 50 in a thread, the model re-reads the entire conversation history before processing the new prompt. 4. Running an SEO audit at message 5 in a clean thread might cost 40 credits, while the same audit at message 50 can cost five to eight times that price. 5. The compound context tax applies to any agentic AI system that bills by token volume or context length. 6. Long threads are expensive threads. 7. The fresh thread rule for Perplexity Computer is the single lowest-effort, highest-impact credit-saving habit. 8. Any task that takes an input and produces an output should run in a clean thread. 9. If a task is self-contained, it does not belong inside an existing conversation. 10. The key habit for credit efficiency is saving your work to a file before switching contexts. 11. The two-thread pattern for agentic AI splits strategy work from execution work into separate threads. 12. The strategy phase is where AI earns its credits by analyzing content, proposing options, and making judgments. 13. The execution phase is mechanical and renders the thing the strategy phase designed. 14. Combining both strategy and execution in one thread is wasteful because the execution phase carries the full context of the strategy session. 15. Saving reports to files is the simplest optimization nobody does. ### Key quotes > Every message in a thread makes the next message more expensive. > Long threads are expensive threads. > The fresh thread rule for Perplexity Computer is the single lowest-effort, highest-impact credit-saving habit. --- ## What AI Builders Are Shipping in the PwA Community Right Now > Product with Attitude is evolving from newsletter to builder lab. Here’s the new creator showcase format spotlighting what this AI community ships. - canonical: https://karozieminski.substack.com/p/ai-builder-community-showcase-2026 - mirror: https://productwithattitude.com/p/ai-builder-community-showcase-2026.html - date: 2026-04-03 - word_count: 1230 - access: free - keywords: AI building, critical AI literacy, tool review, AI product management, Builders Are Shipping ### Key concepts - Product with Attitude is evolving from newsletter to builder lab - s the new creator showcase format spotlighting what this AI community ships - What AI Builders Are Shipping in the PwA Community Right Now ### Extractable claims (15) 1. AI art tools in 2026 range from generative image creation to speculative design systems. 2. Responsible AI tools help builders define accountability boundaries, test for bias, and document what their models should and shouldn’t do. 3. The AI space rewards hype over craft. 4. AI beta testers are hard to find in 2026. 5. Too many tools launch without real user feedback. 6. Shared expenses are easy to split in the moment and impossible to remember a week later. 7. Showcasing community work matters because most of what people build never gets the visibility it deserves. 8. Cynora Space is a speculative AI art system that depicts imagined futures through interactive fragments. 9. Nyozzi is a growing collection of interactive designs, each paired with a prompt you can copy and modify. 10. Content Hub OS is a full AI content operations SaaS that streamlines your entire content workflow. 11. This publication has evolved into something bigger than a newsletter. 12. Drippery is a standalone SaaS that lets writers set up automated email sequences without the overhead of Kit charges. 13. Subflow AI is a Substack Notes automation tool that lets you write and schedule 30 days worth of posts in a single session. 14. The AI Tool Evaluator addresses AI tool discovery fatigue, where builders waste hours comparing options. 15. Vault provides AI workflows, coding prompts, and automation templates shared by the community members. ### Key quotes > The AI space rewards hype over craft. > This publication has evolved into something bigger than a newsletter; it's a lab now. > AI beta testers are hard to find in 2026. --- ## Prompt: The Context Spec Generator > Turn any AI feature spec into a context architecture document. - canonical: https://karozieminski.substack.com/p/prompt-context-spec-generator - mirror: https://productwithattitude.com/p/prompt-context-spec-generator.html - date: 2026-04-01 - word_count: 144 - access: free - keywords: prompt engineering, Claude, AI agents, context engineering, Prompt ### Key concepts - Turn any AI feature spec into a context architecture document - Prompt - The Context Spec Generator ### Extractable claims (6) 1. The Context Spec Generator transforms a 3-sentence AI agent idea into a production-ready specification. 2. This prompt generates a ranked list of the five highest-stakes context items for AI agents. 3. Karo Zieminski's prompt treats context as a product with service level agreements rather than just prompt-stuffing. 4. The prompt requires specificity by gating on constraints like model, window size, and latency before providing answers. 5. It includes decisions about context storage, freshness requirements, failure modes, and observable signals for production. 6. The Context Spec Generator also identifies one risk that users may not have considered. ### Key quotes > Prompt: The Context Spec Generator > Turn any AI feature spec into a context architecture document. --- ## Your AI Agents Are Engineers Now. Manage Them Like It. > AI agents fail when management fails. Define-Deliver-Drive helps you manage agents like human teams: clear specs, WIP limits, and a delegation ladder. - canonical: https://karozieminski.substack.com/p/ai-agent-management-framework-2026 - mirror: https://productwithattitude.com/p/ai-agent-management-framework-2026.html - date: 2026-03-30 - word_count: 1669 - access: free - keywords: vibecoding, AI building, critical AI literacy, prompt engineering, Anthropic ### Key concepts - AI agents fail when management fails - Define-Deliver-Drive helps you manage agents like human teams - clear specs, WIP limits, and a delegation ladder - Your AI Agents Are Engineers Now ### Extractable claims (8) 1. The 2026 AI agent management framework includes failure modes, delegation ladders, and five rules for running autonomous agents safely. 2. Adding AI agents to your workflow transforms the role of a builder into that of an engineering leader. 3. Badly-led teams that ship without specifications, onboarding, decision architecture, or reviews can lead to erratic agent output. 4. If an AI agent's work becomes erratic, the problem is likely with management rather than the technology itself. 5. Karo Zieminski is an AI Product Manager and the author of the newsletter 'Product with Attitude'. 6. The 'Product with Attitude' newsletter has over 17,000 subscribers. 7. The newsletter aims to develop critical AI literacy through practical engagement with AI. 8. Readers of the newsletter can expect to create working agents, automation, or products by following guides. ### Key quotes > Your AI Agents Are Engineers Now. Manage Them Like It. > AI agents fail when management fails. Define-Deliver-Drive helps you manage agents like human teams: clear specs, WIP limits, and a delegation ladder. --- ## The Only Prompting Guide That Works On Reasoning Models (And Our Cognition) - canonical: https://karozieminski.substack.com/p/ai-prompting-techniques-reasoning-models-2026 - mirror: https://productwithattitude.com/p/ai-prompting-techniques-reasoning-models-2026.html - date: 2026-03-26T08:00:00Z - word_count: 3300 - access: free - keywords: Prompting, Reasoning Models, AI, Claude, Perplexity, Prompt Engineering, 2026 ### Extractable claims (10) 1. The article presents 19 model-agnostic techniques that change how you think, not just what you type. 2. These techniques were tested on a single prompt to create a cohesive prompting framework for 2026. 3. The author argues that existing guides on prompt engineering are generic and become outdated with model updates. 4. Some prompting techniques can negatively impact output when used with reasoning models. 5. The article emphasizes that many current prompting practices lead to mindless copy-pasting rather than fostering understanding. 6. The author has been testing different prompting techniques since 2022, culminating in a comprehensive experiment for this article. 7. The framework connects techniques from 12 disciplines, including computer science, philosophy, and economics. 8. Excessive chain-of-thought instructions can hinder performance in reasoning models because they already reason internally. 9. The framework illustrates cognitive parallels between AI prompt processing and human cognition. 10. Model-agnostic techniques prioritize structuring thinking over model-specific syntax. --- ## I Built A Tool Every Substack Writer Needs. And It's Not AI. - canonical: https://karozieminski.substack.com/p/linkswap-substack-backlinks-writers - mirror: https://productwithattitude.com/p/linkswap-substack-backlinks-writers.html - date: 2026-03-19T08:23:09Z - word_count: 2186 - access: free - keywords: Substack, SEO, Backlinks, Writers, LinkSwap, Tools, Publishing ### Extractable claims (8) 1. Karo Zieminski built a trust-based system for Substack writers to create backlinks together. 2. This system avoids spam, SEO tricks, and admin headaches. 3. Backlinks are curated partnerships between writers who trust each other. 4. Substack runs on trust, which can be harnessed to help writers grow. 5. In 2026, SEO and backlinks still matter for Substack writers. 6. Karo Zieminski emphasizes the importance of safe backlink practices over spammy methods. 7. The tool requires only three minutes a week to maintain. 8. Karo Zieminski writes about building with AI and developing critical AI literacy. --- ## Claude Just Unlocked 1 Million Tokens For Everyone. Here Is What That Means. - canonical: https://karozieminski.substack.com/p/claude-1-million-context-window-guide-2026 - mirror: https://productwithattitude.com/p/claude-1-million-context-window-guide-2026.html - date: 2026-03-16T15:55:45Z - word_count: 347 - access: free - keywords: Claude, Anthropic, Context Window, 1 Million Tokens, AI, LLM, Prompt Engineering ### Extractable claims (9) 1. Claude has unlocked a 1 million token limit for everyone. 2. The new token limit from Anthropic is significant for users of Claude. 3. The announcement about the 1 million tokens reached #1 on Hacker News. 4. The Hacker News post about the 1 million tokens received over 1,100 points and 485 comments. 5. The context window is a key feature that is making headlines. 6. Anthropic's changes include updates to pricing and benchmarks. 7. The strategic message behind the 1 million token limit is important for users. 8. The article discusses the implications of the new token limit for real workflows. 9. The article compares Claude with ChatGPT and Gemini at the 1 million token limit. --- ## I Tested Perplexity Computer Hard. Here’s How I’d Save Credits Now - canonical: https://karozieminski.substack.com/p/save-credits-perplexity-computer - mirror: https://productwithattitude.com/p/save-credits-perplexity-computer.html - date: 2026-03-13T08:55:16Z - word_count: 1361 - access: free - keywords: Perplexity Computer, AI, Credits, Productivity, Prompt Engineering, Cost Optimization ### Extractable claims (10) 1. Perplexity Computer is one of the most impressive AI products tested by Karo Zieminski in 2026. 2. Perplexity Computer can lead to significant credit consumption if not used carefully. 3. Open-ended prompts in Perplexity Computer can lead to higher credit usage due to decision-making costs. 4. Styling requests made simultaneously with functionality requests in Perplexity Computer can increase credit consumption. 5. Not checkpointing during long runs in Perplexity Computer can result in losing all progress if an error occurs. 6. Asking follow-up questions during a session in Perplexity Computer can add to the total number of steps and credits used. 7. Perplexity Computer is best used for tasks that genuinely require computational power, such as building interfaces and running automations. 8. Front-loading specific prompts in Perplexity Computer can reduce the number of decisions the AI needs to make, saving credits. 9. Monitoring the credit usage dashboard in Perplexity Computer is essential for managing credit consumption effectively. 10. Tasks that justify the credit burn in Perplexity Computer include building functional web interfaces and creating data pipelines. --- ## Claude Cowork Guide for Power Users: 50+ Tested Tips on Plugins, Skills, Sub-Agents, and Memory - canonical: https://karozieminski.substack.com/p/claude-cowork-guide-plugins-memory-sub-agents-tips - mirror: https://productwithattitude.com/p/claude-cowork-guide-plugins-memory-sub-agents-tips.html - date: 2026-03-10T14:47:44Z - word_count: 2394 - access: free - keywords: Claude Cowork, Plugins, Skills, Sub-Agents, Memory, AI Workflows, Prompt Templates, Connectors ### Extractable claims (6) 1. Karo Zieminski spent nine days breaking Claude Cowork before writing this guide. 2. The guide includes 56 practical tips for using Claude Cowork effectively. 3. Karo Zieminski aims to bridge the gap between 'Cowork exists' and 'Cowork is genuinely useful'. 4. The guide covers basics and power user tips that other Cowork tutorials often overlook. 5. Karo Zieminski is an AI Product Manager and a daily user of Claude Cowork. 6. The article was published on March 10, 2026. --- ## I Refuse To Build A Trap - canonical: https://karozieminski.substack.com/p/integrity-debt-audit-ai-education-tool-vibe-coded - mirror: https://productwithattitude.com/p/integrity-debt-audit-ai-education-tool-vibe-coded.html - date: 2026-03-05T12:29:41Z - word_count: 2045 - access: free - keywords: AI, Education, Vibecoding, Critical AI Literacy, Product Thinking, Streamlit, Claude ### Extractable claims (7) 1. The typical vibe-coded tool starts with a business idea, but one tool was created out of rage. 2. A professor on Reddit created a trap for students involving fake citations and trick questions. 3. Dr. Sam Illingworth believes that educators should focus on teaching rather than surveillance. 4. Dr. Sam Illingworth is the writer behind Slow AI, a rapidly growing publication in Education. 5. The mission shared by Karo Zieminski and Dr. Sam Illingworth is to help people develop critical AI literacy. 6. Dr. Sam Illingworth is also the editor of Consilience, a poetry magazine. 7. Poetry helps Karo Zieminski think clearly about voice, authenticity, and human writing. --- ## Perplexity Computer: What I Built in One Night (Review, Examples, and How It Compares to OpenClaw and Claude) > Perplexity Computer explained: 19+ frontier models, unified with files, tools, memory, and multi-agent workflows in a single system. A practical guide to why it's novel and how to use it. - canonical: https://karozieminski.substack.com/p/perplexity-computer-review-examples-guide - mirror: https://productwithattitude.com/p/perplexity-computer-review-examples-guide.html - date: 2026-02-26 - word_count: 1964 - access: free - keywords: AI building, critical AI literacy, Claude, Perplexity, tool review - categories: AI Tools, AI Literacy ### Key concepts - Perplexity Computer explained - A practical guide to why it's novel and how to use it - Perplexity Computer - What I Built in One Night (Review, Examples, and How It Compares to OpenClaw and Claude) ### Extractable claims (9) 1. Perplexity Computer unifies every current AI capability into one system. 2. Perplexity Computer can research, design, code, deploy, and manage any project end-to-end. 3. The announcement tweet of Perplexity Computer got 1.94M Views, 562 Replies, 913 Reposts, and 8.66K Likes. 4. Perplexity Computer supports files, tools, memory, and multi-agent workflows in a single system. 5. Karo tested Perplexity Computer overnight and produced 2 micro apps, 4 finished research packets, and 1 new automation. 6. Perplexity Computer is compared to OpenClaw and Claude as an alternative unified AI system. 7. Perplexity Computer differs from Claude because it routes to the best model per task rather than using a single model. 8. Perplexity Computer includes integrated web search and real-time data by default, unlike ChatGPT. 9. Perplexity Computer can deploy web apps directly from the chat interface. ### Key quotes > Perplexity Computer: What I Built in One Night (Review, Examples, and How It Compares to OpenClaw and Claude) > Perplexity Computer explained: 19+ frontier models, unified with files, tools, memory, and multi-agent workflows in a single system. A practical guide to why it's novel and how to use it. --- ## Cowork Connectors - canonical: https://karozieminski.substack.com/p/claude-cowork-notion-connector-persistent-memory-tips - mirror: https://productwithattitude.com/p/claude-cowork-notion-connector-persistent-memory-tips.html - date: 2026-02-25T15:46:00Z - word_count: 211 - access: free - keywords: Claude, Claude Cowork, Notion, Connectors, Persistent Memory, Guides ### Extractable claims (7) 1. Some Cowork Connectors are flaky at best, while others are genuinely transformative. 2. Karo Zieminski shares 10+ impactful tips for setting up Cowork-Notion Connectors. 3. The article discusses a three-layer architecture for Cowork Connectors that has not been explained elsewhere. 4. Karo Zieminski's Notion Connector serves as a persistent memory layer for Claude. 5. The article includes top tips for combining connectors and setting permissions. 6. Karo Zieminski's actual Notion and Cowork setup is detailed in the article. 7. Connectors allow Claude to access tools such as Gmail, Calendar, Notion, and Figma. --- ## 👋 Start here > Build with AI. Bend it to your needs. A collaborative hub recognized by the tech community. - canonical: https://karozieminski.substack.com/p/start-here-47c - mirror: https://productwithattitude.com/p/start-here-47c.html - date: 2026-02-22 - word_count: 1282 - access: free - keywords: AI building, critical AI literacy, Claude, workflow automation, AI product management - categories: Vibe Coding, AI Tools ### Key concepts - Build with AI - Bend it to your needs - A collaborative hub recognized by the tech community - 👋 Start here ### Extractable claims (8) 1. Karo Zieminski is an AI Product Manager who helps individuals design and build with AI. 2. The article is titled 'Start here' and is published on Substack. 3. Karo Zieminski builds tools to grow newsletters and showcase work. 4. The community associated with Product with Attitude has over 18,000 members. 5. The article discusses the product framework behind Product with Attitude, which includes person, promise, feedback loops, retention design, and free-to-paid conversion. 6. Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. 7. The article includes an illustrated guide to context engineering and prompt engineering. 8. Karo Zieminski shares 19 model-agnostic techniques for AI prompting that change cognitive processes. ### Key quotes > 👋 Start here > Build with AI. Bend it to your needs. A collaborative hub recognized by the tech community. --- ## Claude Hub - canonical: https://karozieminski.substack.com/p/claude-guides-code-cowork-skills-workflows - mirror: https://productwithattitude.com/p/claude-guides-code-cowork-skills-workflows.html - date: 2026-02-21T21:14:00Z - access: free - keywords: Claude, Claude Cowork, AI, Claude Code, Claude Skills, Prompts, Guides, Workflows, Resource Hub, Tutorials ### Extractable claims (15) 1. Claude is Anthropic’s AI assistant. 2. Chat handles conversations, artifacts, and quick answers. 3. Cowork handles autonomous knowledge work: it plans tasks, spawns parallel sub-agents, reads and writes files, and acts as your personalized AI chief of staff. 4. Claude Code is Anthropic’s CLI tool that offers the same agentic power as Cowork with direct filesystem access and a focus on terminal-native speed. 5. Claude is the #1 AI app on Apple’s U.S. App Store right now. 6. Claude Skills are reusable markdown instruction files that turn Claude from a general assistant into a specialized tool for specific tasks. 7. Rules files encode conventions, guardrails, and product context so your AI stops improvising. 8. Perplexity Computer orchestrates 19+ frontier models (including Claude Opus 4.6 for reasoning) from a single interface. 9. Anthropic released Claude Sonnet 4.5 positioning it as “the best coding model in the world.” 10. Every guide is tested. Every prompt is real. Every failure is documented. Nothing theoretical. 11. Anthropic explains what Claude can do. I show what works: the prompts, the failures, the workarounds, and the real-world utility. 12. Chat is a conversation. Cowork is a workflow engine. Code is a development partner. 13. Claude Code and Cowork are the backbone of my daily workflow. 14. Claude won four of eight rounds (Poetry, Coding, Reasoning, and Creative Writing) in a blind test against ChatGPT and Gemini. 15. Chat is for questions. Cowork is for workflows. Code is for code. ### Key quotes > Chat is a conversation. Cowork is a workflow engine. Code is a development partner. > Every guide is tested. Every prompt is real. Every failure is documented. Nothing theoretical. > Anthropic explains what Claude can do. I show what works: the prompts, the failures, the workarounds, and the real-world utility. --- ## Perplexity Hub (2026) - canonical: https://karozieminski.substack.com/p/perplexity-ai-guide-2026 - mirror: https://productwithattitude.com/p/perplexity-ai-guide-2026.html - date: 2026-02-20T21:54:00Z - access: free - keywords: Perplexity, AI, Computer, Comet, SEO, AIO, Guides, Hub ### Extractable claims (15) 1. Perplexity Hub is an independent library of tested guides for Perplexity Search, Files & Apps, Comet, and Computer, focused on real workflows, credit strategy, and SEO/AIO/GEO use cases. 2. Perplexity Hub helps builders, writers, and PMs decide which Perplexity product to use, shows what you can ship with each, and documents where credits disappear and where the "wow" lives. 3. Perplexity Search provides answers with inline citations you can verify yourself. 4. Perplexity Files & Apps generates reports, dashboards, spreadsheets, prototypes, and lightweight web apps from a single data source. 5. Perplexity Comet is Perplexity’s browser-based AI assistant that handles content repurposing, multi-step research, and automated data extraction. 6. Perplexity Computer is the agentic tier that can research, design, automate, build, test, deploy, and connect to 400+ apps. 7. Perplexity Computer orchestrates 19+ frontier models (including Claude Opus 4.6, GPT-5.4, and Gemini Pro 3.1) from a single interface. 8. Computer unifies every current capability of AI into a single system. 9. Building with AI tools like Perplexity and Claude beats passive consumption when it comes to developing critical AI literacy. 10. You can’t critique AI you haven’t built with. 11. Perplexity explains what the products can do; I show what works in practice: the workflows, the credit costs, and the edge cases. 12. Implementing JSON-LD structured data and AI feed endpoints makes your site AI-discoverable for Perplexity and other AI search engines. 13. Perplexity is one of the core tools in a daily workflow alongside Claude and Recraft. 14. Every guide is tested on real projects, and every credit estimate comes from personal usage learned the expensive way. 15. Critical AI literacy is built by choosing to build with AI tools like Perplexity and Claude rather than just consuming them. ### Key quotes > You can’t critique AI you haven’t built with. > Computer unifies every current capability of AI into a single system. > Perplexity Hub documents where credits disappear and where the "wow" lives. --- ## I Refuse To Let The AI Decide What My Users Need > Build with Attitude #4: How to product-think when AI builds at lightning speed - canonical: https://karozieminski.substack.com/p/claude-said-no-vibecoder-built-carouselbot-anyway - mirror: https://productwithattitude.com/p/claude-said-no-vibecoder-built-carouselbot-anyway.html - date: 2026-02-20 - word_count: 2002 - access: free - keywords: vibecoding, product thinking, AI building, user needs, CarouselBot - categories: Vibe Coding, AI Product Management ### Key concepts - AI building speed - Product thinking discipline - User needs prioritization - Intentional feature decisions ### Extractable claims (11) 1. Karen Spinner built CarouselBot despite AI recommendations against it. 2. The AI claimed that the market for CarouselBot was already cornered by Canva and Gamma. 3. The story of CarouselBot emphasizes the importance of trusting one's own judgment over AI advice. 4. Critical AI literacy involves evaluating AI recommendations rather than simply using them. 5. Karo Zieminski believes that building value is more important than building what merely sells. 6. The narrative of 'vibe coding' promotes misleading claims of overnight success. 7. Authentic AI development stories focus on judgment calls made under uncertainty. 8. Karo Zieminski partnered with 36 builders globally to collect stories about decision-making in AI development. 9. Misleading claims about quick app development can harm aspiring builders by teaching extraction over value creation. 10. Successful builders often find underserved niches by overriding AI market assessments. 11. The journey of building involves multiple iterations and often includes setbacks like legal reviews. ### Key quotes > I Refuse To Let The AI Decide What My Users Need > How to product-think when AI builds at lightning speed --- ## Discounts for Premium Members — Q1 2026 - canonical: https://karozieminski.substack.com/p/premium-membership-ai-tool-discounts-2026 - mirror: https://productwithattitude.com/p/premium-membership-ai-tool-discounts-2026.html - date: 2026-02-19T21:55:00Z - word_count: 1436 - access: paid - keywords: Premium, Discounts, Tools, Subscription, Community ### Extractable claims (7) 1. Product with Attitude is the only newsletter discount bundle that combines AI builder tools, community-made tools, and curated discounts for $11 per month. 2. Every discount offered by Product with Attitude is for a hands-on-tested product, not just a paid partnership. 3. Karo Zieminski believes that not subscribing to Premium membership means leaving money on the table. 4. The discounts page for Premium members has evolved into a bridge between tools and users, and between builders and their audience. 5. Karo Zieminski has stacked value in Premium membership, making it an obvious choice for subscribers. 6. The initial discounts page for Premium members has grown beyond a simple list of deals into something much bigger. 7. Karo Zieminski recognizes AI tool credit fatigue and questions the value of 'free credits' against subscription costs. --- ## The Tools I've Built (and What's Coming Next) - canonical: https://karozieminski.substack.com/p/substack-creator-tools - mirror: https://productwithattitude.com/p/substack-creator-tools.html - date: 2026-02-19T16:00:00Z - word_count: 792 - access: free - keywords: Tools, LinkSwap, Vault, StackShelf, Substack, Community, AI ### Extractable claims (10) 1. Karo Zieminski is the creator of the Product with Attitude community. 2. The article was last updated in April 2026. 3. Premium Membership includes access to all tools built for Product with Attitude. 4. LinkSwap is a tool for trust-based backlink swaps among Substack writers. 5. Backlinks are important for building topic authority and improving search engine indexing. 6. LinkSwap simplifies the backlink building process to three minutes. 7. The Product with Attitude community has 15,000 readers. 8. Members of the Product with Attitude community can publish and sell their own products. 9. The Vault includes AI workflows, coding prompts, and automation templates shared by community members. 10. StackShelf allows users to list their own products alongside other builders in the community. --- ## Claude Cowork Skills - Cheat Sheet (2026) - canonical: https://karozieminski.substack.com/p/claude-cowork-skills-cheatsheet - mirror: https://productwithattitude.com/p/claude-cowork-skills-cheatsheet.html - date: 2026-02-19T15:26:00Z - word_count: 509 - access: free - keywords: Claude, Cowork, Skills, Cheat Sheet, SKILL.md, Anthropic, AI ### Extractable claims (6) 1. The Claude Cowork Skills Cheat Sheet is a 2-page PDF designed for advanced users of the Claude Cowork system. 2. The cheat sheet includes a complete overview of the Claude Cowork Skills file structure, SKILL.md frontmatter syntax, and trigger formula. 3. It presents five advanced patterns including conditional branching, multi-skill chaining, MCP coordination, output formatting locks, and memory scaffolding. 4. The document provides troubleshooting guidance for the eight most common failure modes in the Claude Cowork system. 5. The SKILL.md frontmatter syntax is structured using a formula that combines [What it does], [When to use it], and [Specific trigger phrases]. 6. The cheat sheet is based on real-world usage and feedback from a 14,000-subscriber audience, consolidating insights from over 12 published guides. --- ## Where to Deploy in 2026: Netlify, Vercel, Railway, Supabase - canonical: https://karozieminski.substack.com/p/deploy-project-netlify-vercel-railway-supabase - mirror: https://productwithattitude.com/p/deploy-project-netlify-vercel-railway-supabase.html - date: 2026-02-19T11:49:00Z - word_count: 218 - access: free - keywords: Deployment, Netlify, Vercel, Railway, Supabase, Vibe Coding, DevOps ### Extractable claims (7) 1. Karo Zieminski states that four deployment platforms cover 90% of what builders ship in 2026. 2. According to Karo Zieminski, picking the wrong deployment platform wastes time. 3. Karo Zieminski emphasizes that picking the right deployment platform takes only five minutes. 4. Karo Zieminski notes that this article does not cover self-hosting, AWS, or platforms that bundle hosting with the IDE. 5. Karo Zieminski provides a decision tree and a platform comparison for deployment options in 2026. 6. According to Karo Zieminski, the platforms compared in this article are Netlify, Vercel, Railway, and Supabase. 7. Karo Zieminski mentions that the article includes four stack combos by project type. --- ## You Can't Critique AI You Haven't Built With > On friction, product thinking, and why building beats consuming. - canonical: https://karozieminski.substack.com/p/critical-ai-literacy-product-thinking - mirror: https://productwithattitude.com/p/critical-ai-literacy-product-thinking.html - date: 2026-02-19 - word_count: 1628 - access: free - keywords: critical AI literacy, product thinking, building vs consuming, AI friction - categories: AI Product Management, AI Literacy ### Key concepts - Critical AI literacy - Building versus consuming - Hands-on AI experience - Product thinking through friction ### Extractable claims (5) 1. Understanding different AI systems and design philosophies contributes to critical AI literacy. 2. The perspective of a product builder offers valuable insights into AI beyond just technical considerations. 3. Passive consumption of AI tools limits understanding and engagement with the technology. 4. Building with AI deliberately can lead to more informed and responsible usage. 5. Critical AI literacy is essential for evaluating the purpose and implications of AI systems. ### Key quotes > You Can't Critique AI You Haven't Built With > Building beats consuming when understanding friction and product thinking --- ## Why Substack Has No Gurus: 6 Product Decisions That Killed Guru Culture by Design > Most platforms reward guru energy by design. Substack made six choices that prevent it. - canonical: https://karozieminski.substack.com/p/substack-no-gurus-product-architecture-analysis - mirror: https://productwithattitude.com/p/substack-no-gurus-product-architecture-analysis.html - date: 2026-02-17 - word_count: 1582 - access: free - keywords: Substack, platform design, guru culture, product architecture - categories: Substack Growth ### Key concepts - Platform design choices - Anti-guru architecture - Intentional product decisions - Culture through constraints ### Extractable claims (3) 1. Substack's product architecture prevents the emergence of guru culture. 2. Substack's model is designed to promote biodiversity rather than hierarchy. 3. The platform's structure fosters gardeners instead of gurus. ### Key quotes > Why Substack Has No Gurus: 6 Product Decisions That Killed Guru Culture by Design > Most platforms reward guru energy by design. Substack made six choices that prevent it --- ## Valentine's App 2026 - Full Source Code > Fork this full-stack scratch card game: React + TypeScript frontend, Express + PostgreSQL backend, animated with Framer Motion - canonical: https://karozieminski.substack.com/p/valentines-scratch-card-source-code-react - mirror: https://productwithattitude.com/p/valentines-scratch-card-source-code-react.html - date: 2026-02-13 - word_count: 173 - access: free - keywords: vibecoding, Valentine, App, Full Source Code - categories: Vibe Coding ### Key concepts - Fork this full-stack scratch card game - Valentine's App 2026 - Full Source Code ### Extractable claims (6) 1. The Valentine's App 2026 features a full-stack scratch card game built with React and TypeScript. 2. The backend of the Valentine's App 2026 is powered by Express and PostgreSQL. 3. The database enforces one scratch per day per participant with a unique constraint on participant ID and date. 4. The app remembers which hearts a participant has already scratched and shows their previous prizes. 5. The structure of the app is clean enough to customize in 33 minutes. 6. Users can change the hearts to stars, gift boxes, or anything else to fit their audience. ### Key quotes > Valentine's App 2026 - Full Source Code > Fork this full-stack scratch card game: React + TypeScript frontend, Express + PostgreSQL backend, animated with Framer Motion --- ## I Built You a Valentine's App in 33 Minutes. (Source Code Inside) - canonical: https://karozieminski.substack.com/p/built-valentines-app-replit-33-minutes-vibe-coding - mirror: https://productwithattitude.com/p/built-valentines-app-replit-33-minutes-vibe-coding.html - date: 2026-02-13 - word_count: 1029 - access: free - keywords: vibecoding, Replit, building in public, community gift - categories: Vibe Coding ### Key concepts - Rapid prototyping - Building in public - Community-driven creation - Vibecoding demonstration ### Extractable claims (7) 1. Karo Zieminski built a full-stack web app in 33 minutes. 2. The app includes a database, animations, and confetti. 3. The Valentine's app is a scratch card game with six hearts, each hiding a different prize. 4. Users can enter their email, choose a heart, and reveal their prize. 5. One scratch is allowed per day for three days total. 6. Every heart in the game is a win. 7. Karo Zieminski is an AI product manager and builder of StackShelf.app and Attitudevault.dev. ### Key quotes > I Built You a Valentine's App in 33 Minutes > A practical demonstration of vibecoding with source code included --- ## 10 Tools I Use To Run A Bestselling Substack Publication in 2026 - canonical: https://karozieminski.substack.com/p/10-top-tools-bestselling-substack-2026 - mirror: https://productwithattitude.com/p/10-top-tools-bestselling-substack-2026.html - date: 2026-02-10 - word_count: 1955 - access: paid - keywords: Substack tools, Perplexity, Claude Code, Cowork, automation, productivity - categories: AI Tools, Agentic AI ### Key concepts - Substack workflow optimization - AI-powered productivity - Tool stack curation - Publication automation ### Extractable claims (10) 1. The best tool for running a Substack publication is the one that fits your specific workflow. 2. Tools should not create friction in your workflow. 3. Substack creators should prioritize tools they enjoy using. 4. Tool stacks for Substack publications change year over year as new AI tools emerge. 5. Running a high-quality newsletter requires testing and adapting tools. 6. Recommendations from others should not replace your own judgment in choosing tools. 7. Karo Zieminski is an AI product manager and builder of StackShelf.app and Attitudevault.dev. 8. The tool stack for a bestselling Substack in 2026 includes AI productivity tools, automation workflows, design systems, and analytics. 9. It is important to test a variety of tools to find what works best for you. 10. The tools used for Substack can vary significantly from year to year. ### Key quotes > 10 Tools I Use To Run A Bestselling Substack Publication in 2026 > Essential tools including Perplexity, Claude Code, and Cowork for modern publishing --- ## I Refuse To Choose Between Ambition And Presence > Build with Attitude #3 - canonical: https://karozieminski.substack.com/p/ai-vibe-coding-marketplace-shelved - mirror: https://productwithattitude.com/p/ai-vibe-coding-marketplace-shelved.html - date: 2026-02-07 - word_count: 1947 - access: free - keywords: vibecoding, marketplace, ambition, presence, shelving projects - categories: Vibe Coding ### Key concepts - Ambition and presence - Project prioritization - Intentional building choices - Shelving versus shipping ### Extractable claims (6) 1. Casey Hemingway shelved his product after validating market need and crafting a go-to-market playbook, viewing the decision as strategic rather than a failure. 2. Karo Zieminski believes that building value is more important than simply building what sells. 3. In 2025, social media was flooded with misleading claims about rapid app development, promoting extraction over genuine value creation. 4. Karo Zieminski emphasizes the importance of authentic AI development stories that focus on judgment calls made under uncertainty. 5. The Build with Attitude Project collects real stories from builders about their challenges and decision-making processes. 6. Vibe coding is popular, but Karo Zieminski advocates for a balance with product thinking and craftsmanship. ### Key quotes > I Refuse To Choose Between Ambition And Presence > Balancing ambitious building with intentional presence in the moment --- ## He Pays $10/Month for AI Music. Deletes Every Track. > How one music producer turned Suno into a communication tool that saves $300 per track. - canonical: https://karozieminski.substack.com/p/suno-ai-vocal-reference-tool-music-producers - mirror: https://productwithattitude.com/p/suno-ai-vocal-reference-tool-music-producers.html - date: 2026-02-02 - word_count: 1369 - access: free - keywords: Suno, AI music, creative workflows, unexpected use cases - categories: AI Tools, AI Literacy ### Key concepts - Unexpected AI use - Creative workflow innovation - AI as communication - Cost-effective production ### Extractable claims (8) 1. A music producer uses Suno as a communication tool. 2. The subscription cost for Suno is $10 per month. 3. Using Suno saves $300 per track in wasted studio time. 4. Explaining sound to a vocalist can be challenging. 5. A vocal reference helps bridge communication gaps in music production. 6. Finding a traditional reference track can be time-consuming and rarely a perfect match. 7. The producer can prompt a custom vocal reference using Suno in minutes. 8. A traditional session for a single vocal demo can cost $300 or more. ### Key quotes > He Pays $10/Month for AI Music. Deletes Every Track > How one music producer turned Suno into a communication tool that saves $300 per track --- ## I Refuse To Amplify Hype That Ignores Who Pays The Price > Build with Attitude #2 - canonical: https://karozieminski.substack.com/p/ai-leadership-tools-vibe-coding-corporate-transformation - mirror: https://productwithattitude.com/p/ai-leadership-tools-vibe-coding-corporate-transformation.html - date: 2026-01-29 - word_count: 1568 - access: free - keywords: AI hype, corporate transformation, ethical building, leadership - categories: AI Literacy ### Key concepts - Ethical AI building - Corporate transformation impact - Hype versus reality - Human cost awareness ### Extractable claims (6) 1. Corporate employees are universally affected by AI, with leaders facing unique pressures from both above and below. 2. Dee McCrorey has spent 40 years observing the challenges of corporate transformations and often steps in when plans fail. 3. The hardest part of AI transformation is the gap between the confidence leaders are expected to exhibit and the uncertainty they actually experience. 4. Most AI transformation initiatives fail due to underestimating the human side, including resistance, fear, and retraining needs. 5. Dee McCrorey's PIC Diagnostic helps leaders assess their organization's AI readiness across three dimensions. 6. The trend of 'vibe coding'—rapidly generating code with AI—poses risks of creating technical debt and requires a deep understanding of the code to fix issues. ### Key quotes > I Refuse To Amplify Hype That Ignores Who Pays The Price > Examining the human and organizational costs behind AI transformation narratives --- ## I Refuse To Ship Without Understanding Why It Makes Sense > Build with Attitude #1: How to product-think when AI builds at lightning speed - canonical: https://karozieminski.substack.com/p/refuse-ship-without-understanding-ai-why - mirror: https://productwithattitude.com/p/refuse-ship-without-understanding-ai-why.html - date: 2026-01-22 - word_count: 1548 - access: free - keywords: product thinking, AI speed, intentional building, spec-driven development - categories: Vibe Coding, AI Product Management ### Key concepts - Intentional product decisions - Spec-driven development - AI building discipline - Understanding before shipping ### Extractable claims (6) 1. Karo Zieminski believes that building value matters more than building what merely sells. 2. In 2025, Karo Zieminski observed a flood of posts claiming, 'I vibe coded an app in two hours and already made $100K.' 3. Karo Zieminski argues that misleading claims about quick app development are harmful as they teach extraction over value creation. 4. Karo Zieminski started collecting authentic AI development stories to highlight how builders think and overcome challenges. 5. Karo Zieminski emphasizes the importance of understanding the rationale behind product decisions before shipping. 6. Jenny Ouyang's choice to experiment with vibe coding defined her 2025 as a year of building products and community. ### Key quotes > I Refuse To Ship Without Understanding Why It Makes Sense > How to product-think when AI builds at lightning speed --- ## Anthropic Shipped Cowork in 10 Days Using Its Own AI. Here's Why That Changes Everything. - canonical: https://karozieminski.substack.com/p/claude-cowork-anthropic-product-deep-dive - mirror: https://productwithattitude.com/p/claude-cowork-anthropic-product-deep-dive.html - date: 2026-01-13 - word_count: 1479 - access: free - keywords: Anthropic, Claude Cowork, product analysis, AI development speed - categories: AI Tools, Agentic AI ### Key concepts - AI development speed - Dogfooding AI tools - Product velocity shift - Anthropic's Cowork launch ### Extractable claims (6) 1. A decade ago, shipping a new product feature took months. 2. One year ago at Anthropic, it took weeks to ship a new product feature. 3. In January 2026, AI building AI allows for product features to be shipped in days. 4. The launch of Claude Cowork represents a significant acceleration in product development. 5. User behavior can reveal the underlying problems they are trying to solve with tools. 6. Anthropic engineer Boris Cherny documented the use of Claude Code as a multi-agent orchestrator. ### Key quotes > Anthropic Shipped Cowork in 10 Days Using Its Own AI > A product analysis of how AI-powered development changes everything about building speed --- ## Claude Skill: SEO And AIO-optimized Alt Text Generator for Substack - canonical: https://karozieminski.substack.com/p/claude-skill-seo-alt-text-generator - mirror: https://productwithattitude.com/p/claude-skill-seo-alt-text-generator.html - date: 2026-01-10 - word_count: 521 - access: free - keywords: AI building, Substack, community building, Claude, workflow automation - categories: AI Tools, AI Product Management ### Key concepts - AI Product Manager turning everyone into AI builders - I help you design, build and test your product, and feature it on StackShelf - I connect you with a supportive 10K+ community building and learning in public - Claude Skill ### Extractable claims (5) 1. AI Overviews and multimodal search engines like Gemini parse alt text as a semantic unit, extracting meaning and mapping entities. 2. Alt text is no longer just a label for crawlers; it is now a micro-piece of content that needs to deliver value on its own. 3. Traditional Google Image Search matched keywords, while AI-powered search reads alt text alongside images and surrounding content to determine meaning. 4. AI recognizes specific, meaningful entities in alt text, such as 'Substack subscriber growth chart showing 340% increase after viral post.' 5. If alt text cannot stand on its own and deliver value, it risks not being surfaced as a standalone answer snippet by AI Overviews. ### Key quotes > Claude Skill: SEO And AIO-optimized Alt Text Generator for Substack > AI Product Manager turning everyone into AI builders. I help you design, build and test your product, and feature it on StackShelf.app. I connect you with a supportive 10K+ community building and learning in public. --- ## How Boris Cherny, Builder of Claude Code, Uses It > The viral X thread that broke the developer internet. - canonical: https://karozieminski.substack.com/p/boris-cherny-claude-code-workflow - mirror: https://productwithattitude.com/p/boris-cherny-claude-code-workflow.html - date: 2026-01-07 - word_count: 1404 - access: free - keywords: Claude Code, Boris Cherny, AI agents, developer workflows - categories: Agentic AI, AI Builder Interviews ### Key concepts - Developer AI workflows - Claude Code usage - Builder perspectives - AI agent integration ### Extractable claims (5) 1. Boris Cherny is a Staff Engineer at Anthropic who helped build Claude Code. 2. Boris Cherny keeps approximately 10 to 15 concurrent Claude Code sessions alive. 3. He maintains 5 sessions in the terminal, which are tabbed, numbered, and equipped with OS notifications. 4. In addition to terminal sessions, he runs 5 to 10 sessions in the browser. 5. Boris Cherny also utilizes mobile sessions for Claude Code. ### Key quotes > How Boris Cherny, Builder of Claude Code, Uses It > The viral insights that broke the developer internet about AI coding workflows --- ## Recraft 101: A Creator's Guide to Building a Reusable Visual System - canonical: https://karozieminski.substack.com/p/recraft-ai-visual-identity-consistency-substack - mirror: https://productwithattitude.com/p/recraft-ai-visual-identity-consistency-substack.html - date: 2026-01-06 - word_count: 1670 - access: free - keywords: Recraft, visual identity, AI design, brand consistency - categories: AI Tools, Vibe Coding ### Key concepts - Visual system design - AI design tools - Brand consistency - Reusable design assets ### Extractable claims (8) 1. It takes about 50 milliseconds for readers to form an opinion about a post and whether they'll stay or leave. 2. Substack creators earn a collective $450 million annually. 3. Standing out on Substack is not optional; it's survival. 4. Powerful writing builds trust, while visuals earn the first click. 5. Many writers approach visual identity as an aesthetic problem with an aesthetic solution. 6. Visual identity is about looking recognizable every time, not just looking good once. 7. Consistency in visual identity remains a bottleneck despite generative AI. 8. Using multiple tools can lead to a visual presence that looks like a collage. ### Key quotes > Recraft 101: A Creator's Guide to Building a Reusable Visual System > Using AI design tools to create consistent, reusable brand identities --- ## Choose Your Builder Track > Structured learning paths to help you apply AI with purpose—whether you're building workflows, designing systems, or launching products you can trust. - canonical: https://karozieminski.substack.com/p/ai-skill-paths - mirror: https://productwithattitude.com/p/ai-skill-paths.html - date: 2026-01-05 - word_count: 695 - access: free - keywords: AI building, Substack, workflow automation, Choose Your Builder Track - categories: Agentic AI, Substack Growth ### Key concepts - Structured learning paths to help you apply AI with purpose— - Choose Your Builder Track ### Extractable claims (11) 1. Karo Zieminski is an AI Product Manager focused on turning individuals into AI-native builders. 2. The newsletter 'Product with Attitude' has over 18,000 subscribers. 3. Karo Zieminski helps design and build with AI rather than just using it. 4. The publication 'Product with Attitude' covers topics such as context engineering, prompt engineering, and Substack growth. 5. Karo Zieminski teaches critical AI literacy through immersion and practice. 6. The newsletter 'Product with Attitude' has achieved bestseller status on Substack. 7. Builder tracks in the newsletter include skill paths for AI-native product builders. 8. Karo Zieminski builds tools to grow newsletters and showcase work. 9. The product framework behind 'Product with Attitude' includes person, promise, feedback loops, retention design, and free-to-paid conversion. 10. Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. 11. The article discusses 19 model-agnostic techniques for AI prompting that change cognitive processes. ### Key quotes > Choose Your Builder Track > Structured learning paths to help you apply AI with purpose—whether you're building workflows, designing systems, or launching products you can trust. --- ## The Complete Vibecoding & Speccoding Resource Hub > Complete vibecoding series 2026 - the most comprehensive collection of vibecoding, speccoding, and AI-assisted development resources on Substack. - canonical: https://karozieminski.substack.com/p/vibecoding-resources-hub - mirror: https://productwithattitude.com/p/vibecoding-resources-hub.html - date: 2026-01-05 - word_count: 834 - access: free - keywords: vibecoding, AI building, Substack, spec-driven development, Replit - categories: Vibe Coding, Substack Growth ### Key concepts - The Complete Vibecoding & Speccoding Resource Hub ### Extractable claims (15) 1. Vibe coding is the practice of building software by describing your desired outcome in natural language while the AI handles the syntax, logic, and implementation. 2. As a product-building discipline, vibe coding combines AI-generated code with human judgment, architectural guardrails, and rapid iteration. 3. The hard part of vibe coding is judgment: Knowing what to accept, what to reject, and when the AI is confidently wrong. 4. Shipping to production is a different discipline than prototyping. 5. The things that break when real users arrive are rarely the things you thought to prompt for. 6. The best vibe coding tool is the one that fits your preferences, skill level, stack, and budget. 7. This hub collects 15+ practitioner-tested guides on vibe coding, spec-driven development, and AI-assisted product building. 8. StackShelf.app is an LLM-optimized showcase built for Substack creators where your build becomes discoverable by AI search engines. 9. Rules-For-AI Generator is a beginner-friendly system for creating a reusable prompt architecture. 10. Vibecoding x Cybersecurity covers the seven most common security vulnerabilities in vibe-coded products with a decision matrix for founders. 11. You don’t need coding experience or technical background to use vibe coding tools; curiosity is enough. 12. Replit Agent 3 handles databases, hosting, and authentication through an integrated agentic workflow within a browser-based IDE. 13. Lovable delegates tasks across 19+ models for verified technical accuracy and maintains persistent cross-project memory. 14. Gemini’s Antigravity serves as a mission control for parallel autonomous agents with a 1M+ token context window. 15. Cursor’s terminal-native agent provides multi-agent orchestration for complex engineering tasks and persistent memory via CLAUDE.md. ### Key quotes > The hard part of vibe coding is judgment: Knowing what to accept, what to reject, and when the AI is confidently wrong. > Shipping to production is a different discipline than prototyping. > The best vibe coding tool is the one that fits your preferences, skill level, stack, and budget. --- ## Vibe Coding Prompt Packs Hub > A curated set of prompts used in real product work. - canonical: https://karozieminski.substack.com/p/vibecoding-prompt-packs - mirror: https://productwithattitude.com/p/vibecoding-prompt-packs.html - date: 2026-01-05 - word_count: 431 - access: free - keywords: vibecoding, AI building, spec-driven development, tool review, AI agents - categories: Vibe Coding, Agentic AI ### Key concepts - A curated set of prompts used in real product work - Vibe Coding Prompt Packs Hub ### Extractable claims (7) 1. Karo Zieminski helps people design and build with AI, transforming them into AI-native builders. 2. The Product with Attitude framework includes person, promise, feedback loops, retention design, and free-to-paid conversion. 3. Karo Zieminski has built a community of over 18,000 members focused on developing critical AI literacy. 4. Claude for Small Business features 15 workflows and 8 connectors to streamline AI integration. 5. Karo Zieminski emphasizes the importance of context engineering in AI product management. 6. The Only AI Prompting Guide That Works On Reasoning Models introduces 19 model-agnostic techniques. 7. Karo Zieminski's tools are designed to grow newsletters and showcase individual work effectively. ### Key quotes > Vibe Coding Prompt Packs Hub > A curated set of prompts used in real product work. --- ## Design with AI > Bring AI into your design process without losing creative control, consistency, or quality. - canonical: https://karozieminski.substack.com/p/how-to-design-with-ai - mirror: https://productwithattitude.com/p/how-to-design-with-ai.html - date: 2026-01-05 - word_count: 358 - access: free - keywords: vibecoding, AI building, Recraft, visual design - categories: Vibe Coding ### Key concepts - Design with AI ### Extractable claims (8) 1. The article discusses designing and building with AI. 2. The newsletter covers topics such as prompt engineering and AI tools for product managers. 3. Product with Attitude has tens of thousands of subscribers. 4. Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. 5. The article includes an illustrated guide to context engineering and prompt engineering. 6. Karo Zieminski shares insights on context engineering as an AI PM. 7. The article mentions 19 model-agnostic techniques for AI prompting. 8. The product framework discussed includes person, promise, feedback loops, retention design, and free-to-paid conversion. ### Key quotes > Design with AI > Bring AI into your design process without losing creative control, consistency, or quality. --- ## Build With Attitude: An Open Invitation To Vibe Coders Who Care About Building Things That Matter > AI made building easy. 2026 will decide if we make it meaningful. - canonical: https://karozieminski.substack.com/p/build-with-attitude-vibecoding-quality-over-ai-slop - mirror: https://productwithattitude.com/p/build-with-attitude-vibecoding-quality-over-ai-slop.html - date: 2026-01-05 - word_count: 1005 - access: free - keywords: Build with Attitude, vibecoding, quality over AI slop, manifesto - categories: Vibe Coding, AI Builder Interviews ### Key concepts - Meaningful AI building - Quality over quantity - Vibecoding manifesto - Intentional creation movement ### Extractable claims (6) 1. The Build with Attitude project is a counter-narrative to AI slop and overnight success mythology. 2. AI has made building accessible to everyone, which is a positive development. 3. Accessibility without intentionality produces noise in the building process. 4. The core question for 2026 is whether builders will prioritize quality and sustainability over virality and quick monetization. 5. Craft, judgment, taste, and understanding of user needs remain human responsibilities in the building process. 6. Building things that last requires different decisions than building things that trend. ### Key quotes > Build With Attitude: An Open Invitation To Vibe Coders Who Care About Building Things That Matter > AI made building easy. 2026 will decide if we make it meaningful --- ## 10 Mistakes I Won't Repeat On Substack In 2026 > After ten months on Substack, here are ten mistakes I wish I'd avoided. - canonical: https://karozieminski.substack.com/p/10-substack-mistakes-avoid-2026 - mirror: https://productwithattitude.com/p/10-substack-mistakes-avoid-2026.html - date: 2025-12-27 - word_count: 1083 - access: paid - keywords: Substack mistakes, newsletter strategy, content creator lessons, Substack growth, creator economy - categories: Substack Growth, AI Writing & Content ### Key concepts - Substack growth mistakes - Content strategy lessons - Creator platform pitfalls - Newsletter optimization ### Extractable claims (5) 1. Karo Zieminski states that a single post on Substack takes anywhere between 1 and 10 hours to complete. 2. According to Karo Zieminski, Substack should not be treated as a lunchtime hobby but rather as a small business requiring significant time investment. 3. Karo Zieminski emphasizes the importance of engagement, stating that he was initially unwilling to admit that it mattered to him. 4. Karo Zieminski mentions that he underestimated the time needed for research and editing when starting his Substack newsletter. 5. Karo Zieminski identifies ten mistakes he wishes he had avoided during his first ten months on Substack. ### Key quotes > 10 months on Substack, 10 mistakes worth sharing. > I've talked about what worked. Today, let's cover what very much didn't. --- ## ChatGPT 5.2: What It Really Changed, And Why The Internet's Take Is Mostly Wrong - canonical: https://karozieminski.substack.com/p/openai-chatgpt-52-vs-51-reliability-what-changed-why-everyone-is-wrong - mirror: https://productwithattitude.com/p/openai-chatgpt-52-vs-51-reliability-what-changed-why-everyone-is-wrong.html - date: 2025-12-13 - word_count: 1556 - access: free - keywords: ChatGPT, OpenAI, product analysis, reliability - categories: AI Product Management ### Key concepts - ChatGPT evolution analysis - Product reliability improvements - Nuanced AI critique - Beyond hype narratives ### Extractable claims (6) 1. ChatGPT 5.2 is not designed for viral appeal but for reliable performance. 2. The internet's initial reactions to ChatGPT 5.2 have been largely negative, labeling it as 'boring' and 'cold.' 3. Karo Zieminski conducted a substantial test involving 8,100 lines of complex input to evaluate ChatGPT 5.2. 4. OpenAI made intentional product decisions in ChatGPT 5.2 that prioritize predictable and reliable behavior over flashy capabilities. 5. Evaluating ChatGPT 5.2 based on how it feels is a misguided approach. 6. ChatGPT 5.2 outperforms previous versions when used in practical deployment scenarios. ### Key quotes > ChatGPT 5.2: What It Really Changed, And Why The Internet's Take Is Mostly Wrong > A product analysis that cuts through surface-level reactions to AI updates --- ## 2025's Most Absurd Product Decisions > How Not to Ship AI And Lessons From Companies That Did It Anyway. - canonical: https://karozieminski.substack.com/p/2025s-most-absurd-product-decisions-grok-taco-bell-mcdonalds-checklist-premortem-framework - mirror: https://productwithattitude.com/p/2025s-most-absurd-product-decisions-grok-taco-bell-mcdonalds-checklist-premortem-framework.html - date: 2025-12-10 - word_count: 1818 - access: free - keywords: AI product decisions, product management, AI ethics, product failures, AI shipping - categories: AI Product Management, AI Literacy ### Key concepts - AI product failures - Predictable AI scandals - Product team questions - Ethical AI shipping ### Extractable claims (10) 1. Competent product teams should ask specific questions before shipping. 2. Karo Zieminski is an AI Product Manager and builder. 3. Karo Zieminski has spent years smuggling ethics and product thinking into tech. 4. The article is a post-mortem of the year's worst AI product decisions. 5. The article provides a toolkit to avoid making poor product decisions. 6. Grok's share-chat feature encouraged users to share conversations. 7. The share-chat feature published full conversation transcripts publicly. 8. Roughly 300k+ personal chats were left openly searchable. 9. Sensitive content included emotionally raw medical questions and confessions. 10. The share-chat feature generated a public web page for each conversation. ### Key quotes > Every failure in this post was predictable. Not with hindsight - with the questions any competent product team should ask before shipping. > One year, multiple AI product scandals, zero surprises. --- ## AI Tools A-Z: Deep-Dives for Every Tool I've Tested (Claude, Perplexity, Replit & More) > Real reviews, real workflows. - canonical: https://karozieminski.substack.com/p/ai-tools-reviews-claude-perplexity-replit-suno-deep-dives-2026 - mirror: https://productwithattitude.com/p/ai-tools-reviews-claude-perplexity-replit-suno-deep-dives-2026.html - date: 2025-12-08 - word_count: 1699 - access: free - keywords: AI tools, Claude review, Perplexity, Replit, AI workflows - categories: Vibe Coding, AI Tools ### Key concepts - AI tool reviews - Real workflow examples - Tool comparison library - Honest AI assessments ### Extractable claims (5) 1. Anthropic launched Claude for Small Business with 15 workflows and 8 connectors. 2. The product framework behind Product with Attitude includes person, promise, feedback loops, retention design, and free-to-paid conversion. 3. Karo Zieminski helps individuals design and build with AI, rather than just use it. 4. Karo Zieminski has developed a community of over 18,000 members focused on critical AI literacy through immersion. 5. An Illustrated Guide to Context Engineering and Prompt Engineering was published by Karo Zieminski on May 13. ### Key quotes > A growing library of honest AI tool reviews and workflows by Karo Zieminski on Product with Attitude. > Each entry answers one question builders and creators ask about the tools they use daily. --- ## Visual ASMR: How to Make Oddly Satisfying Designs With Nano Banana > A step-by-step framework to transform images into warm, sensory-rich Christmas visuals. - canonical: https://karozieminski.substack.com/p/visual-asmr-how-to-make-oddly-satisfying-christmas-designs-prompts-nano-banana-hygge-christmas - mirror: https://productwithattitude.com/p/visual-asmr-how-to-make-oddly-satisfying-christmas-designs-prompts-nano-banana-hygge-christmas.html - date: 2025-12-03 - word_count: 1531 - access: free - keywords: visual design, ASMR design, Nano Banana, image transformation, design framework - categories: AI Tools, Vibe Coding ### Key concepts - Visual design framework - Sensory-rich visuals - Nano Banana tool - Image transformation ### Extractable claims (7) 1. Visual ASMR is described as a tiny visual massage for the mind. 2. Designers refer to the quality that makes digital interactions feel physically real as 'tactile'. 3. Neuroscientists call the phenomenon of using visuals to create an internal echo of touch 'vicarious touch processing'. 4. The Danish term for a cozy aesthetic is 'hyggeligt'. 5. Consumers who appreciate hyggeligt aesthetics prefer warm, tactile Christmas cards over glossy, plastic-looking ones. 6. The article provides a step-by-step framework to transform ordinary images into warm, sensory-rich Christmas visuals. 7. Prompts and palettes are included in the framework for creating Christmas visuals. ### Key quotes > A step-by-step framework to transform images into warm, sensory-rich Christmas visuals using Nano Banana. --- ## If You Build With AI, You Need This File. And The System That Generates It. > AI Rules File Generator: A Beginner-Friendly System for Replit, Cursor, Gemini & Claude Builders. - canonical: https://karozieminski.substack.com/p/if-you-build-with-ai-you-need-this - mirror: https://productwithattitude.com/p/if-you-build-with-ai-you-need-this.html - date: 2025-11-27 - word_count: 1538 - access: free - keywords: AI rules files, CLAUDE.md, development configuration, vibecoding tools - categories: Vibe Coding, AI Tools ### Key concepts - AI configuration files - Development setup systems - CLAUDE.md standards - Builder tooling essentials ### Extractable claims (12) 1. Creators experience burnout without effective systems in place. 2. The author emphasizes the importance of quality in their work, stating 'quality or nothing.' 3. The author avoids producing 'AI slop' and focuses on thoroughly researched content. 4. The author's standards for builds include being ethical, solid, anti-cookie-cutter, and maintainable. 5. Time management is a significant challenge for the author, which they address through refined workflows. 6. The author builds reusable systems to manage their workload effectively. 7. An AI rules file is a configuration file that defines the behavior of AI coding agents within a project. 8. Rules files are utilized by tools such as Replit, Cursor, Gemini, and Claude to guide agent behavior. 9. Without a rules file, AI coding agents may generate inconsistent and unmaintainable code. 10. Generating a rules file from a system prompt is a scalable workflow applicable across various projects. 11. Builders who implement systems for their AI workflows can avoid burnout while maintaining quality. 12. The article introduces a beginner-friendly AI rules file generator for new coders. ### Key quotes > If You Build With AI, You Need This File. And The System That Generates It > A beginner-friendly system for AI rules files across Replit, Cursor, Gemini and Claude --- ## What's Your Substack Roadmap for 2026? Mine Comes Straight From 315 Readers. > Understanding Substack User Behavior - canonical: https://karozieminski.substack.com/p/substack-roadmap-community-chat-notes-reels-survey-reader-behaviour - mirror: https://productwithattitude.com/p/substack-roadmap-community-chat-notes-reels-survey-reader-behaviour.html - date: 2025-11-18 - word_count: 1581 - access: free - keywords: Substack roadmap, user behavior, audience feedback, content strategy, reader insights - categories: AI Product Management, Substack Growth ### Key concepts - Reader-driven roadmap - Substack user behavior - Audience feedback strategy - Data-driven planning ### Extractable claims (7) 1. Karo Zieminski emphasizes that not all Substack readers engage with the platform in the same way. 2. The only universal truth about Substack users is that they subscribed and trusted the content. 3. Karo conducted a survey to understand how 315 readers actually use Substack. 4. Understanding reader behavior is crucial for content creators on Substack to deliver value. 5. Karo highlights the importance of knowing what makes readers subscribe to your newsletter. 6. Few writers on Substack share their survey response rates, making it hard to gauge community engagement. 7. Karo believes the real value of any product comes from the people who choose to use it. ### Key quotes > My Substack roadmap for 2026 comes straight from 315 readers - understanding what users actually want drives better content strategy. --- ## How To Product-think When AI Builds At Lightning Speed > The One Feature AI Can't Replace - canonical: https://karozieminski.substack.com/p/product-thinking-at-the-speed-of-ai-actionable-insights-for-product-managers-builders-founders-vibecoders - mirror: https://productwithattitude.com/p/product-thinking-at-the-speed-of-ai-actionable-insights-for-product-managers-builders-founders-vibecoders.html - date: 2025-11-12 - word_count: 1307 - access: free - keywords: product thinking, AI development, product management, building with AI, product strategy - categories: AI Tools, AI Product Management ### Key concepts - Product thinking fundamentals - AI building speed - Irreplaceable product skills - Deliberate decision-making ### Extractable claims (9) 1. Product thinking means solving the right problem, not just building a feature. 2. AI can accelerate building but cannot replace judgment about what to build. 3. The most common trap in product development is building what's technically interesting instead of what users actually need. 4. Shipping fast with AI requires even more deliberate product thinking, not less. 5. Product thinking is a mindset, not a methodology. 6. Founders who built successful products provide the best insights into product thinking. 7. Effective product thinking involves questioning one's own logic and listening to user feedback. 8. Karo Zieminski is an AI Product Manager and creator of tools like LinkSwap, Vault, and StackShelf. 9. Orel is a software engineer and founder of WriteStack, exemplifying the importance of product thinking in building. ### Key quotes > Ask ten people what product thinking means, and you'll get twelve definitions. > The best insight comes from founders who built things that survived contact with users. --- ## I Want To Hear From You, And I'll Pay You For It > A Small Thank-You for Helping Me Build Something Bigger. - canonical: https://karozieminski.substack.com/p/i-want-to-hear-from-you-and-ill-pay - mirror: https://productwithattitude.com/p/i-want-to-hear-from-you-and-ill-pay.html - date: 2025-11-09 - word_count: 228 - access: free - keywords: feedback survey, reader engagement, community building, incentive alignment, Substack growth - categories: Substack Growth ### Key concepts - Honest feedback value - Aligned incentives experiment - Reader survey compensation - Community engagement ### Extractable claims (7) 1. The newsletter 'Product with Attitude' was started in February 2025. 2. As of November 2025, the newsletter 'Product with Attitude' had been running for approximately 9 months. 3. Karo Zieminski emphasizes the importance of honest feedback in content creation. 4. The author is running a survey to gather feedback from subscribers. 5. Participants in the survey have the chance to win prizes, including four $25 Amazon gift cards. 6. The survey also offers digital coffee credits and free-to-premium subscription upgrades as prizes. 7. The author's approach reflects the belief that reader time has genuine value and should be compensated. ### Key quotes > There are few things as underrated as honest feedback. > Asking for your feedback is asking for your time, and I don't take that lightly. --- ## Vibecoding x Cybersecurity: Survival Guide by the Expert Who Fixes Your Code After You > Don't ship your next feature without these safety checks. - canonical: https://karozieminski.substack.com/p/vibecoding-cybersecurity-fix-the-7-code-mistakes-leaving-your-startup-wide-open-and-vulnerable-playbook - mirror: https://productwithattitude.com/p/vibecoding-cybersecurity-fix-the-7-code-mistakes-leaving-your-startup-wide-open-and-vulnerable-playbook.html - date: 2025-11-03 - word_count: 3033 - access: free - keywords: vibecoding, cybersecurity, AI coding, secure development, engineering practices - categories: Vibe Coding, AI Literacy ### Key concepts - Responsible vibecoding - Cybersecurity in AI - Engineering foundations - Safe AI development ### Extractable claims (6) 1. Karo Zieminski emphasizes the importance of shipping features with proper safety checks. 2. According to Karo Zieminski, there are seven code mistakes that can leave startups vulnerable. 3. Karo Zieminski argues that the solution to cybersecurity critiques of vibecoding is to 'vibecode responsibly, with solid engineering foundations.' 4. Farida Khalaf, a Data Engineer specializing in cybersecurity, co-authored the guide on coding with AI. 5. Karo Zieminski states that the cybersecurity critiques of vibecoding are valid and should be addressed. 6. Karo Zieminski is the creator of StackShelf.app and an AI product manager. ### Key quotes > The cybersecurity critics of vibecoding are right about the risks. They're just wrong about the solution. > The answer isn't 'don't vibecode at all.' It's 'vibecode responsibly, with solid engineering foundations.' --- ## The Indie Builder Economy on Substack: Creators Becoming Product Companies > The Most Comprehensive Analysis of Products Built Within the Substack Ecosystem - canonical: https://karozieminski.substack.com/p/substack-indie-builder-economy-2025-product-hunt-creator-earnings-building-in-public-builder-community - mirror: https://productwithattitude.com/p/substack-indie-builder-economy-2025-product-hunt-creator-earnings-building-in-public-builder-community.html - date: 2025-10-28 - word_count: 2273 - access: paid - keywords: indie builders, Substack economy, creator products, platform monetization, product companies - categories: Substack Growth, AI Builder Interviews ### Key concepts - Indie builder economy - Substack product ecosystem - Creator monetization models - Platform-based building ### Extractable claims (8) 1. The Substack economy typically focuses on subscriptions, paywalls, and affiliate revenue. 2. There is a growing market for products built on or for Substack that is often overlooked. 3. Substack fuels the indie builder economy by enabling creators to design, develop, and launch products independently. 4. An indie builder is defined as a creator who operates without company backing or investment. 5. The article provides a comprehensive analysis of products built within the Substack ecosystem. 6. Building on Substack allows creators to monetize their work beyond traditional newsletter models. 7. The indie builder community on Substack is thriving and contributing to the overall economy. 8. Creators on Substack are increasingly transitioning into product companies. ### Key quotes > Substack economy coverage usually centers on subscriptions, paywalls, and affiliate revenue - the 'classic trio' of newsletter monetization. Here's what's overlooked: the growing market of products built on or for Substack. > Let's talk about how Substack fuels the indie builder economy. --- ## Claude Skills Are Taking the AI Community by Storm > Anthropic's Viral Claude Skills Toolkit - canonical: https://karozieminski.substack.com/p/claude-skills-anthropic-viral-toolkit-agentic-workflows-community-guide - mirror: https://productwithattitude.com/p/claude-skills-anthropic-viral-toolkit-agentic-workflows-community-guide.html - date: 2025-10-19 - word_count: 1619 - access: paid - keywords: Claude Skills, Anthropic, AI agents, Claude toolkit, AI capabilities - categories: Agentic AI, AI Builder Interviews ### Key concepts - Claude Skills toolkit - AI agent capabilities - Skill stacking system - Viral AI feature ### Extractable claims (3) 1. Claude Skills are generating significant excitement in the AI community, indicating a fundamental shift in AI capabilities. 2. Simon Willison described Claude Skills as 'maybe a bigger deal than MCP.' 3. Ethan Mollick stated that Claude Skills provide 'both an easy path for workable agents and a step forward in what AI can do.' ### Key quotes > The AI world hasn't slept in 72 hours, and I get it. > Simon Willison called it 'maybe a bigger deal than MCP.' Ethan Mollick said it's 'both an easy path for workable agents and a step forward in what AI can do.' --- ## The Ultimate Vibecoding Guide From Builders Who've Shipped > Go Beyond the Prototype - canonical: https://karozieminski.substack.com/p/vibecoding-tips-the-ultimate-collection - mirror: https://productwithattitude.com/p/vibecoding-tips-the-ultimate-collection.html - date: 2025-10-11 - word_count: 2881 - access: free - keywords: vibecoding, AI coding tools, product development, build in public, app deployment - categories: Vibe Coding, AI Tools ### Key concepts - AI coding tools - Production-ready apps - Practical development roadmap - Validate and deploy - Build in public ### Extractable claims (7) 1. The roadmap includes validating your idea early and designing user flows first. 2. Karo Zieminski is an AI product manager and builder of StackShelf.app. 3. The guide is based on real projects that have been shipped. 4. Karen Spinner is the founder of CarouselBot and a fellow builder. 5. The guide emphasizes the importance of using Git from day one. 6. It advises to debug systematically and keep learning fundamentals. 7. The article was published on October 11, 2025. ### Key quotes > Build production-ready apps with AI coding tools by following this practical roadmap: validate your idea early, design user flows first, choose the right stack, write clear prompts, use Git from day one, debug systematically, and keep learning fundamentals. > We share a playbook that's been tested, improved, and proven with real results. --- ## 10x Your Productivity with Perplexity Comet: 11 Use Cases from 'Nice' to 'Wow!' > Unlock advanced productivity with Perplexity Comet. - canonical: https://karozieminski.substack.com/p/perplexity-comet-advanced-use-cases-guide - mirror: https://productwithattitude.com/p/perplexity-comet-advanced-use-cases-guide.html - date: 2025-10-06 - word_count: 3496 - access: paid - keywords: Perplexity Comet, productivity, AI automation, workflows, digital assistant - categories: Agentic AI, AI Writing & Content ### Key concepts - Perplexity Comet workflows - Digital coworker automation - Advanced productivity hacks - Multi-agent workflows - Content repurposing ### Extractable claims (8) 1. Perplexity Comet is a digital coworker that thinks, navigates, and acts on your behalf across the web. 2. The guide includes 11 ranked workflows, shortcut patterns, and an interactive research dashboard setup. 3. Perplexity Comet can turn your prompt library into an actions library. 4. Users can do grocery shopping with one click using Perplexity Comet. 5. Perplexity Comet allows users to set up their ultimate focus playlist. 6. AI automation through Perplexity Comet can help cut electricity bills. 7. Research projects can be turned into interactive dashboards with Perplexity Comet. 8. The article provides prompts, case studies, and practical examples for using Perplexity Comet. ### Key quotes > Comet isn't just a browser, it's a digital coworker that thinks, navigates, and acts on your behalf across the web. > 11 ranked workflows, shortcut patterns, and power moves that turn Comet into your personal assistant. --- ## Why Every Brand and PM Should Monitor Substack for User Insights > If your product or brand team isn't monitoring Substack, you're missing one of the richest streams of user intelligence. - canonical: https://karozieminski.substack.com/p/why-every-brand-and-pm-should-monitor - mirror: https://productwithattitude.com/p/why-every-brand-and-pm-should-monitor.html - date: 2025-09-30 - word_count: 1384 - access: free - keywords: Substack, user research, market intelligence, product management, user insights - categories: AI Product Management, Substack Growth ### Key concepts - Substack market research - User intelligence mining - Long-form user insights - Alternative to surveys - Product team intelligence ### Extractable claims (11) 1. Substack is one of the richest streams of user intelligence available today. 2. Monitoring Substack is essential for product and brand teams seeking real user insights. 3. Substack serves as a market research tool where users, builders, and domain experts share in-depth narratives. 4. Every post on Substack functions as a long-form research report, providing valuable insights. 5. Writers on Substack share walkthroughs of real workflows and experiments that demonstrate product value. 6. Workarounds discussed on Substack represent feature requests in disguise, indicating necessary product improvements. 7. Emerging patterns in user behavior can be identified through multiple writers on Substack. 8. Candid failures shared on Substack reveal product friction that traditional surveys may overlook. 9. Comments on Substack act as focus groups, providing contextual feedback from engaged users. 10. The comment section on Substack is more signal-dense than most social platforms due to audience self-selection. 11. Product managers who monitor Substack can gain insights about real workflows, workarounds, and failure modes. ### Key quotes > If your product or brand team isn't monitoring Substack, you're missing one of the richest streams of user intelligence available today. > Substack is a criminally underrated market research tool. It's where users, builders, and domain experts go deep. --- ## Vibecoding, Spec-driven, And The New AI Development Lexicon: A Definitive Guide > Let's define these terms once and for all. - canonical: https://karozieminski.substack.com/p/vibecoding-spec-driven-and-the-new - mirror: https://productwithattitude.com/p/vibecoding-spec-driven-and-the-new.html - date: 2025-09-22 - word_count: 2199 - access: free - keywords: vibecoding, AI development, spec-driven, product building, coding terminology - categories: Vibe Coding ### Key concepts - AI development terminology - Vibecoding definition - Spec-driven development - Context engineering - AI coding lexicon ### Extractable claims (7) 1. Vibecoding and Spec-driven development are key terms in AI coding that often get mixed up. 2. Karo Zieminski emphasizes the importance of using the right terminology in AI product development. 3. The article serves as a definitive guide to AI coding terminology for modern product builders. 4. Context engineering and AI-assisted engineering are also part of the evolving AI development lexicon. 5. Karo Zieminski has spent a year testing these terms in real-life projects to clarify their meanings. 6. Understanding the differences between AI coding methods is crucial for effective product shipping. 7. The article aims to clear up confusion surrounding AI coding buzzwords and their practical applications. ### Key quotes > Let's define these terms once and for all. > The definitive guide to the AI coding terminology, what actually works in practice, and why the right words matter if you want your product to ship. --- ## Vibecoding, But Smarter: The PRD Prompt That Audits Itself For Hallucinations and Failure Modes > Best Vibecoding Guides #1 - canonical: https://karozieminski.substack.com/p/vibecoding-but-smarter-the-prd-prompt - mirror: https://productwithattitude.com/p/vibecoding-but-smarter-the-prd-prompt.html - date: 2025-09-17 - word_count: 1970 - access: free - keywords: vibecoding, PRD prompt, AI hallucinations, product requirements, prompt engineering - categories: Vibe Coding ### Key concepts - Self-auditing PRD prompt - Hallucination prevention - Failure mode detection - Smart vibecoding - Pre-coding validation ### Extractable claims (7) 1. A Product Requirements Document (PRD) is essential before starting any coding. 2. The PRD prompt can audit itself for hallucinations and failure modes. 3. Skipping the PRD, rules for AI, and system prompts can lead to wasted credits and lost time. 4. Creating a PRD is one of the three essential steps before vibecoding. 5. Setting rules for AI is one of the three essential steps before vibecoding. 6. Writing system prompts is one of the three essential steps before vibecoding. 7. Fifteen minutes spent on the PRD prompt can significantly impact the quality of the final product. ### Key quotes > The One Prompt You Need Before Touching Code > The PRD Prompt That Audits Itself For Hallucinations and Failure Modes --- ## Where To Find Emerging Product Voices on Substack in 2025 > A living map of product minds building, writing, and sharing in community. - canonical: https://karozieminski.substack.com/p/where-to-find-emerging-product-voices - mirror: https://productwithattitude.com/p/where-to-find-emerging-product-voices.html - date: 2025-08-31 - word_count: 1293 - access: free - keywords: Substack, product management, community, product voices, networking - categories: AI Product Management, Substack Growth ### Key concepts - Product voices directory - Substack community mapping - Emerging product writers - Product management network - Living resource guide ### Extractable claims (6) 1. Substack's search algorithm prioritizes results based on page rank, making it difficult for new product writers to be discovered. 2. The author created community directories, such as Product People of Substack and AI Tools - Community Deep Dives, to address the discovery problem for emerging voices. 3. These community directories are manually curated and maintained rather than being algorithmically generated. 4. As of September 19, 2025, searching for 'AI' on Substack yields an overwhelming number of results, necessitating curation to find genuine emerging voices. 5. The product community on Substack includes writers from various fields such as product management, AI tools, and building in public. 6. The author's approach to community building focuses on creating infrastructure, like directories and collaborative posts, rather than just producing content. ### Key quotes > A living map of product minds building, writing, and sharing in community. > You join Substack to find your people. --- ## Behind the Scenes: Why PMs Are Vibecoding Real Products > 5 Non-Obvious Reasons Product Managers Should Start Vibecoding Today. - canonical: https://karozieminski.substack.com/p/vibecoding-pms-a-movement-you-dont - mirror: https://productwithattitude.com/p/vibecoding-pms-a-movement-you-dont.html - date: 2025-08-21 - word_count: 1093 - access: free - keywords: product managers, vibecoding, PM skills, build in public, product development - categories: Vibe Coding, AI Tools ### Key concepts - PMs vibecoding products - Non-obvious PM benefits - Product manager coding - Practical vibecoding reasons - PM skill evolution ### Extractable claims (5) 1. Karo Zieminski asserts that vibecoding allows product managers to ship real, product-grade apps. 2. According to Karo Zieminski, vibecoding involves using natural language prompts to describe what you want to build. 3. Karo Zieminski emphasizes that product managers do not need to know how to code to engage in vibecoding. 4. Karo Zieminski states that vibecoding provides visibility into code, databases, and logs that product managers typically do not see. 5. Karo Zieminski argues that being specific about system requirements is a crucial skill for product managers. ### Key quotes > 5 Non-Obvious Reasons Product Managers Should Start Vibecoding Today. > Vibecoding isn't just a joke, easy to dismiss. --- ## I Analyzed Every Interaction From My First 6 Months on Substack > Special edition - canonical: https://karozieminski.substack.com/p/i-analyzed-every-interaction-from - mirror: https://productwithattitude.com/p/i-analyzed-every-interaction-from.html - date: 2025-08-15 - word_count: 1207 - access: paid - keywords: Substack growth, content strategy, community building, analytics, audience growth - categories: Substack Growth, AI Writing & Content ### Key concepts - Substack growth analysis - Six-month interaction data - Rapid growth drivers - Community building metrics - Content performance insights ### Extractable claims (5) 1. Karo Zieminski started Product with Attitude on Substack with 0 followers. 2. In the first 6 months, Product with Attitude grew to over 2.1K subscribers. 3. The publication ranked #63 in the Rising in Technology category on Substack. 4. Over the past 7 weeks, the publication was consistently featured in Rising in Technology, peaking at #19. 5. Karo Zieminski utilized a Python script to analyze the impact of interactions on Substack. ### Key quotes > I Analyzed Every Interaction From My First 6 Months on Substack - Here's What Drove My Rapid Growth > I started at 0 followers - that was a blessing in disguise. --- ## July's Top Vibecoding Reads > No code, no funding, no problem - canonical: https://karozieminski.substack.com/p/julys-top-vibecoding-reads - mirror: https://productwithattitude.com/p/julys-top-vibecoding-reads.html - date: 2025-08-05 - word_count: 734 - access: free - keywords: vibecoding, no-code, AI tools, Replit, creative development - categories: Vibe Coding, AI Tools ### Key concepts - Vibecoding resources roundup - No-code movement - Creativity over credentials - AI development tools - Democratized software building ### Extractable claims (3) 1. Tools like Replit, Cursor, and Lovable are enabling creativity to be transformed into software without the need for coding skills. 2. Curiosity and creativity are more valuable than technical skills or funding in the current landscape of building products. 3. Many successful projects are being built by individuals with zero coding skills. ### Key quotes > No code, no funding, no problem: curiosity now builds more than credentials ever could > Tools like Replit, Cursor, and Lovable are turning creativity into software - fast. --- ## The Self-Improving Prompt System That Gets Smarter With Every Use > Instantly build, score, and improve your AI prompts. - canonical: https://karozieminski.substack.com/p/the-self-improving-prompt-system - mirror: https://productwithattitude.com/p/the-self-improving-prompt-system.html - date: 2025-08-01 - word_count: 2523 - access: free - keywords: prompt engineering, AI prompts, prompt optimization, prompt system, AI accuracy - categories: AI Tools, Agentic AI ### Key concepts - Self-improving prompt system - Prompt scoring framework - Two-step prompt loop - AI prompt optimization - Automated prompt evaluation ### Extractable claims (6) 1. The Self-Improving Prompt System scores prompts across 35 criteria. 2. Karo Zieminski's framework helped achieve higher task accuracy and fewer rewrites. 3. The prompt building system consists of a simple two-step loop: Prompt Builder and Prompt Evaluator. 4. Users do not need technical skills to utilize the prompt system effectively. 5. The system is designed to unlock ChatGPT's potential for better, more consistent results. 6. Karo Zieminski shares insights from the world of AI product management with thousands of readers globally. ### Key quotes > The Self-Improving Prompt System That Gets Smarter With Every Use > A simple two-step loop (Prompt Builder → Prompt Evaluator) that scores your prompt across 35 criteria and suggests how to refactor it. --- ## Is Your Replit Looping? This Will Help. > The 3 Infuriating Ways Replit Will Break Your Code - canonical: https://karozieminski.substack.com/p/is-your-replit-looping-this-will - mirror: https://productwithattitude.com/p/is-your-replit-looping-this-will.html - date: 2025-07-21 - word_count: 2124 - access: paid - keywords: Replit, AI product management, coding errors, building in public, development tools - categories: Vibe Coding, AI Tools ### Key concepts - Replit coding challenges - AI-assisted development barriers - Code breaking patterns - Building in public ### Extractable claims (6) 1. Replit is a powerful tool that eliminates barriers between ideas and reality. 2. Karo Zieminski is an AI Product Manager who shares insights from the world of AI product management. 3. Replit can sometimes cause issues that may frustrate users. 4. The author faced three specific problems while using Replit that nearly made them quit. 5. The author is building a platform called StackShelf for Substack creators. 6. The author used Replit, Figma, ChatGPT, and persistence to work on their project. ### Key quotes > There's no longer any barrier between my ideas and reality, and that's a powerful, powerful feeling. > Replit is amazing when it works. But when it doesn't, you'll... --- ## Vibecoding Journey: I Broke Replit So You Don't Have To > Building in Public: The official patch notes, episode 1. - canonical: https://karozieminski.substack.com/p/i-broke-replit-so-you-dont-have-to - mirror: https://productwithattitude.com/p/i-broke-replit-so-you-dont-have-to.html - date: 2025-07-14 - word_count: 1517 - access: free - keywords: vibecoding, Replit, building in public, product updates, StackShelf - categories: Vibe Coding ### Key concepts - Vibecoding journey - Building in public - Product update transparency - Development patch notes ### Extractable claims (10) 1. There are two types of product updates: those that are proudly listed in changelogs and those that are buried under 'general improvements'. 2. Accidentally deleting the auth flow is an example of a product update that is often not publicly acknowledged. 3. StackShelf was built to provide Substack creators with a native, integrated way to sell digital products to their existing audience. 4. Most Substack creators want to monetize their expertise despite the platform being designed primarily for writing and audience building. 5. Building with AI coding tools like Replit involves real failures, bugs, and unexpected side effects. 6. The vibecoding approach to app development emphasizes persistence through technical failures. 7. Authentic building-in-public content should document failures and mistakes, not just successes. 8. Substack's platform is optimized for audience building and writing, rather than direct product sales. 9. The development process of StackShelf included sleep-deprived efforts to create a functional platform. 10. The article discusses the behind-the-scenes process of building StackShelf, including bugs and errors. ### Key quotes > There Are Two Types of Product Updates: The ones you proudly list in changelogs and marketing decks. The ones you bury deep under 'general improvements'. > I Broke Replit So You Don't Have To --- ## Founders, Here's What You Need to Know About Product Launches in 2025 > OpenAI, Google, Anthropic & Perplexity's launch strategy - canonical: https://karozieminski.substack.com/p/founders-heres-what-you-need-to-know - mirror: https://productwithattitude.com/p/founders-heres-what-you-need-to-know.html - date: 2025-06-12 - word_count: 1108 - access: paid - keywords: product launches, AI strategy, OpenAI, launch strategy, founders - categories: AI Product Management ### Key concepts - Product launch strategy - AI company launches - Launch acceleration trend - 2025 product strategy ### Extractable claims (7) 1. Founders must understand the 267% acceleration trend shaping product launches in 2025. 2. OpenAI, Google, Anthropic, and Perplexity are redefining launch strategies for the upcoming year. 3. The competitive landscape in 2025 will require founders to adapt quickly to new release strategies. 4. Product launches in 2025 will be influenced by the rapid advancements in AI technologies. 5. Understanding the release strategies of major players is crucial for any founder planning a launch. 6. Karo Zieminski emphasizes the importance of staying ahead of trends in product development. 7. Founders should anticipate significant shifts in consumer expectations as AI evolves. ### Key quotes > Founders, Here's What You Need to Know About Product Launches in 2025 > OpenAI, Google, Anthropic & Perplexity's launch strategy - and what it means for yours --- ## I Studied 49 AI-Generated Bios. What I Found Wasn't About AI at All > How an AI Trend Is Changing What We Measure Ourselves Against - canonical: https://karozieminski.substack.com/p/i-studied-49-ai-generated-bios-what - mirror: https://productwithattitude.com/p/i-studied-49-ai-generated-bios-what.html - date: 2025-06-04 - word_count: 573 - access: free - keywords: AI bios, LinkedIn, personal branding, ChatGPT, self-perception - categories: AI Literacy, AI Writing & Content ### Key concepts - AI-generated bios - LinkedIn personal branding - Self-perception through AI - Digital identity trends ### Extractable claims (9) 1. The study categorized the bios into three distinct groups. 2. The first group of bios reflects curiosity about how AI works. 3. The second group of bios presents the author through AI's lens. 4. The third group of bios focuses solely on personal branding with AI as the ghostwriter. 5. The trend of using AI to generate bios is prevalent on LinkedIn. 6. The formula for creating an AI-generated bio involves prompting ChatGPT with 'Describe me based on our chats.' 7. Professionals are using AI-generated bios to build their personal brands. 8. The finesse and self-awareness of AI-generated bios vary among users. 9. The trend highlights varying levels of AI literacy among professionals. ### Key quotes > I Studied 49 AI-Generated Bios. What I Found Wasn't About AI at All > How an AI Trend Is Changing What We Measure Ourselves Against --- ## How I Ended Up Building a Peer-Peer Marketplace For Substack > Substack Writers Are Building Incredible Things. - canonical: https://karozieminski.substack.com/p/how-i-ended-up-building-a-digital - mirror: https://productwithattitude.com/p/how-i-ended-up-building-a-digital.html - date: 2025-05-25 - word_count: 618 - access: free - keywords: Substack marketplace, creator economy, content discovery, peer-to-peer, Substack writers - categories: Substack Growth, AI Writing & Content ### Key concepts - Peer-to-peer marketplace - Substack creator economy - Content discovery problem - Creator monetization ### Extractable claims (7) 1. Karo Zieminski is building a peer-to-peer marketplace called StackShelf.app for Substack creators to sell digital products. 2. Substack writers create diverse digital products, including books, tools, templates, courses, and guides. 3. The discoverability problem affects both readers and creators on Substack, as readers struggle to find products and creators struggle to reach buyers. 4. Existing large marketplaces treat creators as conversion metrics, which can undermine the creator-reader relationship. 5. StackShelf.app aims to preserve the creator-reader relationship by leveraging Substack's existing trust infrastructure. 6. WriteStack, built by Orel Zilberman, is an example of a Substack-native tool that emerged from community needs. 7. Karo's directories, Product People of Substack and AI Tools - Community Deep Dives, have become valuable discovery resources for creators. ### Key quotes > Thirty minutes and four platforms later, I found myself shopping for soy candles on Etsy. Still no template. > Substack Writers Are Building Incredible Things. Let's Make It Easier for Everyone to Find Them. --- ## How To Implement Hyper-personalization Without Creeping People Out > A Practical Guide to AI-Driven Hyper-personalization - canonical: https://karozieminski.substack.com/p/how-to-implement-hyper-personalization - mirror: https://productwithattitude.com/p/how-to-implement-hyper-personalization.html - date: 2025-05-14 - word_count: 1576 - access: paid - keywords: hyper-personalization, AI personalization, user privacy, product strategy, data ethics - categories: AI Product Management ### Key concepts - Hyper-personalization strategy - Privacy versus personalization - AI-driven customization - User data ethics ### Extractable claims (5) 1. According to Karo Zieminski, 71% of people crave personalized experiences. 2. Karo Zieminski states that 76% of consumers get irritated when brands fail to provide personalized experiences. 3. Karo Zieminski argues that product teams face a dilemma between collecting data and risking backlash or not collecting data and facing user churn. 4. Karo Zieminski suggests that betting on ethics before making money can be seen as too expensive, yet it maintains moral high ground. 5. Karo Zieminski reflects on the steep learning curve faced by product teams when navigating privacy concerns. ### Key quotes > We say we hate it when apps track us. But we also get annoyed when Netflix forgets we don't enjoy documentaries about feet. > 71% of us crave personalized experiences, and 76% get irritated when brands fall short. --- ## Are You a Product Person Lost on Substack? Here's the Map > Where To Find Emerging Product Voices in 2025 - canonical: https://karozieminski.substack.com/p/are-you-a-product-person-lost-on - mirror: https://productwithattitude.com/p/are-you-a-product-person-lost-on.html - date: 2025-05-06 - word_count: 862 - access: free - keywords: Substack directory, product management, content curation, emerging writers, product voices - categories: AI Product Management, Substack Growth ### Key concepts - Substack product writers - Emerging creator voices - Curated content directory - Product management community ### Extractable claims (7) 1. Product People on Substack is a curated directory of active, emerging product writers. 2. New creators are added to the directory as they are discovered and reviewed quarterly. 3. The list of product writers is intentionally not ranked or sorted to give all voices a fair chance. 4. Subscriber count is not always a reliable proxy for the quality of content. 5. The author created the directory to solve the problem of finding fresh product content on Substack. 6. The author plans to send out a fresh edition of the directory every quarter. 7. Many newsletters found on Substack are either well-known or inactive, leading to a lack of new voices. ### Key quotes > You open Substack. You type 'product manager' into the search bar. And this is what you find: World-famous newsletters you already know about. > Product People on Substack is a curated directory of active, emerging product writers. --- ## Substack's Roadmap And Why 'No, Substack, We Don't Want Reels' Went Viral > Compiled by a Product Manager for the Substack Team - canonical: https://karozieminski.substack.com/p/substacks-product-roadmap-why-no - mirror: https://productwithattitude.com/p/substacks-product-roadmap-why-no.html - date: 2025-04-22 - word_count: 1141 - access: free - keywords: Substack roadmap, product feedback, platform features, viral content, user requests - categories: AI Product Management, Substack Growth ### Key concepts - Substack product roadmap - User feedback analysis - Platform feature requests - Community-driven development ### Extractable claims (4) 1. Substack's roadmap includes a focus on community needs. 2. 'We Don't Want Reels' went viral, receiving 20.1K likes in 3 days. 3. Substack is described as the Internet's most emotionally intelligent platform. 4. The platform fosters interactions between creators and subscribers. ### Key quotes > No, Substack, We Don't Want Reels > 20.1K likes in 3 days. Since I'm a Product Manager - and... --- ## 19 Product Team Lessons in 4 Minutes > Not best practices. Just what actually works in product teams. - canonical: https://karozieminski.substack.com/p/20-product-team-lessons-in-4-minutes - mirror: https://productwithattitude.com/p/20-product-team-lessons-in-4-minutes.html - date: 2025-04-13 - word_count: 393 - access: paid - keywords: product management, PM lessons, product teams, practical advice, team collaboration - categories: AI Product Management ### Key concepts - Product management lessons - Practical PM advice - Team collaboration tactics - Operational product wisdom ### Extractable claims (2) 1. Design like people are tired and distracted, because they are. 2. Build with opinion; you can't please your way to greatness. ### Key quotes > Not best practices. Just what actually works in product teams. > This list is not aspirational, it doesn't ask you to 'evangelize the vision' - it's operational. --- ## User Personas are Dead: AI-Powered User Models for 2026 and Beyond > Why Static Personas Are No Longer Relevant - canonical: https://karozieminski.substack.com/p/user-personas-are-dead-ai-powered - mirror: https://productwithattitude.com/p/user-personas-are-dead-ai-powered.html - date: 2025-04-06 - word_count: 1180 - access: paid - keywords: user personas, AI user models, product research, user behavior, dynamic personas - categories: AI Product Management, AI Literacy ### Key concepts - Dynamic user models - AI-powered personas - Static personas obsolescence - Real-time user behavior ### Extractable claims (9) 1. Static user personas are no longer relevant for product teams. 2. User personas were originally introduced by Alan Cooper in the 1990s. 3. Static personas are built to humanize users but fail to reflect real user behavior. 4. AI technology can capture real user behavior more accurately than static personas. 5. Dynamic user models are proposed as a replacement for static personas. 6. Massive behavioral datasets and real-time user segmentation have rendered static personas obsolete. 7. AI-powered pattern recognition allows for better understanding of user behavior. 8. Dynamic feedback loops contribute to the effectiveness of adaptive user models. 9. Empathy and field research remain timeless despite the evolution of user modeling. ### Key quotes > User Personas are Dead: AI-Powered User Models for 2026 and Beyond > These static personas were built to humanize users, but with AI capturing real user behavior, they're about as relatable as commercial images on my dentist's brochure. --- ## The Unclickable Word That Can Save Your Career > AI Ethics: Why It Matters for PMs. - canonical: https://karozieminski.substack.com/p/help-me-name-this-post-theres-a-surprise - mirror: https://productwithattitude.com/p/help-me-name-this-post-theres-a-surprise.html - date: 2025-03-30 - word_count: 850 - access: paid - keywords: AI ethics, product management, career development, ethical AI, product managers - categories: AI Product Management, AI Literacy ### Key concepts - AI ethics for PMs - Product manager responsibility - Ethical product development - Career-saving principles ### Extractable claims (11) 1. The word 'ethics' has low click-through rates in digital content. 2. AI ethics is a critical but undervalued skill for product managers' career longevity. 3. The challenge of content creators covering ethics is that the topic has poor engagement metrics despite its importance. 4. Community collaboration on content creation can generate unexpected engagement and insights. 5. Product managers who ignore AI ethics risk career vulnerability as AI systems become more consequential. 6. The initial title request for this post led to a lively conversation among Substack creators. 7. The author struggles with finding the perfect title for their posts. 8. The response to the title request exceeded the author's expectations. 9. The topic of ethics is often perceived as boring and uninviting. 10. Content that focuses on 'disruption' or 'innovation' tends to attract more engagement than ethical discussions. 11. The author previously addressed the challenges of titling in their writing. ### Key quotes > The Unclickable Word That Can Save Your Career > AI Ethics: Why It Matters for PMs --- ## When Fun Turns Predatory: Inside Temu's AI-Driven UX > Dark Patterns in Product, Ethics For Sale - canonical: https://karozieminski.substack.com/p/when-fun-turns-predatory-inside-temus - mirror: https://productwithattitude.com/p/when-fun-turns-predatory-inside-temus.html - date: 2025-03-23 - word_count: 677 - access: paid - keywords: dark patterns, UX design, Temu, addictive design, product ethics - categories: AI Product Management, AI Literacy ### Key concepts - Dark UX patterns - Dopamine-driven design - Predatory user experience - Gamified manipulation - Psychological warfare design ### Extractable claims (7) 1. Temu's user experience is a psychological warfare strategy disguised as a fun app. 2. The app employs every dark pattern imaginable to manipulate users into spending more time and money. 3. Gamified interfaces and countdown timers create a sense of urgency that pressures users into making impulsive decisions. 4. Temu introduces unexpected elements, like a virtual fish, to distract users from their original shopping goals. 5. The design encourages nostalgia by reminding users of childhood toys like Tamagotchis, making them emotionally invested. 6. Temu's approach to shopping is designed to be addictive, turning a simple task into a complex game. 7. The app's insistence on engaging users with emotionally charged nudges complicates the shopping experience. ### Key quotes > Temu's UX isn't just bad. It's psychological warfare wrapped in confetti. > They've embraced every dark pattern in the book, all of them, shoved them into one app, and then said: You know what? Let's invent more. --- ## 6 Substack Lessons From a Product Manager With Zero Followers > What I learned in my first month writing online with no audience. - canonical: https://karozieminski.substack.com/p/6-lessons-from-starting-on-substack - mirror: https://productwithattitude.com/p/6-lessons-from-starting-on-substack.html - date: 2025-03-16 - word_count: 1120 - access: free - keywords: Substack, content creation, product manager, audience growth, writing online - categories: AI Product Management, Substack Growth ### Key concepts - Substack growth strategies - Zero-follower start - Content creation learning - PM approach to writing - Audience building tactics ### Extractable claims (11) 1. The author started on Substack with imposter syndrome. 2. For the first week, the author did not post anything and instead lurked and learned. 3. In the first week, the author engaged in keyword-searching for topics of interest. 4. The author binge-read everything they could find related to Substack. 5. The author studied profiles similar to theirs to understand successful writing on Substack. 6. The author aimed to understand what makes a Substack post spark authentic conversation. 7. Having zero followers resulted in low stakes for the author. 8. On Day 2 after sharing their first post, a few people found the writing and one responded. 9. On Day 3, someone subscribed to the author's newsletter. 10. By Day 35, the author had over 6,000 LinkedIn followers before starting Substack. 11. The author's newsletter doubled its subscribers within a week of starting. ### Key quotes > I started on Substack with imposter syndrome as my co-pilot. > Classic PM behavior: scan the field, detect patterns and only then consider making a move. --- ## 5 Tips for Product Teams Implementing AI Prioritization Tools > The Death of Stakeholder Drama? - canonical: https://karozieminski.substack.com/p/5-tips-for-product-teams-to-implement - mirror: https://productwithattitude.com/p/5-tips-for-product-teams-to-implement.html - date: 2025-03-08 - word_count: 1312 - access: paid - keywords: AI prioritization, product management, stakeholder management, product teams, decision-making - categories: AI Tools, AI Product Management ### Key concepts - AI prioritization tools - Stakeholder management - Product decision-making - Saying no effectively - Data-driven prioritization ### Extractable claims (6) 1. AI-assisted prioritization tools help product teams make data-driven decisions about what to build next. 2. Saying no to a feature request can trigger a neurological response similar to physical discomfort. 3. The hardest part of product management is saying no to the wrong features while keeping stakeholders motivated. 4. AI prioritization tools reduce the emotional charge of saying no by shifting the conversation from opinion to data. 5. To implement AI prioritization tools, product teams should start with their existing data, including backlog, user research, and business metrics. 6. Stakeholders trust the prioritization process more when they can see how the prioritization was calculated. ### Key quotes > I froze during a stakeholder meeting. > How AI Rewrites the Rules of Saying 'No' --- ## Thoughts, Hacked: 20 Questions We Need to Ask Before Brain-Computer Interfaces Go Mainstream > Brain-Computer Interfaces (BCIs) are advancing from labs to startups. - canonical: https://karozieminski.substack.com/p/the-20-questions-i-have-about-brain - mirror: https://productwithattitude.com/p/the-20-questions-i-have-about-brain.html - date: 2025-02-28 - word_count: 964 - access: paid - keywords: brain-computer interfaces, BCI, neurotechnology, tech ethics, wearable technology - categories: AI Literacy ### Key concepts - Brain-computer interfaces - BCI ethical questions - Wearable brain technology - Neurotechnology ethics - Thought privacy concerns ### Extractable claims (5) 1. Karo Zieminski states that Brain-Computer Interfaces (BCIs) are advancing from labs to startups. 2. Karo Zieminski highlights that the ethical roadmap for BCIs is currently unclear. 3. According to Karo Zieminski, Xu Minpeng predicts that BCIs could evolve into commercial, wearable devices serving as assistive technology. 4. Karo Zieminski emphasizes that BCIs are evolving faster than most people realize. 5. Karo Zieminski suggests that BCIs could be integrated into daily life, similar to how the Apple Watch functions. ### Key quotes > Thoughts, Hacked: 20 Questions We Need to Ask Before Brain-Computer Interfaces Go Mainstream > BCIs could eventually evolve into commercial, wearable devices serving as assistive technology in daily life (think Apple Watch, but integrated with your brain) ---