What Happened

Fortune reported on Monday, September 7, 2026, that Ineffable Intelligence has awarded cofounder titles to six people who joined the company this summer — none of whom founded it. The appointments became visible when all six updated their LinkedIn profiles; Ineffable confirmed the hires but declined to comment further.

The four from Google DeepMind are Chris Apps, Wojciech Czarnecki, Lasse Espeholt, and Junhyuk Oh. Alexandre Laterre arrives from InstaDeep, the London company BioNTech acquired in 2023, where he led research. Heather Gorham was a partner at Flying Fish, the Seattle venture firm that wrote Ineffable's first cheque.

The corporate history is unusual. Companies House records show a Canadian entrepreneur, George Rose, incorporated Ineffable in November 2025. David Silver became a director in January 2026. Rose ceased to be a person with significant control in February and resigned as a director in July. A company spokesperson said Rose helped Silver get the company running before moving to an advisory role, and that Silver does not consider him a cofounder.

So the man who registered the company is a founding adviser, and six people who arrived months later are cofounders. The title functions as compensation — equity and status move researchers between labs, and a founding credit costs less than cash.

Why It Matters

The hires are not a general-purpose research team. They are a specific team from a specific era of DeepMind — the era before the lab reorganized around shipping Gemini.

Apps, Czarnecki, and Oh all worked with Silver on AlphaStar, the system that reached grandmaster level at StarCraft II in 2019. Espeholt built weather models on Google's MetNet. This is Silver's thesis made into an org chart: large language models cannot reach superintelligence because they only recombine what humans have already written. His answer is reinforcement learning — agents that learn by acting inside simulated worlds and discovering things no human wrote down.

The hire that says the most is Oh, who will run reinforcement learning at Ineffable. In October 2025, he was joint first author on a Nature paper Silver also signed. The team used machine learning to discover a new reinforcement learning rule rather than designing one by hand. They called it DiscoRL, and it beat every existing rule on the Atari benchmark. The paper's closing line suggests that agents may soon find the algorithms for advanced AI themselves, out of their own experience — researchers would stop writing them by hand.

Every frontier lab is now chasing some version of that idea, usually filed under recursive self-improvement. Ineffable has hired one of the few people with a published result in it. The thesis is no longer eccentric — it is crowded. Richard Sutton left Keen in July to start his own RL lab. Recursive Superintelligence raised $650M in May. Ineffable raised $1.1B in April.

Meanwhile, DeepMind's talent drain continues. Per Zeki Data, DeepMind's share of research and advanced-engineering hires across EMEA fell from 49% in 2022-2023 to 18.6% in 2025-2026. The ratio of arrivals to departures fell from roughly 12:1 in mid-2023 to about 2:1 this quarter. Jeff Dean, Oriol Vinyals, and Quoc Le reportedly left on the same afternoon in August.

Who Is Affected

AI startup founders — particularly those building in reinforcement learning — now face a competitor with $1.1B in capital, a clear research mandate, and a proven recruitment strategy using cofounder equity. Expect bidding pressure on RL talent to intensify.

Enterprise AI buyers — Ineffable has no product, no API, and no announced timeline. This is a multi-year research bet. Nothing here changes vendor selection decisions in 2026.

Google DeepMind — The lab is still a net hirer globally, but the trend line is clear. Silver is systematically extracting the specific researchers who built AlphaStar, and the Zeki data suggests the broader talent moat is narrowing.

Strategic Implications

For AI startup founders: Cofounder titles as compensation work when you have a compelling thesis and a famous principal investigator. Silver is using them to reconstruct a specific team. If you're competing for the same talent pool, you need either a comparably compelling research mission or significantly more cash — and Ineffable has $1.1B of that too.

For developers/operators building with AI APIs: Nothing here changes your stack. Ineffable is not shipping anything you can call. The DiscoRL paper is worth reading if you work in RL, but it is not a product. Monitor whether their research produces techniques that filter into open-source frameworks, but don't hold your breath.

For non-technical business owners evaluating AI tools: A $5B valuation on a company with no product tells you that investors are betting on a 5-10 year research horizon, not a near-term commercial outcome. This is not a vendor you'll be evaluating this year or next.

What to Watch Next

Watch for whether Ineffable announces a compute partnership beyond the Google Cloud deal Gorham reportedly secured in July, and whether Czarnecki's role resolves cleanly — Fortune's August reporting placed him as CTO at a company called Fundamental, which may indicate a recent move or a reporting discrepancy. Also monitor whether DeepMind responds to the accelerating talent drain with retention packages or structural changes.

Frequently Asked Questions

Q: What is Ineffable Intelligence and who is behind it?

A: Ineffable Intelligence is a London-based AI startup founded around David Silver, formerly of Google DeepMind. It raised $1.1B in seed funding in April 2026 at a $5B valuation. The company is building AI based on reinforcement learning rather than large language models, with no product yet released.

Q: Why did Ineffable Intelligence name six cofounders who didn't found the company?

A: The cofounder titles function as compensation and recruitment leverage. In AI research, equity and status move talent between labs. By offering cofounder titles to senior researchers from DeepMind and InstaDeep, Silver is assembling a specific team — the AlphaStar-era RL group — without paying Big Tech-level cash salaries.

Q: What is DiscoRL and why does it matter?

A: DiscoRL is a reinforcement learning rule discovered by machine learning rather than designed by hand, published in Nature in October 2025 by Junhyuk Oh and David Silver. It beat every existing RL rule on the Atari benchmark. The paper suggests AI agents may soon discover their own learning algorithms — a form of recursive self-improvement that every frontier lab is now pursuing.