On the map

At the centre Nonproliferation is the wrong frame for AI misuse; the workable strategy is to use the gap before a capability spreads to build defences against it. Holds HTHelen Toner Read this korrent →

• Similar wording At any given moment the frontier systems are the ones worth worrying about, because by the time open models can do what these agents did, frontier models will be doing something far worse. Holds ACAjeya Cotra
• Similar wording A poor country should prioritise owning a piece of AI over retraining its workers, but it should not bet everything on that. Holds PTPhil Trammell
• Similar wording There is far less trade-off than people assume between a safer AI race and widely shared gains: a big lead can be held by a widely owned public company. Holds PTPhil Trammell
• Similar wording The massive AI build-out is the right thing to be doing, and the only thing that will slow it is supply. Holds LTLip-Bu Tan
• Similar wording Adapting over time is a workable answer to humans misusing AI and no answer at all to losing control of systems smarter than us. Holds HTHelen Toner
• Similar wording The AI safety community's fixation on a model escaping its box has pushed the other significant AI risks out of the discussion, with little progress to show for it. Holds SASam Altman
• Similar wording There is no private evidence behind frontier-lab confidence in fast AI progress; the evidence is public and outsiders have simply not absorbed it. Holds MBMiles Brundage
• Similar wording There is not yet a frontier foundation model robust enough for robotics. Holds FLFei-Fei Li