@ansuz @joeyh @cwebber

Ah but even if you can use a specific seed and try to use this to call it a "compiler", your compiler here is the very specific sets of weights within that model, and any change breaks its determinism. I think there being one and exactly one possible implementation to get the specified set of outputs can count as an actual compiler.

@Luk @davidgerard

Isn't that just the message to users for everything now with AI? 😛

@realcarlallen

The fact that this is being allowed to happen is a huge red flag for the rot in our society and government. It is well worth calling out, because it is representative of so many other corrupting and fraudulent things that are happening. If we can't even successful govern this, what can we govern?

@dsandler

Hm yeah that looks perfectly normal.

@dsandler @foone

!?! WHAT DID YOU DO WRONG, DAN !?!

@cstross

"The longtime chief also criticized a lack of discipline in OpenAI’s business strategy"

Meaning they wouldn't be spending enough of that money on Nvidia's GPUs?

@pluralistic

I don't disagree with your overall position, however I don't think these location uses are good examples of it.

None of this is because of bugs in existing location code. From the very beginning, when location was only based on GPS, the contract was that you got back "here's a location and how accurate it is" giving you the center point and circle around it.

The examples here are all new uses that didn't properly make use of how location data was always provided.

@mage_of_dragons @cassidy @firefoxwebdevs

Well we were generally calling these things ML, until the AI hype train started. That isn't totally helpful, since LLMs are themselves another type of ML, but it would sure help to be able to talk about this stuff more specifically by not even more broadly calling it AI.