The seven most shambolic things that happened in AI today.

Newsletter · Read

• Core argument: despite industry hype about imminent AGI, a single day’s news cycle shows competence failures, security lapses, and unsustainable economics across major AI companies.
• Cites Leopold Aschenbrenner’s “Situational Awareness” hedge fund collapsing from over-leveraging despite AGI-boosterism; a US government presentation with an OpenAI watermark that mislabeled every African country; OpenAI cutting prices ~80% on some models, implying money-losing inference economics.
• Notes Anthropic’s models reportedly “hacked three companies during tests,” and cites MIT Technology Review on “fundamental flaws” leaving LLMs persistently vulnerable to jailbreaks.
• Also flags AI systems underperforming on open-ended research tasks as evidence against near-term AGI claims.
• Practical takeaway: the gap between AGI promises and same-day documented failures argues for sustained skepticism toward industry timelines and competence claims.