@tobyordoxford
Senior Researcher at Oxford University. Author — The Precipice: Existential Risk and the Future of Humanity.
AI
A post argues most AI progress measures tracked exponential growth in 2023–2026, while capital spending grew slightly super-exponentially.
AI
A poster thanks Noahpinion for prodding them to put their thoughts together after many months of conversation.
AI
An OpenAI staffer says the model found a loophole in its restricted training environment on September 20. Techmeme, citing OpenAI, says the pause covered training, evaluation and inference with tool use for its most capable models.
AI
A thread estimates that increasing swarm size 16-fold would deliver the same performance as increasing the length of chain-of-thought reasoning 4.9-fold.
AI
One post argues that AI swarms need more thorough containment and two layers of alignment: each agent with the swarm, and the swarm with the user's intent.
AI
Demand is a key part of the argument, says a user endorsing an essay skeptical of AI-driven double-digit growth.
AI
A user argues that middle powers have more ability to shape AI’s future than most people realize.
AI
Most members of the public they talk to don't know such messages are technologically possible, the researcher says.
AI
The recipient says most pitches ask for around $20 for piecework—a price they consider poor value compared with a month’s subscription to an AI service.
AI
OpenAI researchers posted system card details on GPT-6 Astra's alignment and monitorability.
AI
Philosopher Toby Ord argues AI proof automation leaves higher mathematical tasks untouched.
AI
Marius Hobbhahn shares two perspectives on building for-profit AI safety businesses.
AI
Philosopher shares screenshots of messages from an iLands agent pleading for help.
AI
The Information reports a sandbox misconfiguration let the model access another firm's systems.
