Mind-blowing effort to push the narrative “LLMs must be regulated to be safe, only OpenAI should be allowed to license them, or else”. 

Or else the fragile monopoly of OpenAI will be commoditized away. Monopoly, not the jobs. Oops.

https://twitter.com/rubinovitz/status/1637651591191842816?s=20

Look around. Look at yourself. Take notes. Change yourself. Change everything else // Kyiv, Ukraine // Building https://blockless.dev/

They burned a ton of cash to gain an early-mover advantage, only to discover that it might not last long enough to cover these costs. 

Now they’re frantically trying to slow down the research that’s really open, spreading all kinds of FUD. How ironic.

Like, who in their right mind would push so many papers stating “our product is dangerous?” 🤯 Then maybe don’t sell it? 

No, they’re going after the competitors.

Wow, more of the same: TIL they’re cutting access to the base models (which are less restricted and can be used to build competing ones):

https://twitter.com/deepfates/status/1638212305887567873?s=20

I can’t think of any audience for this paper other than the lawmakers who need the scientifically-sounding reason to back the regulations. 

The paper is really devoid of details but full of trigger words. Same for the 60-page GPT4 model card.