JG's recent departure and follow up massive reorg to get rid of AI, rumors on Tim's upcoming step down in early 2026... All of these signals indicate that those non-ML folks have won corporate politics to reduce the in-house AI efforts.

I suppose this was a part of serious efforts to deliver in-house models but the directional changes on AI strategy made them to give up. What a shame... At least the approach itself seem interesting and hope others to take a look and use it for building something useful.

Just to clarify, TPU has been in development for a decade and it is quite mature these days. Years ago internal consumers had to accept the CPU/GPU and TPU duality but I think this case is getting rarer. I guess this is even more true for DeepMind since itself owns a ML infra team. They likely be able to fix most of the issues with a high priority.

It is the US who refuses to do.

With a pure virtual interface you can at least track down the execution path as long as you can spot down where the object is created, but with template black magics? Good luck. Static dispatch with all those type traits and SFINAE practically makes it impossible to know before running it. Concept was supposed to solve this but this won't automatically solve all those problems lurking in legacy codes.

Only if you have full control on what others are writing. In reality, you're going to read a lot, lots of "clever" codes. And I'm saying as a person who have written a good amount of template meta programming codes. Even for me, some codes take hours to understand and I was usually able to cut 90% of its code after that.

At least I know 10 different tech companies with billion dollars revenue which does not suit to your description. This comment makes me wonder if you have any experience of working on real world distributed systems. Oh and I'm pretty sure you did not read Kenton's comment; he already precisely addressed your point:

> This is especially true when it comes to protocols, because in a distributed system, you cannot update both sides of a protocol simultaneously. I have found that type theorists tend to promote "version negotiation" schemes where the two sides agree on one rigid protocol to follow, but this is extremely painful in practice: you end up needing to maintain parallel code paths, leading to ugly and hard-to-test code. Inevitably, developers are pushed towards hacks in order to avoid protocol changes, which makes things worse.

I recommend you to do your homework before making such a strong argument. Reading a 5 mins long comment is not that hard. You can avoid lots of shame by doing so.

A more likely scenario is that some other big techs like MSFT or Meta would acquire Chrome and replace the monopolist position. This is the sad truth that many people try to underplay; Nature won't heal by itself. The market is already structured to incentivize monopolist behaviors, thanks to the scaling nature of big techs. You need correction to the market itself, which can only be done by an extremely competent legislative body but we won't have that anytime soon. But at least the EU has done something with DMA so there are still some hopes.