Replying to @VladZhidkov
Looking at Anthropic’s Economic Index report, we know that SWE usage dominates.
Other industries are stickier but their usage will have to grow a lot to replace SWE revenue.
Also the report hints that SWE is API-pricing-heavy while the rest are likely lower-margin subs.
1
31
Very clever. The key implementation detail:
People keep saying 2026 will be the year of continual learning.
But there are still major technical challenges to making it a reality.
Today we take the next step towards that goal — a new on-policy learning algorithm, suitable for continual learning!
(1/n)
1
74
Replying to @dan_hues
1lb of ground-beef/steaks/salmon, and on most days at least 1/2 of that was beef of some kind.
1
1
36
Replying to @lishali88 @Rosebud_AI
Got it couple days ago to test-drive. It's great hardware: pass-through/MR is killer for productivity, hand gestures work well, it's super-compatible with eye-glasses. Don't think I will keep it but I am certain it will have a prominent place in the future of productivity.
1
2
61
1/3 What impresses me the most about ChatGPT is that iterative UX works so well: if I, the operator, am not happy with the output, I don't have to start from scratch, instead I can point to that "one thing" which I want the model to change in its output.
1
1
that's the one: youtu.be/ppFyPUx9RIU?t=220
1























