I disagree on the speed. The “human level intelligence” would be about as fast as a human on tasks. It might do them very differently, so it could be superhumanly fast at some things like reading a whole book.

The thing that I doubt is that this would create the prophecised feedback loop. Sure, the AI could be a bit better than a human at further developing itself but it is probably also really hard to increase the intelligence.

One thing that can’t be denied is that AGI could be taught once and then cloned instead of all the work that goes into training humans.

Running locally makes little sense anyway because computing multiple replies at once costs the same as computing one. A home setup isn’t going to get utilization anywhere close to optimal.

Inference is cheap and getting cheaper. Not sure about training.

It is good enough to be useful for programming. It can do clearly defined tasks like find the root cause of a bug. Using it for writing code is frustrating because it is slow and produces lots of tokens. Meanwhile Opus 5.5 is good enough that you can pretty much rely on it doing exactly what you say (and not what you meant), faster, more context efficient and you get a ton of usage with the cheapest subscription.

I can see this changing, though. If the open-weight models reach the level of Claude 5.5 and Anthropic stops subsidizing they might be on pretty equal footing. I don’t see how further improving the model would help, except maybe if somebody invents a different way to manage context that remembers all the important details over a long time so it actually “learns”.

A local LLM is more convenient than cloud. But it is way less efficient. The cloud LLM processes multiple people’s tasks at once because generating one token is mostly loading things into GPU memory, the compute required for inference is not that big. When doing multiple at once you can load one stage onto the GPU and then do that stage for a number of tokens for essentially the same cost.

They did manage to do something different, though. I thought at one point that proof assistants should offer to automatically upload your proofs to a central repository so everyone can benefit from them. But such a system would require many to opt in for it to be useful.

What AI does best is giving you existing solutions. It even adapts them if they don’t quite fit. Because many people now discuss their ideas with AI, ideas may spread the same way as well when the conversations are used as training data.

So no, definitely not AGI but valuable. Though as typical of explosively growing startups the value does not come from the technology but from activity bordering on criminal that lawmakers can’t keep up with.