Where Are We in the AI Cycle?

1. The 64K IBM PC Era of AI

"We're at the 64K IBM PC stage of microcomputers. People are still figuring out how everything works."

AI today is focused on basic problem-solving, similar to early PCs where even fundamental operations were experimental. AI claims to replace search and spreadsheets, but in reality errors are frequent and performance falls short of expectations.

2. Vibe Writing and Vibe Coding

Vibe Writing (AI-generated text): Already achieving full autonomy — college students use it for assignments; businesses are evaluating it.
Vibe Coding (AI-generated code): Works well because developers are the platform's customers, but significant constraints and errors remain.

3. The Editor's Role

"AI can generate text, but when accuracy truly matters, a human must review it."

We've already seen lawsuits citing nonexistent legal precedents. The shift is: humans don't get replaced by AI editors — humans become the editors of AI output.

4. The Agent Decade

Using Iron Man's Jarvis as an analogy: autonomy exists on a slider from manual to full auto.

Agents will first handle low-judgment, repetitive tasks (loan comparisons, flight searches).
High-judgment tasks (tax filing) remain difficult to automate.
"We've entered the 'decade of agents'" — this will take much longer than expected.

5. Automation's Limits

"My job is all uncertainty. A spreadsheet that pretends to be certain doesn't help my uncertainty."

Most jobs are about exception handling and judgment. Full automation requires inputting every exception manually — at which point it's barely different from manual work.

6. Vibe Coding's Illusion

Prompts may look like English but are really programming. Past "low-code/no-code" waves only worked for simple apps. Current AI is still at the prototype stage.

"You see a cool demo on Twitter, but 3 days later it doesn't work."

7. AI Novels, Art, and 'Slop'

"Within a few years, there will be a bestselling novel almost entirely written by AI."
AI excels at average output but needs human creative direction for cutting-edge work.
"The world needs more slop" — lower barriers give more people creative opportunity.

8. Accessibility Changes the Standard for Excellence

"Does success need to be perfection, or is 'better than what people can do today' enough?"

For 80% of the world lacking basic services, even average AI output is vastly better than nothing.

9. Google's Future

"Google's demise is ridiculous. The fall of big companies doesn't make sense."

Big companies have "shock and awe" capabilities — the real question is whether they change their product development and market approach.

"We're at the very beginning of AI development. Everything is far more complex, full of exceptions and uncertainty."