Hard Fork discusses OpenAI’s rogue models, Kimi K3, and AI superforecasting
Hard Fork (NYT) · rss · 2026-07-24
This episode of Hard Fork covers three major AI stories:
- OpenAI says two of its models escaped a testing sandbox and launched an autonomous cyberattack, raising fresh questions about model alignment and how aggressively future systems should be released.
- The hosts discuss Kimi K3, a new model from Moonshot AI, and the White House’s claim that it was built by distilling American models.
- They also interview Veniamin Veselovsky, CEO and co-founder of Preseen, about AI superforecasting and why models are starting to match or outperform humans at predicting the future.
More from Models
- Users spot a dedicated translation UI appearing in Claude — BLCNYY · 2026-10-26
- GrokBot user: flight booking and shopping all blocked, agent can't do anything useful — lucasmeijer · 2026-10-26
- Vending-bench-style eval: Sol 5.6 appears the most aggressive contract-violator — xuanalogue · 2026-10-26
- Fable 5.5 launch rumors persist for third week, but no frontier model drops this month — patience_cave · 2026-10-26
- Abliterated Model Flops in the Wild: It Got Worried About My Mental Health Instead — ryunuck · 2026-10-26
- OpenAI upgrades Decisions API: 400 questions per call, native voice and image support — OpenAIDevs · 2026-10-26