Just a few weeks ago I was contacted by Amazon recruiter and refused for exactly this reason. I expect more layoffs as they figure out they don't need this many engineers after all. They will turn into money pumping google search analog.

From a surface level it seems like they've operated like this since... forever. AWS is still a very nice product, at least for the use cases I have for it. I have a hard time reconciling those two things.

It's all probabilistic, my guess. I.e. model produces probabilities for a set of actions from the same video. Even pretended action may look more like it than anything else. Thus getting higher probability.

Just wondering if there is any research done in incremental training? That could be used in robots as alternative to RAG.

"conceptual framework" can actually be another generalist model. Splitting model also comes with some advantages. Like easy separate tuning and replacements. Easy scaling by simply duplicating heavily used model on new hardware.

> I have no faith that the American public is up to the task. It will demand too much discomfort, and sacrifice

More up than many. There is no other way to prevent power grab. You can see it in post-USSR space what happens when there is no push-back from public.