micro1

What changes when AI moves from answering questions to completing work?

On November 11, @daveholtz, Head of AGI Economics Research at @OpenAI, joins the micro1 forum to explore the shift from chat to agents.

We’ll discuss how companies use AI today, where agents are changing workflows, and the data infrastructure and evaluations needed to put them to work.

Featuring:
• @daveholtz, Head of AGI Economics Research at @OpenAI
• @Exp_Mark, PhD, Chief Economist at micro1, Host
• Liu Zhang, PhD, Economist & Member of Technical Staff at micro1

November 11th · 10am PT / 1pm ET

Register through the link in the comments

Meet Kai, micro1's AI partnership specialist.

His job is to get companies paid for their data, fast.

In his first month alone, Kai has:
- Closed $3M+ in data licensing deals
- Conducted 836 partnership calls
- Saved nearly 400 hours of meetings

Calls with Kai take just 7–10 minutes, and eligible companies can go from that conversation to a six- or seven-figure licensing offer in as little as 36 hours.

These partnerships help us build RL environments where frontier AI learns from the work happening inside real companies.

Learn more in comments below.

We gave 10 frontier AI models control of a surgical robot in a simulated operating room.

Across 800 trials, only two models completed the task at all: Gemini 4 Argon nine times, and GPT-6 Astra twice.

Models claimed success 213 times. Every claim was wrong. In 94 of them, the robot never touched the object.

Introducing Dual-Arm Sim, our new robotics benchmark evaluating how well frontier models can control embodiments in simulated environments. More info in thread around what we tested and found.

Can we trust the judgment of personal AI agents?

We took another look at our PersonalAgentBench results to see where personal agents were falling short.

One agent flagged an overdue payment that had already been paid. Another included private account details that weren’t needed. A third said two email drafts were ready, but neither existed in Gmail.

In each case, the agent had access to the right information or account. The trouble came in deciding what mattered, what to share, and whether the work was actually done.

Our latest update digs into these failures and why giving agents permission is only part of learning to trust them. Learn more below.

micro1.ai/blog/where-per…

Government reporting demands more than factual accuracy.

Introducing MissionReport, our latest benchmark testing whether frontier models can turn working files into reports commanders and senior leaders can rely on.