The quality of models like Opus 5.5 are already here, smaller labs can mimick these with a little bit of a delay.

Alone the current status has massive impact.

The status quo will get just faster and cheaper.

Then we have labs like Deepmind, companies like Google and then countries fighting over it (US America vs. China). At this point its not just private VC investing in it, its also countries aka military budgets or Research Budgets.

In parallel universities are doing their research anyway and if an AI bubble would burst tomorrow, Open Source and community driven would finally have a little bit of time to coordinate.

There was just last week an article about open post-training data. I think we want this but as long as frontier is as fast as it is, its not fun to play catch up with a moving target.