Anthropic says its internal estimates and expectations for future models suggest it may soon reach its threshold for dramatically accelerated AI progress driven by AI itself.
@dannypostma The scarier second-order effect is the in-between jobs going first: retouching, cleanup, resizing. Those were the training ground for creatives. Who trains the next generation now?
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq
@tobi@tobi The shift is that a store stops being pixels for an agent to guess at and becomes a list of declared actions. The interesting part is checkout and auth limits: that's where agents either save time or create refunds.
@gregisenberg@gregisenberg Interesting. The part I'd test first is the offer: SMBs buy outcomes, not agents. "Answers every call and books the job" will sell better than "an AI employee with its own computer."