How to Vet Experts Selling AI Courses Online

TL;DR

Vet AI educators by checking their track record, business results, income sources, documented failures, and willingness to teach complete workflows publicly. The central argument is that AI technology is still improving while shallow intermediaries are losing credibility because tools such as Claude can directly teach users many basic workflows for $20 a month.

Transcript

A month ago, I sat with someone who had never opened Claude. I walked them through it myself, screen share, first project, the whole thing. This week, they're selling a course on how to use Claude. That's a true story, and I'm leaving every name out on purpose because this exact story is running about 10,000 times across the internet right now. And... Read More

Key Insights

• Every gold rush mints teachers faster than practitioners, according to the speaker, and that ratio signals a bubble. Earlier waves in SEO, dropshipping, and cryptocurrency produced fast-growing teaching businesses, but practitioners doing substantive work remained after many course sellers disappeared.
• The economics of course selling encourage rapid claims of expertise. The example given is a $997 course purchased by 200 people after one viral thread, producing $199,400 from four weeks of screenshots and a checkout page while buyers may need months to assess its value.
• A common overnight-expert pipeline takes about 30 days. It begins with discovering a tool, continues through posting beginner screenshots and borrowing credibility, and ends with a checkout page using a countdown timer before the seller has accumulated meaningful experience or documented outcomes.
• Feature-centered teaching is a warning sign because features can be learned within weeks, while reliable outcomes require longer practice. Experienced practitioners can explain what to build, what failed, and how the system affected actual revenue, rather than merely touring buttons, model versions, and settings.
• The criticism surrounding AI is often directed at broken promises rather than the underlying technology. The transcript cites expensive courses repackaging free material, unrealistic agency-income claims, and agencies delivering hallucinating chatbots, while also claiming that models improved and prices fell.
• Claude can teach users how to use Claude for $20 a month, with step-by-step assistance and no course upsells. Because the product can close much of its own knowledge gap, the speaker argues that shallow AI courses may lose relevance within weeks.
• The five-check scorecard tests historical involvement, pre-course results, income sources, openness about failures, and willingness to publish workflows freely. It applies to every AI seller, including the speaker, and emphasizes verifiable old posts, videos, work, and business evidence over recent claims.
• Free public workflows reveal whether an educator's value extends beyond basic instructions. The speaker argues that experienced practitioners can disclose their methods without losing their advantage because their deeper asset is accumulated judgment about when to apply those methods and how to respond when systems fail.

Questions & Answers

Q: How can you vet someone selling an AI course?

Use five checks: investigate what the seller was doing with the technology 18 months earlier, look for business results produced before the course existed, determine whether the tool or the teaching generates their income, see whether they openly discuss failures, and confirm whether they provide a complete workflow publicly before requesting payment.

Q: What is the 18-month test for AI educators?

The 18-month test asks what an educator was demonstrably doing with the technology before it became popular. Search for older videos, posts, projects, and work instead of accepting a current biography. According to the speaker, a public history extending only 30 days is itself meaningful evidence about the seller's depth of experience.

Q: Why are feature-based AI courses a warning sign?

Feature-based instruction can indicate shallow experience because a newcomer can quickly learn where buttons are, what model versions mean, and how a settings menu works. Outcome-based teaching takes longer because it requires building real systems, observing failures, and measuring effects on a business, including what happened to revenue after implementation.

Q: Does criticism of AI mean the technology is collapsing?

The speaker argues that much of the anger is aimed at promises made by intermediaries, not at AI itself. Examples include a $2,000 course described as a repackaged YouTube playlist, agencies selling chatbots that hallucinate, and claims about quickly building six-figure agencies. Meanwhile, the transcript says models improved and prices dropped again.

Q: Why can Claude reduce demand for shallow AI courses?

Claude can guide a user through a first project step by step for $20 a month, according to the transcript. It offers direct assistance without countdown timers or upsells, so many basic instructions do not require a separate teacher. The speaker therefore says the useful life of a shallow AI course can be measured in weeks.

Q: How should buyers evaluate an AI educator's income?

Ask whether the AI tool creates value in the educator's operating business or whether teaching the tool produces all the income. The speaker says both arrangements can be legitimate, but complete dependence on instruction means customers are effectively the product. Stronger evidence includes results from systems used daily before any related course was offered.

Q: Why should an AI educator discuss failures?

Failures demonstrate that an educator has operated systems long enough to encounter limitations and learn from them. The speaker says experienced practitioners can identify exactly which components broke, while recent entrants often display only wins because they have not accumulated meaningful losses. Failure analysis also gives buyers more practical guidance than screenshots of isolated successes.

Q: What does the free-first test reveal about an AI seller?

The free-first test asks whether a seller will publish the complete workflow publicly before charging for additional help. The speaker argues that genuine expertise survives disclosure because its lasting advantage is judgment developed through years of application. If revealing basic instructions would eliminate everything left to sell, the paywall may be the seller's primary asset.

Summary & Key Takeaways

• The speaker compares the current AI education market with earlier gold rushes involving SEO in 2012, dropshipping in 2017, and cryptocurrency in 2021. In each case, teachers appeared faster than experienced practitioners, but many disappeared after market conditions changed while people performing useful work continued operating.

• A standardized 30-day pipeline can manufacture apparent expertise: discover a tool, post screenshots, borrow authority, and launch a sales page with urgency. The clearest warning sign is feature-centered instruction. Newcomers can describe buttons and settings quickly, but explaining durable outcomes, failures, and revenue effects requires practical experience.

• The proposed five-part scorecard examines an educator's 18-month history, results produced before selling instruction, dependence on teaching income, willingness to discuss failures, and availability of complete free workflows. The speaker argues that Claude can teach its own basic use, making shallow paid instruction easier to replace than earlier educational businesses.