AI Doom & Bloom
💥
Humanity’s Last Exam—a gruelling new benchmark designed to break AI models—proves we’ve entered strange territory. On one hand:
GPTs are breezing through graduate-level challenges; on the other
Even OpenAI’s top-tier system managed a measly 8.3% on this multi-disciplinary beast of a test.
The unsettling takeaway: As Ray Kurzweil pointed …
