🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Current agents only do 30% of complex real company tasks in this paper. Though note benchmarks are a floor, not a ceiling, if: 1) More recent models show improvement in the benchmark, suggesting future models may do it 2) Better prompting/tools would make the AI perform better.

So, um, is it really the case that the headline AI agents COMPLETED 30%-48% of real-real world professional office tasks? That’s, um, *checks notes* a lot, right?

My sample is biased (obviously), but ChatGPT use (and, rapidly others as well) is beginning to feel ambient. The oracles are running amok amongst us, well maybe not quite amok, but quickly thriving towards ends unknown for sure.

👀👀

Any attempts at counting or estimating anything in this direction anyone knows about?

👀👀

👀👀

Ethan Mollick rohit We pushed the system prompt earlier today. Feel free to take a look! github.com/xai-org/grok-p…

This is a step in a good direction for xAI, transparency in their system prompt.

👀👀

So what to do? We say: focus on the handful of large AI developers truly at at the frontier. That's where the most distinctive potential risks of frontier AI development are most likely to arise, and where the need for transparency, evidence, and understanding are most pressing.

Ketan Ramakrishnan Dean W. Ball Carnegie Endowment I did not expect to agree with Dean W. Ball but yes, this is the path forward. We do need governance on frontier companies. "One of the chief tasks of a frontier AI regulatory regime—arguably the chief task, at least for now—is to put society in a position to reduce such

Keep an eye for these projects to be published! Deric and I were very impressed with the quality of these fellows and their work. There’s not nearly enough work being done to understand what components are needed for an AGI Social Contract. With this incredible team of fellows

Kudos to Anthropic for their work on transparency! This a big step in the right direction. Happy to see it! Let’s have the public conversation about this incredibly important piece of AI policy! “Frontier AI development needs greater transparency to ensure public safety and

Lots to like here. Kudos to Ketan Ramakrishnan and Dean W. Ball for their work here! Moving the discussion forward, in a sensible direction. Love to see it.

👀👀

How could you not be compelled to watch.

I'm giving a public talk Tuesday July 8th, 7:30 pm at Mox in SF. Title: "Can goodness compete?". It's about long-term equilibrium outcomes post-AGI. More info at link in thread.

I know I’m new to DC, because when I look around, I see opportunity everywhere. And, even more than opportunity, there’s maybe even a hint of a touch of momentum.

Always nice to see rohit making an obvious observation that I completely missed. Well put.