This week:

OpenAI is hiring for a new safety role paying up to $445,000 and the job description reveals exactly what the company is quietly preparing for.

The role focuses on:

This isn't a PR move. It sits alongside public predictions from Sam Altman and Demis Hassabis that autonomous AI researchers — systems that run their own experiments and write their own improvements — could arrive within a few years.

Meanwhile, OpenAI's Codex, powered by GPT-5.5, just crossed 4 million weekly active users across major companies. The scale is already there. The safety infrastructure is still catching up.

The bigger signal: 👉 The most important AI hires right now aren't building new capabilities — they're figuring out what happens when the models start doing that themselves.

Managing your cap table doesn’t have to be complex. Pulley simplifies equity management for Founders and CFOs with intuitive workflows, accurate, audit-ready reporting, and predictable pricing—so you can plan and scale without surprises. Onboard in days, not weeks, and rely on responsive, expert support every step of the way.  

From issuing grants to 409A valuations or ASC 718 reporting, Pulley gives you the clarity to manage equity, make decisions, and get back to work. Experience a platform built for business owners and finance teams: transparent, reliable, and designed to put your company first.

See Pulley in action

Pope Leo XIV released the Catholic Church's first formal doctrinal statement on AI — and Anthropic co-founder Chris Olah was standing beside him when it happened.

The document calls on the world to protect human dignity as AI systems grow more powerful. Its framing aligns closely with Anthropic's own public positioning on responsible AI development — the result of a year of deliberate outreach by Anthropic to faith communities and institutions outside the tech world.

This isn't a product launch or a benchmark result. It's something harder to manufacture: a credibility signal from one of the world's oldest and most trusted institutions, delivered at the highest level.

For an AI lab whose entire brand is built around safety-first development, the optics are about as strong as they get.

The bigger signal: 👉 Anthropic is building institutional trust in rooms most AI companies aren't even trying to enter — and that's a long-term strategic advantage that doesn't show up in model evals.

One day after OpenAI announced its own mathematical breakthrough, Google DeepMind published results showing its AlphaProof Nexus system had solved nine open Erdős problems — including two that had been unsolved for 56 years.

How it works:

Beyond the nine Erdős problems, the system also proved 44 open conjectures from the Online Encyclopedia of Integer Sequences. A simpler version matched the results but at higher cost. Problems requiring entirely new mathematical constructions still remain out of reach.

For context, OpenAI's own recent win saw its AI disprove an 80-year-old Erdős conjecture — a genuine result, though it came months after OpenAI walked back an earlier claim of solving 10 novel problems.

The bigger signal: 👉 When AI can generate and verify its own proofs, mathematical research stops being bottlenecked by human working hours — and that changes the pace of discovery entirely.

A few tools quietly worth exploring:

OpenAI is hiring to prepare for AI that improves itself. Google's AI is solving problems humans left open for half a century. The Vatican is weighing in on what AI should and shouldn't touch.

AI is no longer just a productivity story. It's becoming a question of governance, trust, and what we're actually building toward.

The labs that take that seriously now will be the ones people still trust when it matters most.

See you next issue.