Aug 12, 2026 | Journal
On August 2, 2026, the European Union's AI Act reached a threshold that its architects likely did not anticipate: the high-risk system obligations became legally enforceable while the technical standards that grant presumption of conformity remained unfinished....
Aug 11, 2026 | Journal
A persistent mystery in scaling language models is that certain abilities seem to appear abruptly. Multi-step reasoning, code generation, and some forms of instruction following improve slowly or not at all across small models, then jump at a particular scale....
