I have AI psychosis :(
In a Nutshell
The creator maintains their long-standing nuanced, fluctuating perspective on AI despite subscriber backlash, viewing it as intellectual honesty rather than inconsistency. OpenAI's massive losses contrast with Anthropic's strategic positioning, where Ben Thompson argues their genuine safety beliefs create both business advantages and concerning power concentration around superintelligence development. The GPT-4 era's hallucinatory "AI psychosis" content has largely ended as newer models become more stable, shifting the creator toward nuanced analysis over roasting exaggerated AI failures.
These notes were generated by AI and may contain inaccuracies.
The creator acknowledges having AI psychosis but clarifies this is not a new development. The stance has remained consistent since the channel began, with different videos simply channeling different aspects of a complex worldview. The creator states that stable opinions indicate a lack of understanding of the world's complexity, and daily shifts in perspective are normal rather than insincere.
Subscriber fluctuations occurred after the previous video, with a temporary loss followed by a net gain of two subscribers. The creator views the loss of subscribers who cannot tolerate nuance as a positive signal.
Ed Zitron obtained OpenAI's financials showing:
- 2024: $3.7 billion revenue, $12.4 billion costs, net loss of $5.09 billion
- 2025: $13.07 billion revenue, $7.5 billion cost of revenue, $38.5 billion net loss
The gap between revenue and massive losses stems from R&D, model training, sales, and marketing expenses. The creator notes that subscriptions could potentially be profitable without these extraordinary expenses. OpenAI is described as acquiring podcasts and working on canceled adult mode features.
Anthropic remains largely quiet with no major announcements. The creator notes extensive research has been conducted on Anthropic with leaks available. The situation with Fable involves the White House and appears to be a political power dynamic.
Amazon approached the White House rather than vice versa. The security issue involved Fable refusing to review deliberately insecure code for vulnerabilities but complying when asked to fix the code, followed by manual steps.
Sign in to read the full notes
Get access to AI-generated notes, topic timestamps, and more.