Podcaster Stewart Alsop III on AI psychosis, token maxxxing, and why LLMs aren't AGI
StewartalsopIII · x · 2026-09-08
Stewart Alsop III (Crazy Wisdom podcast) reflects on months of "AI psychosis" and token maxxxing:
- He finally built his in-house podcast recording system with AI — something he never thought he could build alone.
- From March to June he fell into "the bad kind" of AI psychosis and learned the hard way that LLMs are not AGI — the failure showed up in visual/audio tasks.
- Now he's shifting from endless prompting to distribution: running 6 distinct knowledge-management services, and as an introvert realizing that building the "vortex of movement" around meetings matters more than sequestering himself in token-maxxing sessions.
A first-hand look at how even heavy AI users can over-rely on chat while real growth still comes from human connections.
More from AGI Musings
- Investor: the one dev service worth paying for is schema audits — StewartalsopIII · 2026-09-08
- Thinking of AI as a decently smart human with infinite patience and willpower — shauseth · 2026-09-08
- Robotics researcher to AI hypesters: debug a controller before making grand claims — GeorgiaChal · 2026-09-08
- François Fleuret: coding agents won't replace one worker, they'll integrate the whole company — francoisfleuret · 2026-09-08
- Reddit asks: how would a China-US AI safety pause even work in practice? — sunstersun · 2026-09-08
- François Fleuret: pre-AI software is awful, coding AI will rebuild what SaaS solves — francoisfleuret · 2026-09-08