Microsoft’s 3 billion bet on Copilot is assessed to be a disaster

[Norion]

curl-6 said:

Norion said:

Computers in the 1950's were not doing something as advanced as disproving something mathematicians thought was the case for close to a century. Thinking it's overhyped is fine but I do highly suggest looking up what mathematicians are saying on it if you haven't yet. There is variance in opinion with some thinking it's a bigger or smaller deal than others but it being significant has broad agreement with some worried that their degree will be mostly worthless in a few years due to AI completely taking over the field.

What AI companies said we'd have ASI by 2026? It's still flawed but top end models and in particular more powerful internal models like this OpenAI one are reliable enough now to be a big deal and do major things. Another big thing is in the Ukraine war drones are beginning to have autonomous capabilities impacting the war. More than anything a country fighting for its continued existence trusting the technology and benefitting from it proves just how big of a deal this stuff has already become I'd say.

Computers in the 50s were vastly less powerful and solving simpler problems, yeah, but that's the progression of tech, more powerful over time.

Musk has claimed Singularity will happen this year, he's famously full of shit, but still. Anthropic's Dario Amodei also said a while back AGI could happen in 2026.

Autonomous weapons aren't strictly a new thing either, drones have been fighting in wars for years now. A country fighting for survival's obviously going to take more risks with this stuff, that's doesn't mean it's inherently a good idea or a sustainable path forwards.

Right though what I'm saying is that only very recently has AI gained this level of capability in math so has now become capable of advancing mathematics. There are absolutely over optimistic predictions but something notable is the typical expectation among experts of when it'll happen has become a lot closer to the present the past few years. Due to the rapid advancement there is increasing expectation of it being achieved in the 2030's.

Them getting increasing autonomous capabilities is a recent development though and the usage of those capabilities has been making an impact so Ukraine isn't slowing down in continuing to develop those capabilities further. I think this is a very bad aspect of the technology to be clear but stuff like fully autonomous drones is the future of warfare and this seems to be an early taste of that.

To break my main point down fully the huge implications of drones in warfare starting to become autonomous combined with AI now being advanced enough to have a notable impact on things like coding, mathematics, cybersecurity and so on is why the technology is already a big deal. The main AI optimism argument is it'll continue improving over time but an increasing amount of big stuff is now real so aren't future hypotheticals any more.

Last edited by Norion - on 25 May 2026