It was all telegraphed well in advance. For those willing to see.
As I've mentioned in several of my posts yesterday,
just released their first "reasoning" model(s), o1-preview and o1-mini. These are the first sets of models that use some kind of "reasoning" on top of LLMs to greatly augment the abilities of the AI systems. Now, there is a lot to be said about all of those claims, and I may address many of those in my future posts. However what I want to address in this article is the fact that none of the announcements yesterday came as any kind of surprise to anyone who paid even cursory attention to what has been going on in AI over the past year. I spent my Summer in Silicon Valley (as is my custom) and tried to listen to what the rumors are "on the street". However, most of those rumors did not go far beyond what we already knew based on scattered posts here on X. We knew about "Q*/Strawberry" at least since the November of lat year. We know that most of the leading AI labs are actively doing research on "reasoning", and are getting very good at it. We know that the next generation of the leading LLM models has already been trained, or is close to. We even know, thanks to some well informed insiders, when the next big AI announcements are about to drop. I even tweeted about it in one of my posts yesterday. We are at the place where all the details may not be in place, but the big picture is already well established. Now we are going through the moves of the endgame.
Scaling laws work. They are far from being exhausted. Lots of compute and engineering progress will get us really far, as far as we can humanly take it. Much better and far more powerful models have already been built, and are tested and used by all the leading AI labs. View all the weaknesses, apparent and real, of the current publicly available models and services as an opportunity to add your own insights and value. Do something. Build something. Sure, you can be a stick in the mud naysayer douche, but that grift is gonna feel really stale and exhausted really soon.
Leading labs are not stupid. They know what's at stake. They are all trying to play their cards the best they can. In the near future enormous fortune and riches depend on it. In the long term we are literally talking about the future of humanity. No, your normie neighbor will not appreciate the enormity of the situation, but you can do yourself a service and try to see how best to position yourself. X is probably the best public platform that can inform you about all of his. Learn who the reliable sources are, and follow them.