AI agents are changing not just how much work we can automate, but what work itself feels like.
I wrote about trying to put agents everywhere, the strange problem of waiting for them, and the endless optimization problems that follow. https://t.co/FF64FvtD7F
This is a nice publicity stunt, and is an exciting headline, but the harsh truth is this just a waste of Anthropic’s compute.
Let me explain.
1. Solving math proofs wont help you RSI faster
2. The amount of compute spent to solve this could have been used for RSI workloads
3. The RL required for this type of task is large and you could have spent it on RSI
just tried out the t3 (simulator) device hub and the remote part of that, and I have to say that this exceeds my expectations.
that is literally the cooles thing for mobile dev one could image! go try it, it's worth it.
Look at Colette Kress' facial expressions when she has to address that the board won't allow Jensen Huang coming back.
One could expect a bit of empathy about the man who built NVIDIA from nothing, but what do you see in her expression?
It's as hostile as it gets. Look at the instant she says that 'Jensen Huang resigned from his position at NVIDIA':
Introducing Gemini 4 Neon! Pushing the frontier across domains and punching above its model class in a meaningful way. To me, Neon feels like the right balance of intelligence, speed, and usability. Available now across @GoogleAIStudio, the API, and more!