Phase I // 2023–2024

CHATBOT

Fascinating at first.

Hallucinations Inaccuracies Flattery

It flattens out.

Phase II // 2025

Vibe Coding

Mixed results.

MCP

Phase III // 2026

Autonomous Agents

Summary /// Andrija Arapovic

Will AI-Automation last forever?

The founder of a large AI automation community says the business model won’t keep running the way it has. And yet it’s still the best industry to be in.

The thesis

No business as usual.

Nick, founder of “Maker School” (by his own account around $10 million in revenue from AI automation), sees two reasons why the business is fundamentally changing.

Reason A

Benchmark satu­ration

0→60% Automation Bench & GDPval
<6 months to get there
+5–10% per month, extrapolated linearly
Dec–Mar saturated

Agents already automate large parts of a business, with no human in the loop.

Reason B

The knowledge gap is shrinking.

You used to get paid for knowing more than your client. That edge is getting smaller.

He expects a sudden jump in awareness once demos of AGI-level capabilities hit the media.

Analogy

See: web design.

Site builders and agents have commoditized web design. The only people still earning well there are exceptional: at sales and marketing, or at the product itself.

The floor keeps rising.

Stand still and you fall behind, relatively.

And yet

The industry that auto­mates stays.

Every white-collar industry is being transformed. According to him, the industry that delivers the automation is the most durable.

Forecast

Timeline

  1. H1 / 2027 Many business owners realize what AI can do and look for human help. Probably the most profitable year yet. But possibly the beginning of the decline.
  2. After that ??? He says openly that he doesn’t know.

Recommendations

5 Moves.

  1. Move away from drag-and-drop. Make, Zapier and n8n get thinned out. More Claude Code, internal web apps.
  2. Higher-bandwidth outputs. Images, video, 3D, simulations.
  3. From implementation to enablement. Workshops, team training, consulting, upskilling.
  4. Seize short-term opportunities. Growth hacking, while it lasts.
  5. Sales, marketing, delivery > tech. 100×→1000×

The FUD moment

Fear­monger­ing?

At the end of the video, Nick anticipates the accusation that he’s spreading fear, uncertainty and doubt (FUD).

He stresses that the video isn’t meant to scare people but to help them move to a more effective business model. As evidence he points to recent progress: agents working on mathematical Millennium Problems like Navier–Stokes, and agents creating works of art. You should at least think about what that means for the next few years.

Right after that comes the sales pitch for Maker School (guarantee: first client within 90 days, over 2,000 members).