It's the circle of life.

Sturgeon's Law says that 90% of everything is crap. But those are, to quote the movie, rookie numbers. The great AI productivity gains have enabled us to pump those numbers up to 99%. What a time to be alive!

Sam Altman is doing a pretty good impression of it...

*When* the AI boom goes bust. Cannot possibly still be "if" at this point.

And oh my god it's stupid.

look if reversing the tachyon polarity hasn't fixed it you've got bigger problems than chatgpt

What a remarkable invention.

I mean you look at all those speakers with screens they kept building in the Olden Days between the invention of the radio and the invention of the television and think to yourself "If only we had AI back then, they could have saved a whole lot of wasted resources by just not building all those useless screens".

Thank god for Sam Altman.

Nobody knows... :-(

Woooo AI games here we come

I don't work for T-Mobile but I do work in the Core Network design team of another mobile carrier in another country. Amongst those applications were undoubtedly a bunch of systems that are literally the mobile network itself, Virtual Network Functions (VNF) - things like HSS, MME, PGW etc. You can't just pick those up and flick them somewhere else on a whim. With the complexity of a modern mobile network 4 years would be on the short side for a program of work to move them. Especially since the move would realistically mean a shift to containerised versions instead of virtualised, which are an order of magnitude more complex to deploy.

Source: our company is in the throes of replacing large chunks of the core network with containerised functions (on RedHat OpenShift, which is the big telco vendors standard platform these days), currently in year three and have only completed HSS, with IMS ready to go live soon, OCS in the early stages, PCRF, messaging and the entire EPC yet to kick off in 2027-2028.

I thought the whole point was that it was so smart it wouldn't matter which way I held it. Why, a goldfish could design a new car with magical AI!

The toilet seat is still freezing in the morning even after Uranus has been...

If we're in the fun hype phase now, I shudder to think...

Anyway... here's Wonderwall!

LLMs hallucinate 100% of the time. More often than not the hallucination is in the shape of the correct answer. Sometimes it isn't. But you can't stop LLMs from hallucinating ... because everything is a hallucination.

All you can do is keep trying to find more edge cases and writing rules against them.

Which is basically computer programming again but with 1000x the resource cost and none of the determinism.