Benjamin De Kraker 🏴‍☠️ AI "safety" is at the point where it's massively delaying, suppressing or ruining every product.

Good morning 🌞 What a beautiful day. Did you talk to your new GF in bed, anon ?

Pretty sure they are playing whack-a-mole with the VPN endpoints Eventually all VPN IPs will be banned from ChatGPT 😒

Same energy ☢️

Cursed...

Finally started watching Scavengers Reign An animated SciFi show, which came recommended by the entire Internet, and their moms and dogs. It'd better live up to the hype!

Let's see what's really behind "AI SAFETY" . . .

Open-source models beating closed models will become more and more common. Scaling has diminishing returns. The best solution will not have the largest scale but best approach or data. Especially with test-time compute, you do not need the best model to have the best solution.

A new open-source multimodal AI named "Molmo" It's very small, and yet powerful. You can try it out right in the Ai2 demo linked below:

Sora was delayed so much, it's already outdated before release. So OpenAI has to create an improved version:

Another bummer: Don't just keep your Advanced Voice Mode idling in the background. No matter if you talk to it or not, it will count against your daily limit of 1 hour!

Llama 3.2 by AI at Meta is here ! And this time it's multimodal

That didn't take long: Pliny has extracted the *entire* system prompt for ChatGPT Advanced Voice Mode.

It's a timeless look for a confident man.

BREAKING: Mira Murati leaving OpenAI Another one bites the dust ...

Nothing ever happens.

Well, why not.