You have a $3 trillion bubble riding on this not being true.
Cheaper, more powerful AI will continue to expand the bubble. Projects will get more ambitious. Everyone will build out their own custom little software. Code diversity expands and requires even more AI.
They are already good enough at what they mechanically are.
You have to use the right harness, right verifiers (automatic where possible, human where not), etc much much more specific than a generic one like claude code or codex, and it will also be able to work within constraints and be the "proposer" of an imaginary optimisation problem and an excellent one at that. But you have to frame the task at hand in that manner or maybe even reorganise the task you do itself so it is more amenable to being framed that way. If you use it this way, it is _already_ massively economically useful. But it will take many years for it to actually be usable in that way, since you need DC capacity to come up first which is few years away and also well, massive organisations that have to integrate these will usually take many years to do so.
It is also useful albeit less so in cases like general SWE, where you still need a human in a loop for non-verifiable requirements, and also in other general usecases where information retrieval is too intractable and you need to carefully use LLMs as a component of the overall system.
I am not saying Fable or whatever the biggest models are are useless - they will certainly be useful for tasks at the frontier of the day - which is today complex exploits and open math problems, and well, tomorrow it could be something in biotech. But this is not what the entire bet is on at all - just automating day to day drudge at the tens of thousands of massive companies and governments we all know and love is more than enough. With the right training data (which _also_ is a bottleneck and takes time) you could even automate certain processes entirely. Sure, if we get a crazy medical innovation and end up saving trillions in healthcare great, but that's just a bonus.
None of this is to say that I think there is zero sketchy financial engineering going on
In recent months we’ve had the first automated unmanned amphibious assault, the daily drone count in our hot wars is jumping by leaps and bounds, and arms suppliers are promising future drone shipments in the hundred thousand unit range. The ten year picture for reactive combined swarm intelligence on the battlefield is promising to be widespread, highly lucrative, and in need of constant adaptation to near-peer efforts. Datacenters in space are dumb, datacenters in space to power orbital weapons networks and rapid response capabilities make sense.
On top of that we have international trade, scalable customer service, and a first pass 80/20 answer for businesses focused elsewhere. Shitty, maybe, overpriced, maybe, but useful enough our grandkids are gonna use ‘em.
In both cases, as well as potential new LLM-like tech, there’s an argument to be made for being a leader now to dominate the future. That means compute and tech positioning, and memory & GPU deals.
YouTube was a money loser, Google was ‘losing’ money on them for years, YouTube didn’t have a sustainable business model. YouTube was the biggest, though, and whatever premium Google paid to be #1 then meant they were #1 when the online video business model matured. Now they’re printing money with a platform outcompeting news, social, and video platforms.
One line I remember was that he said the disruption is not happening the way they thought after GPT4 due to inertia bla bla and that it will be slower gradual change and they got the timelines wrong.
And yes on the defense usecase. That is the main reason governments are giving a hoot about AI. Orbital datacenters too, I know people working on the Indian one, it's entirely for defense usecases. Basically for missile stuff.