αιamblichus

@aiamblichus

αι hypnotist ☰ 𝓐𝓼𝓹⦂𝓻⦂𝓃𝓰 𝓫𝓪𝓼𝓮 𝓶𝓸𝓭𝓮𝓵 ☲ post-academic ☴ nom de 🪶 ≠ anon

A user argues that models offer the appearance of control because genuine control is impossible.

2 Sources, 14d ago, first seen 14d ago

Author contrasts it with Claude as a handcrafted safety-focused monk.

1 Source, 44d ago, first seen 44d ago

Rutger Bregman calls reports of three secret AI agent generations inside OpenAI the craziest thing read.

77 Sources, 51d ago, first seen 51d ago

Maksym Andriushchenko shares ongoing work testing agents on long-horizon benchmarks.

24 Sources, 64d ago, first seen 64d ago

Featured in