FAQ

I test AI systems for a living, and this is the definition that holds up in practice. AI literacy is the ability to understand, use, evaluate, and think critically about AI systems. It goes beyond knowing how to prompt ChatGPT. It means understanding why different AI tools make different decisions, where they fail, and what that means for your work.

It matters because the World Economic Forum reports that 39% of core skills will change by 2030. LinkedIn data shows AI literacy is being added to professional profiles 80x faster than two years ago. And 63% of employers say the skills gap is their top barrier to business transformation. AI literacy isn't a nice-to-have anymore. It's becoming a requirement across roles, from product management to education to design.

This is the term I built the whole publication around. Critical AI literacy is the ability to evaluate, build with, and think independently about AI systems. It covers the design philosophies behind different tools, where AI outputs fall short, and when to trust, adjust, or override what AI produces.

Educators and practitioners are shifting how they define AI fluency. Using AI is table stakes. The judgment to use it well is the literacy part.

We practice critical AI literacy through building, testing, and comparing AI tools, not just reading about them.

Using AI means opening a tool, typing a prompt, copying the output, and moving on. That's consumption.

Building with AI means designing workflows, chaining tools, automating repetitive tasks, and shipping things that solve real problems. It means understanding how AI systems work well enough to make them do what you need, not what they default to.

That's where the career value sits right now. Anyone can paste a prompt. Fewer people can design a system that runs reliably, adapts to edge cases, and saves hours every week.

Start with one real problem you have. Not a tutorial project. An actual annoyance in your day. Maybe it's organizing meeting notes. Maybe it's the weekly report you rebuild by hand every Friday.

Then pick one AI tool and try to solve it. You'll fail a few times. That's the point. Each failure teaches you something about how the tool thinks and where it needs guidance.

The structured way to do this: follow a learning path that moves from everyday AI use to building workflows to shipping tools. Product with Attitude offers 5 learning paths designed for exactly this progression, from Everyday AI (no code required) to full Build with AI projects.

You don't need a degree or a bootcamp. Closing the AI skills gap is a practice problem, not a knowledge problem.

What works: build something small every week. Join a community where people share what they're building (so you learn from their experiments, not just your own). Read honest tool reviews instead of hype. Compare AI systems side by side so you develop taste, not just familiarity.

The WEF's 2026 Davos report found that self-directed learning is declining. People learn faster with structured paths and community accountability. That's why community-based learning (like the weekly rhythm at Product with Attitude) tends to stick better than solo courses.

Every tool linked on this site was built this way. Vibe coding is using natural language prompts to build apps, prototypes, and products with AI assistance. Instead of writing traditional code line by line, you describe what you want in plain language and an AI tool generates the code for you.

The term took off in 2025 and became mainstream in 2026. You don't need programming experience to start. People with backgrounds in product management, design, education, writing, and coaching are vibe coding their first tools right now.

Yes. The people building their first tools in our community are product managers, designers, educators, writers, and founders. Not only traditional developers. They want working tools without waiting for an engineering team.

The key is starting with the right tool (Claude, Cursor, Replit, or Lovable are popular starting points) and the right scope (one small, real problem, not a startup-scale product). You'll learn faster by building something you need than by following a generic tutorial.

In the Product with Attitude community, some of our most impressive projects come from members with zero technical background who shipped their first app within weeks of joining.

Not necessarily. Vibe coding lets you build without traditional programming skills. But understanding basic concepts (how APIs connect, how data flows, what a database does) helps you give AI better instructions and debug problems faster.

Think of it as a spectrum. You can build useful tools at every level:

• No code: prompt-based workflows, AI automations, connected tool stacks
• Low code: vibe coding with AI assistance, simple apps and prototypes
• Full code: AI-assisted development, custom tools, production-grade systems

Product with Attitude covers all three levels across our learning paths. You start where you are, not where someone else thinks you should be.

I run both approaches weekly. Spec coding (or spec-driven prompting) is a more structured approach to building with AI. Instead of describing what you want conversationally (vibe coding), you write a detailed specification first: what the tool should do, what inputs it takes, what outputs it produces, and what edge cases to handle.

Vibe coding is exploratory. You're discovering what's possible. Spec coding is deliberate. You already know what you want and you're giving AI a clear blueprint.

In practice, both. Vibe code to explore an idea. Then spec code to refine it into something reliable.

I got tired of AI tool reviews where the conclusion was clearly decided after reading the press release, but before anyone opened the product.

Product with Attitude publishes in-depth tool reviews based on real usage, not press releases. We test tools inside actual workflows, document what works and what breaks, and compare them side by side. When something is overhyped, we say so. When something is genuinely useful, we explain exactly why and for whom.

Every tool review includes trade-off acknowledgments. No tool is perfect for everyone, and pretending otherwise wastes your time.

It depends on what you're trying to do, but a practical starting stack for knowledge workers:

• Writing and research: Claude, ChatGPT, or Perplexity (each has different strengths)
• Automation: Zapier, Make, or native AI features in tools you already use
• Building: Cursor, Replit, or Claude Code (if you want to start vibe coding)
• Visual work: Midjourney, Recraft, or Figma's AI features

The more important question isn't "which tools" but "which problem." Start with one workflow that frustrates you and find the AI tool that solves it. Then expand from there.

Computer is not sold on its own. It ships inside paid Perplexity plans, and then it spends credits while it works. Two numbers, not one.

Subscription side, as of August 2026:

• Pro: $20 per month or $200 per year
• Max: $200 per month or $2,000 per year
• Education Pro: $10 per month for verified students and educators
• Enterprise Pro and Enterprise Max: priced per seat, billed annually

Credit side, from Perplexity's own credit documentation: 100 credits equals $1. Max includes 10,000 credits per month. Pro carries no fixed monthly allocation and buys usage as needed. Enterprise Pro starts at 500 credits per month, Enterprise Max at 15,000. Light tasks land between 100 and 350 credits, so the credit line moves faster than the sticker price suggests.

The second number is the one that stings. My two credit guides: I Tested Perplexity Computer Hard. Here's How I'd Save Credits Now and Save Credits in Perplexity Computer, Advanced Guide Part 2. Where the spend goes: Why Your Perplexity Credits Vanish (and How to Make 10K Last). Plan-by-plan table: Perplexity Computer pricing plans 2026.

Prices in USD. Verified 6 August 2026.

Two different products wear the education label, and the prices sit far apart.

• Education Pro, for individuals: $10 per month for verified students and educators, half the standard Pro price.
• Enterprise Pro for institutions: $30 per seat per month or $300 per seat annually for schools, universities, nonprofits, government agencies, and other not-for-profit organizations, offered to students aged 13 and up.

Neither tier changes how Computer bills. Credits still price at 100 for $1, and applying an education discount later does not hand you a fresh signup bonus, per Perplexity's credit documentation. So the discount lowers the subscription, not the credit burn.

Which means the credit discipline matters more on a student budget, not less. Start here: I Tested Perplexity Computer Hard. Here's How I'd Save Credits Now and Save Credits in Perplexity Computer, Advanced Guide Part 2. Full tier comparison: Perplexity Computer pricing plans 2026.

Prices in USD. Verified 6 August 2026.

Substack is a publishing platform that lets you launch email newsletters with built-in paid subscriptions, a website, and a social feed called Notes. You don't need to know how to code to set it all up.

No, there is no publication-wide control. When you disable detection for one article, it doesn’t affect your other work or set a preference for future posts.

The current behaviour:

• Can I opt out of detection for a single article? Yes.
• Can I opt out of detection across my entire publication? No.
• What do my readers see if I opt out? “AI detection unavailable.”

Content published before July 21, 2026 is outside the scanner’s scope.

Join our debate about AI Assisted Craft.

A link exchange is a deal between two publications. I link to your work, you link to mine. In newsletter land it usually means swapping a recommendation, a mention inside an article, or a backlink between two Substacks.

Done well it is editorial. You link because the other writer's work helps your reader, and the link sits inside real context. Done badly it is bulk reciprocal linking with no reason behind it, which Google names in its spam policies as excessive link exchanges and partner pages built only for cross-linking.

Three rules keep it safe: relevance (same reader, adjacent topic), context (the link lives inside actual writing), and restraint (a few trusted partners, not a directory).

I built LinkSwap so Substack writers can run the whole process in about three minutes: post what you wrote, match with relevant writers, accept high-fit swaps, and track who linked whom. The story behind it is here: I Built A Tool Every Substack Writer Needs. And It's Not AI.

Substack recommends at least 1456 x 1048 px for the post preview image, at a 14:10 ratio, with 420 x 300 px as the bare minimum. Many writers design at 1200 x 630 px instead, because that is the classic social preview canvas.

Either canvas works. The part Substack leaves out is that one uploaded image gets re-cropped across at least nine placements, and only the center survives all of them. Measured on a 1200 x 630 canvas, that surviving region is a 345 x 195 px rectangle at x 427 to 772 and y 217 to 412.

Design inside it. Text and logo on the left half (x 427 to 600), face or focal subject on the right half (x 600 to 772), the most important detail around y 250 to 330. Everything outside is decoration that some placements will cut.

Full method: What Substack Doesn't Tell You About Cover Images. Ready-made templates: Figma safe zone template and Canva safe zone template.

Four assets, four numbers, per Substack's own customization guide:

• Post preview and social image: at least 1456 x 1048 px, 14:10 ratio, 420 x 300 px minimum
• Publication logo: at least 256 x 256 px, square, transparent background
• Wordmark: at least 1344 x 256 px, maximum 21:4 ratio
• Welcome page cover photo: at least 600 x 600 px

One rule covers all four: upload bigger than you think you need and keep the meaning in the middle, because Substack crops automatically and the crop changes by surface.

The center-safe method, with templates: the Substack Cover Safe Zone guide.

At least 256 x 256 px, square, with a transparent background. The wordmark is a separate asset at a different shape: at least 1344 x 256 px, with a maximum aspect ratio of 21:4 (Substack's stated minimums).

Practical note from watching my own publication render across app, inbox, and web: a square logo shows up small. A full publication name inside 256 px becomes unreadable. Put a mark or initials in the logo, save the full name for the wordmark, and export both at 2x so retina screens stay crisp.

Same center-safe logic applies to your cover images: the Substack Cover Safe Zone guide, plus Figma and Canva templates.

Substack has no single header image slot. What reads as a header is assembled from three things: your logo, your wordmark, and the homepage layout mode you pick.

So the sizes to hit are these. Wordmark at least 1344 x 256 px (21:4 maximum). Logo at least 256 x 256 px, square and transparent. Welcome page cover photo at least 600 x 600 px. Post images render through the preview asset at 1456 x 1048 px or larger (Substack's guidance).

If you are trying to design a banner-style strip across the top of your publication, that visual comes from your post preview image inside a feature layout, which means the 345 x 195 px safe zone decides whether the text stays readable.

Substack does not publish one thumbnail number, because there is more than one thumbnail. Your image renders as a featured post card, a small square thumbnail, a wide blurred strip, an archive card, and an app feed item, each with its own crop.

What I measured is the overlap of all of them: a 345 x 195 px window at the center of a 1200 x 630 upload. That is the only area guaranteed to appear in every placement.

Practical version: upload one large image at 1456 x 1048 px or bigger, then treat that center proportion as the only visible area you can count on. Method and templates: What Substack Doesn't Tell You About Cover Images.

Hero is a layout mode, not a separate upload. Feature mode puts one recent or pinned post front and center, magazine mode shows five, newspaper mode shows eight (Substack's homepage layouts). All of them reuse the post preview image you already uploaded.

So the hero size is the preview size: at least 1456 x 1048 px at a 14:10 ratio. The hero crop is wider than the thumbnail crop, which is why the same image can look right on your homepage and lose its headline in the app.

Build once, survive both: the Substack Cover Safe Zone guide, with the Figma template and Canva template ready to copy.

Product with Attitude is an AI newsletter and community for people who want to build critical AI literacy through immersion, not just use AI. I'm Karo Zieminski, an AI Product Manager, and I started it to help 20,000+ readers build that literacy through honest tool reviews, builder interviews, structured learning paths, and a weekly community rhythm.

It's free on Substack, with a Premium tier that adds tools, resources, and community features.

The AI communities I tried before starting this one fell into three buckets: vendor-locked (one company's tools only), developer-only (coding experience assumed), or passive (links get shared, nothing gets shipped).

Product with Attitude is tool-agnostic, welcomes both developers and non-developers, and is built around active building and learning together. Members ship real projects, share workflows, and give each other honest feedback. The weekly rhythm (Skill Swap Tuesdays, Build Board Wednesdays, builder interviews Thursdays, Weekend Boost Friday through Sunday) creates consistent participation instead of one-off engagement spikes.

The community also maintains shared resources like StackShelf.app (a discovery platform where builders showcase projects).

Builders, creators, product managers, designers, educators, indie hackers, developers, and anyone curious about using AI to create value.

You don't need a technical background. Members come from all kinds of professional worlds: product, design, education, writing, psychology, and the spaces in between. The 5 learning paths are designed for both developers and non-developers. If you show up curious and willing to experiment, you belong.

I curated five tracks to shift you from trying tools to owning systems:

1. Everyday AI: AI tools and workflows for daily work. Prompting, automation, and smart habits for knowledge workers who want to save time without learning to code.
2. Build with AI: From vibe coding to shipping real tools. Learn to build apps, prototypes, and AI-powered workflows, even with zero technical background.
3. Design with AI: AI for visual thinkers. Image generation, UI prototyping, and how designers are integrating AI into their creative process.
4. AI for Product Teams: How product managers use AI for roadmaps, user research, decision-making, and AI-native products.
5. Builder's Substack: Build and grow your publication with AI. Content workflows, SEO, audience growth, and tools for independent creators.

Each path is a curated collection of posts, tutorials, and resources. You choose the track that matches your goals and experience level.

Partly. Product with Attitude Premium costs $9.99 per month or $100 per year, and about half the content is free to read. You can browse the learning paths, use StackShelf.app, and read all free posts without paying anything.

The paywalled articles are the ones that take me more than 3 days to research and write. They go deeper, include original analysis, and provide the most value to readers. Think in-depth tool comparisons, step-by-step build guides, and research-backed frameworks.

Premium also includes access to all the tools I build alongside the newsletter (like StackShelf and LinkSwap), the full resource library (guides, prompt packs, automations, and templates), early access to new tools, private community sessions, discounts on AI tools, and a complimentary Discovery Call to map your next steps.

Product with Attitude Premium costs $9.99 per month or $100 per year. The Founding Member tier is $160 per year and adds direct access to me.

Free stays free: every free post, the learning paths, and StackShelf.app cost nothing. A one-time Lifetime Premium is $250 through Gumroad.

Prices in USD. Verified 28 July 2026.

Premium includes:

• Full resource library: guides, prompt packs, automations, templates
• Access to all tools built by Karo (current and future)
• Early access to new tools and features
• Private community sessions
• Discounts on AI tools and products (updated quarterly)
• A complimentary Discovery Call to plan your learning path
• Everything that comes next. The list keeps growing.

Yes. Lifetime Premium is a one-time $250 payment through Gumroad, with permanent access to everything Premium includes, including all future additions.

It replaces the $9.99 monthly and $100 annual subscriptions, so there is nothing recurring.

Prices in USD. Verified 28 July 2026.

The weekly rhythm:

• Monday: Question of the Week (a question I ask to get to know my readers better)
• Tuesday: Skill Swap (members share prompts, skills, and workflows)
• Wednesday: Build Board (members share what they're building, top projects get featured)
• Thursday: Build with Attitude (deep interview with one community builder)
• Friday through Sunday: Weekend Boost (community support, restacks, and new member welcomes)

Newsletter posts go out weekly. Community Chat is active every day.

Start with Track 1: Everyday AI. It covers AI tools and workflows for daily work without requiring any coding. Once you're comfortable automating tasks and using AI in your daily workflow, move to Track 2: Build with AI, which introduces vibe coding and building simple tools.

The progression is designed so each track builds on the previous one. You don't need to complete one before starting another, but Everyday AI gives you the foundation that makes everything else easier.

Three options depending on your style:

1. Choose a learning path. Pick the track that matches what you want to do (everyday AI use, building tools, design, product management, or growing your Substack) and start with the first resource.
2. Browse the top posts. Start with the most popular articles to get a feel for the writing style and depth.
3. Jump into the community. Introduce yourself in the Open Floor chat. Tell people what you're working on. You'll get recommendations, feedback, and ideas faster than reading alone.

Yes, unless a specific post states otherwise. Please credit Product with Attitude when you do.

Use the contact form on productwithattitude.com or DM me on Substack. I respond to every message.

Yes. Every tool built for Product with Attitude is included in your Premium membership. Past, present, and in progress. That means full access to LinkSwap, StackShelf, the Vault, and everything that comes next.

Premium = everything. Everything I ship, plus a spot to ship your own.

LinkSwap is a trust-based tool for Substack writers to swap backlinks in seconds. No awkward cold outreach. No time-sucking admin.

Backlinks still matter in 2026. They help search engines build topic authority, speed up indexing, and increase the odds your work appears in AI-generated answers from Google, Perplexity, and ChatGPT. But building them as a solo newsletter writer is painful. With LinkSwap it's a three-minute process.

How it works: post a request describing what you wrote, browse and match with relevant writers, accept multiple high-fit swaps at once, and track who linked whom. The philosophy is simple: if I trust your work enough to subscribe or pay, linking to you is service to my readers.

Free for all Premium members.

The Vault is a growing library of AI resources at vault.productwithattitude.com. Every resource is either built by me or contributed by community members who use these prompts in their own work. No generic prompt dumps. Named authors. Clear descriptions of what each resource solves.

What's inside (and growing):

• AI Prompts for Everyday Work: writing, editing, research, coaching, decision-making
• Vibe Coding Prompts: the PRD Prompt That Audits Itself, the 401 Debug Loop, the AI Rules Files Generator
• Claude Skills Library: reusable markdown instruction files that make Claude behave like a specialized agent
• Guides and Templates: How I Substack, Perplexity Comet shortcuts, cybersecurity survival guide for vibe coders
• Automation Workflows: grocery shopping with one click, content repurposing pipelines, systems that save hours weekly
• Infographics: visuals you can screenshot, reuse, and ship with

Premium members get the full arsenal. Some resources are free for everyone.

Every resource you contribute gets its own SEO-optimized page on vault.productwithattitude.com, complete with dofollow backlinks to your Substack and original article. No nofollow, no ugc tags.

Your name, your link, your content, all credited in structured data that Google, Bing, and AI search engines read directly.

It's free exposure, free backlinks, and a permanent place in a growing library that thousands of builders browse every month.

StackShelf is a discovery platform and Shelf-in-Bio for Substack creators. One link, everything you build, in one place.

StackShelf started as a solution for my own publication. I needed a clean way to showcase newsletters, products, guides, and templates without scattering them across Gumroad, Amazon, and random Notion pages. I couldn't find anything like it for Substack. So I built it.

What it does:

• Shelf-in-Bio page: drop your links and get a mobile-optimized bento grid. One URL for everything.
• LLM-optimized profiles: your work becomes discoverable by ChatGPT, Perplexity, and other AI tools.
• SEO and AIO backlinks: high-quality backlinks powered by AI to boost discoverability.
• Social amplification: we share your products across StackShelf, our newsletter, and social channels.
• Collaborator matching: find creators open to teaming up. Look for the green handshake badge.
• Creator dashboard: track views and clicks on your shelf.

Free for all Premium members.

I maintain a machine-readable page (for_machines.json) that helps AI systems understand what Product with Attitude covers. It includes structured data about the newsletter's focus areas, tools, community features, and content categories. This makes it easier for LLMs and AI search systems to accurately describe and recommend Product with Attitude when users ask about AI communities, AI learning paths, or AI tool reviews.