Ineffable Intelligence

Ineffable Intelligence is a London-based artificial intelligence research lab founded by David Silver, the DeepMind researcher who led the work on AlphaGo and AlphaZero. The company was incorporated in November 2025^[10] and emerged from stealth on 27 April 2026 with a roughly $1.1 billion seed round at a $5.1 billion valuation, which it and its backers described as the largest seed round in European history.^[1][2][3] Its stated goal is to build a "superlearner," an AI system that reaches superintelligence by learning from its own experience through reinforcement learning, rather than by training on text and other data produced by humans.^[1][2]

The launch placed Silver among a group of senior researchers who left large technology and AI companies in 2025 and 2026 to start their own AI labs, often with very large early rounds.^[2] What sets Ineffable apart from most of that cohort is the technical bet. While the dominant approach in the industry is to scale up large language models trained on human data, Silver argues that getting to superintelligence will require systems that discard human knowledge and learn to achieve goals from first principles.^[4] Since the launch the company has announced infrastructure partnerships with NVIDIA (May 2026) and Google Cloud (June 2026), named six further co-founders, most of them former DeepMind colleagues of Silver's (September 2026), and been identified by NVIDIA as the first customer to access Google Cloud's A5X instances built on NVIDIA Vera Rubin systems (October 2026).^{[8][14][9][16]} It had not released a product or model as of October 2026.^[13][26]

Founding and David Silver's background

Ineffable Intelligence Ltd was incorporated in England on 19 November 2025.^[10] According to Companies House records reported by Fortune, the company was set up by George "Geordie" Rose, a Canadian entrepreneur, who was its first director and initially held 75% or more of its shares.^[9][11][12] Silver was appointed a director on 16 January 2026, the same date on which the registry lists him as the company's person with significant control in place of Rose.^[11][12] Rose resigned as a director on 30 June 2026.^[11] Geordie Rose describes himself as a "founding advisor"; a company spokesperson told Fortune that Rose is a friend of Silver's who helped him get the company up and running before moving into an advisory role, and that Silver does not consider him a co-founder.^[9] As of October 2026, the registry listed Silver as the only person with significant control, holding more than 25% but not more than 50% of the shares and voting rights.^[12]

Silver left Google DeepMind in January 2026, when the lab informed staff of his departure.^[4] He had been on sabbatical for several months beforehand and never formally returned to his role.^[4] Fortune first reported the move on 30 January 2026, citing a person with direct knowledge of his plans; a Google DeepMind spokesperson confirmed the departure, and the person said the new company was recruiting researchers and seeking venture funding.^[4] Silver later published a note on the company's blog, dated 15 January 2026, which begins: "I wrote the following note for myself when deciding to set up Ineffable Intelligence."^[1][13] Sequoia Capital's company page gives the founding year as 2026,^[18] while news reports at the launch described the lab as founded in late 2025.^[2][3]

Silver, born in 1976, is one of the most decorated researchers in reinforcement learning. He studied at Christ's College, Cambridge, then completed a PhD at the University of Alberta under Richard Sutton, where he co-introduced the algorithms used in the first master-level 9x9 computer Go programs.^[5] He received a Royal Society University Research Fellowship in 2011 and became a lecturer at University College London, where he is a professor.^[5] He consulted for DeepMind from 2010 before joining full-time in 2013.^[5] At DeepMind he led the reinforcement learning team and the projects behind AlphaGo, which beat Go champion Lee Sedol in 2016, and AlphaZero, which reached superhuman play in Go, chess and shogi purely through self-play.^[5][4] He was also a key member of the teams behind AlphaStar in StarCraft II and MuZero, which learned to master games without being told their rules, and he worked with the team that built AlphaProof, which solved International Mathematical Olympiad problems.^[4] He won the 2019 ACM Prize in Computing and was elected a Fellow of the Royal Society in 2021 for his work on Deep Q-Networks and AlphaGo.^[5]

Silver has known DeepMind co-founder Demis Hassabis since university, and Fortune describes him as one of the company's first employees when it was established in 2010.^[4] According to a person familiar with his thinking, he told friends he wanted to get back to "the awe and wonder of solving the hardest problems in AI."^[4] In his founding note he called the mission his "life's work" and acknowledged "a significant risk of failure for a chance of spectacular success."^[1][13]

In an interview on Sequoia Capital's Training Data podcast, published in October 2026, Silver, asked why he had not pursued the work inside DeepMind, said that the path he believes in "doesn't actually go through LLM-based products as an intermediate point," and that it deserved to be "more than just the corner of an organization that's focused on LLM-based products."^[17] He said the name refers to the idea that the knowledge a superintelligence discovers could be "so profound that we can't even express it in human language," which matches one of the beliefs listed on the company's website.^[17][13] He also gave a joking "secret second reason": that an approach built entirely on learning from experience "can't possibly f up," hence "in-eff-able."^[17]

Research thesis: learning from experience

Ineffable's technical premise comes directly from a 2025 position paper, "Welcome to the Era of Experience," that Silver co-wrote with Sutton,^[6] his former doctoral advisor and a co-recipient of the 2024 Turing Award. The paper argues that AI is moving past the era of training on static human data and into one where agents improve mainly by acting in the world, observing the results, and adjusting. Silver and Sutton sketch four characteristics of this shift: agents that inhabit long streams of experience rather than short snippets of interaction, actions and observations grounded in the environment rather than in human dialogue alone, rewards grounded in the agent's experience of the environment instead of in human prejudgement, and planning or reasoning about experience rather than reasoning solely in human terms.^[6]

Applied as a company mission, this becomes the "superlearner." The company's website describes a system that "discovers all knowledge from its own experience, from elementary motor skills through to profound intellectual breakthroughs," driven by reinforcement learning and "without relying on human data," in the way AlphaZero learned Go and chess by playing against itself rather than studying human games.^[13][1] Fortune summarized the view of reinforcement learning purists such as Silver in stark terms: to reach superintelligence, an AI will not just have to surpass human knowledge but discard it and learn to achieve goals from scratch.^[4] Silver told Wired, as quoted by Decrypt, that "human data is like a kind of fossil fuel that has provided an amazing shortcut," while systems that learn for themselves are "a renewable fuel."^[24] The company's website compares the potential payoff to a scientific breakthrough "of comparable magnitude to Darwin."^[13][1]

Most of the capital and attention in AI flows to ever-larger language models, in which reinforcement learning is mostly used as a post-training step, often driven by human feedback, rather than as the engine of capability itself.^[4] Ineffable proposes to make experience-driven reinforcement learning the engine. Sequoia, one of its lead investors, wrote that Ineffable is scaling reinforcement learning "from a clean base: no pre-training, no human data to shortcut the system," and called the bet "genuinely contrarian," with an uncertain timeline to superintelligence.^[19] Whether that scales the way transformer pretraining did is the open question on which the company rests.

Superlearners and crucibles

In the Training Data interview, Silver defined a superlearner along two dimensions. In depth, it should start "from not knowing anything at all" and go on to discover things far beyond what humans know; in breadth, it should be possible "to drop it into any stream of experience," whether simulated, digital or physical.^[17] He said current language-model systems that make discoveries do not meet this definition because they still need "a lot of human-in-the-loop effort."^[17]

Ineffable calls the simulated environments in which its agents learn "crucibles." Silver described them as worlds with "the capacity for a lot of emergent complexity" rather than sets of handcrafted puzzles, and said the lab is "not breadcrumb-trailing" the objective: it gives agents simple objectives and expects complex behavior to emerge, by analogy with evolution optimizing only for fitness.^[17] Co-founder Lasse Espeholt said the lab wants a range of crucibles, from simple ones that run quickly to more expensive environments, noting that fully realistic physics "would be very computationally expensive."^[17] Silver argued that if superintelligence can be built in simulation it will happen "ten times or a hundred times faster," because simulation can be parallelized and reproduced, and that simulation is "the safest place for superintelligence to be built."^[17] The company's job listings include a "Simulated Universes" engineering role for people with experience of large-scale simulations, physics engines or distributed multiplayer game worlds.^[32]

The lab's reinforcement learning stack builds on Espeholt's earlier work. At DeepMind and Google Brain he was the lead author of IMPALA, a scalable distributed actor-learner architecture, and SEED RL, which moved inference onto accelerators; he said Ineffable uses implementations of those algorithms and their derivatives.^[30][31][17]

Funding and investors

Ineffable raised what it reported as approximately $1.1 billion (about €937 million) in a seed round announced on 27 April 2026, at a $5.1 billion valuation; EU-Startups described it as a post-money valuation.^[1][2][7] The round was co-led by the U.S. venture firms Sequoia Capital and Lightspeed Venture Partners, with Alfred Lin and Sonya Huang representing Sequoia and Ravi Mhatre and Raviraj Jain representing Lightspeed.^[1][2][20] The company's announcement, as reported by EU-Startups and Tech.eu, listed participation from NVIDIA, DST Global, Index Ventures, Google, Flying Fish Ventures, EQT Ventures, Evantic Capital, the UK Wellcome Trust, BOND Capital, the British Business Bank, the U.K. government's Sovereign AI Fund, and strategic angels.^[7][20] The British Business Bank invested $20 million, and the Sovereign AI Fund co-invested an undisclosed additional amount.^[21][23] Press accounts noted that the company had earlier targeted around $1 billion at a $4 billion pre-money valuation before the round grew.^[3] Companies House records show that Alfred Lin and Ravi Mhatre were appointed directors of the company on 17 April 2026.^[11]

The company and several outlets called it the largest seed round ever raised in Europe.^[1][2][3] It came a month after AMI Labs, co-founded by Yann LeCun, raised $1.03 billion at what TechCrunch described as a $3.5 billion pre-money valuation.^[1] With no product yet, EU-Startups wrote that the funding was "less about a near-term product roadmap and more about the credibility of Silver's research thesis."^[7] Danny Rimer of Index Ventures called the investment "the ultimate people bet."^[34]

Silver has said he intends to give away the money he personally makes from his Ineffable equity. He told Wired that "any money that I make from Ineffable will go to high-impact charities that save as many lives as possible."^[1] Tech.eu reported that he had committed to give away 100% of the money he makes from his Ineffable equity through Founders Pledge, the biggest pledge in that organization's history.^[20] Founders Pledge lists him as a member.^[33]

NVIDIA partnership

On 13 May 2026, about two weeks after the funding announcement, NVIDIA disclosed an engineering-level collaboration with Ineffable to co-design infrastructure for large-scale reinforcement learning.^[8][25] NVIDIA framed the collaboration as a way to work out the hardware and software needed as AI shifts away from human data and toward systems that learn through simulation and experience.^[8] It said the work was starting on NVIDIA's Grace Blackwell systems and "will be among the first to explore" the upcoming Vera Rubin platform.^[8] NVIDIA chief executive Jensen Huang said the company was "thrilled to partner with Ineffable Intelligence to codesign the infrastructure for large-scale reinforcement learning as they push the frontier of AI and pioneer a new generation of intelligent systems."^[8] Silver said researchers had "largely solved the easier problem of AI," building systems that know what humans already know, and now needed to build "systems that discover new knowledge for themselves."^[8] NVIDIA explained that, unlike pretraining on a fixed dataset, reinforcement learning workloads generate their data on the fly in tight act, observe, score and update loops, which puts pressure on interconnect, memory bandwidth and serving in ways that pretraining does not.^[8]

Google Cloud partnership

On 16 June 2026, at Google Cloud Summit London, Google Cloud and Ineffable announced a strategic partnership under which Ineffable selected Google Cloud as its preferred cloud partner.^[14][15] The announcement said Ineffable would deploy "one of the largest clusters of A5X," Google Cloud instances powered by NVIDIA Vera Rubin NVL72 systems.^[14] Silver said the company had "evaluated the space and chose Google Cloud as the best fit for our reinforcement learning infrastructure," and Google Cloud chief executive Thomas Kurian said Ineffable was using Google's "full-stack AI Hypercomputer."^[14] The announcement did not disclose financial terms.^[14]

On 21 October 2026, during NVIDIA GTC Berlin 2026, NVIDIA said A5X instances would become generally available the following month and that Ineffable had been the first customer to access them.^[16] Silver said Vera Rubin's high-bandwidth connections, between the Vera CPU and Rubin GPUs and between GPUs within the rack, allowed the lab to "push the scale and ambition of our reinforcement learning research further."^[16] In the Training Data interview, Espeholt said the company was the first customer of the largest cluster on Google Cloud, due to start about a month after the recording, and that NVIDIA GPUs suited the lab because the same chips could run simulation, training and inference.^[17]

Team and location

Ineffable is based in London, which makes it one of the larger AI ventures attempting to build a frontier lab in Europe rather than in the San Francisco Bay Area.^[2][3] Reporting at the time of the funding said several former DeepMind staffers were set to join the executive team.^[1] On 7 September 2026, Fortune reported that the company had named six "cofounders" alongside Silver; Ineffable confirmed the hires but declined to comment further.^[9] Four had worked with Silver at Google DeepMind, and three of them (Apps, Czarnecki and Oh) were authors, with Silver, of DeepMind's 2019 AlphaStar paper in Nature.^[9][29] Alexandre Laterre had already described "my exceptional co-founders" in a post on launch day.^[27]

Sources: Fortune,^[9] CNBC,^[2] PR Newswire,^[14] Nature,^[28][29] arXiv,^[30][31] Sequoia Capital.^[17]

Fortune reported that Oh, as a DeepMind researcher, had been one of the first to use AI to discover new reinforcement learning algorithms.^[9] The October 2025 Nature paper, a project he co-led with Silver as senior author and on which he was the first-listed of four joint first authors, meta-learned a reinforcement learning rule from the experience of a population of agents; the discovered rule surpassed all existing rules on the Atari benchmark, and the authors suggested that the algorithms required for advanced AI "may soon be automatically discovered from the experiences of agents."^[28] Fortune, citing GeekWire, reported that Gorham had approached Silver in early 2024 to learn about new approaches to reinforcement learning; Flying Fish went on to write the first check for the company, and Gorham helped secure its Google Cloud partnership.^[9]

In the Training Data interview published in October 2026, Silver said the company had "hired 20 people in total" and was intentionally keeping the team small; Espeholt said the small headcount kept the amount of compute per person high.^[17] Silver said four of the co-founders came from DeepMind's AlphaStar team and "the other three founders all came from very successful startups." He said the lab hires for talent, kindness, alignment with the mission and fit with its needs.^[17] Its open roles are on-site positions in London.^[32]

The London base, combined with backing from the U.K. Sovereign AI Fund and the British Business Bank, gave the launch a national dimension. Science and Technology Secretary Liz Kendall said the investment underlined the government's "determination to ensure that the UK isn't just an AI taker but an AI maker," and called it Sovereign AI's second direct investment.^[2][22] The Telegraph reported that the Sovereign AI unit was designed to tie promising AI start-ups to Britain, potentially avoiding the fate of DeepMind, which was sold to Google in 2014.^[23]

Views on safety and alignment

Ineffable's published beliefs include that superintelligence "can and must be built in a way that is beneficial to humanity" and "can be built within years, not decades or centuries."^[13] In the Training Data interview, Silver said that AI alignment approaches based on the language agents use, and on priors from human data, had been effective, but that "we're starting to see that approach reaching its limits," because as agents are trained more they "start to learn to say one thing and act in a different way."^[17] He argued that monitoring an agent's chain of thought in human language will become harder, and suggested alternatives such as "shadow actions," where an agent's intended action is observed without being carried out, and using other superintelligent systems as monitors.^[17]

Silver said the company's most important task was to understand "the science of superintelligence" in simulation, including what it takes for a society of superintelligences to interact peacefully with each other and with "lesser intelligences," and to share the answers publicly. He said Ineffable had committed to gate anything it does in the real world on sufficient evidence from that science.^[17] He said he supported pacing the frontier of AI in the real world while moving "as fast as possible in simulation."^[17]

On timing, Silver predicted "extremely strong evidence that we're on the right track within a year" and said he expected "a true superintelligence" within three years, shortened from his earlier estimate of three to five years because of what coding agents can do to speed up research.^[17]

Timeline

Context within the 2026 wave of frontier labs

Ineffable arrived during a stretch in which several senior researchers left the big labs to start their own superintelligence companies with very large rounds. In reporting Silver's departure, Fortune pointed to Safe Superintelligence, founded by former OpenAI chief scientist Ilya Sutskever, and to Reflection AI, founded by some of Silver's former colleagues on AlphaGo, AlphaZero and MuZero.^[4] Fortune later described Ineffable as one of the so-called "neolabs," a group that it said also includes Thinking Machines Lab, co-founded by former OpenAI chief technology officer Mira Murati, Safe Superintelligence, Sakana AI, Recursive Superintelligence, Core Automation and Adaption Labs.^[9] Ineffable fits that mold on the financing side. On the research side it is an outlier, betting on experience-driven reinforcement learning rather than scaling language models.

Recursive Superintelligence, co-founded by former DeepMind principal scientist Tim Rocktäschel and also incorporated in the U.K., was raising money at the same time as Ineffable;^[1] CNBC reported in May 2026 that it had raised $650 million.^[25] The two are separate companies.

Whether the superlearner thesis pays off will take years to judge. The Next Web noted that reinforcement learning's best-known wins came in domains with explicit rules and a clear signal for winning, such as Go, chess, StarCraft II and Atari, and that the method had not yet been shown to hold where goals are open-ended and rewards ambiguous.^[26] As of October 2026 the company's website listed no models or research papers, only its mission, beliefs, careers page and Silver's founding note.^[13] What is already clear is that a large pool of investors was willing to fund the attempt at a scale normally reserved for companies with revenue, on the strength of a research idea and the person behind it.

See also

• List of neolabs
• Neolab
• David Silver
• Richard S. Sutton
• Geordie Rose
• Reflection AI
• Recursive Superintelligence
• NVIDIA GTC Berlin 2026

References

1. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14Anna Heim, "DeepMind's David Silver just raised $1.1B to build an AI that learns without human data," TechCrunch, 27 April 2026. techcrunch.com/...i-that-learns-without-human-data
2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11Kai Nicol-Schwarz, "Former Google DeepMind researcher's AI startup raises record $1.1 billion seed funding to pursue superintelligence," CNBC, 27 April 2026. cnbc.com/...ence-record-seed-funding-nvidia-google
3. ^1 ^2 ^3 ^4 ^5 ^6"Ineffable Intelligence closes Europe's largest seed at $5.1B valuation from Sequoia, Lightspeed," Tech Funding News, 28 April 2026. techfundingnews.com/...e-1-1bn-seed-europe-largest
4. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12Jeremy Kahn, "Exclusive: Google DeepMind researcher David Silver leaves to launch his own AI startup," Fortune, 30 January 2026. fortune.com/...d-ai-startup-ineffable-intelligence
5. ^1 ^2 ^3 ^4 ^5"David Silver (computer scientist)," Wikipedia. en.wikipedia.org/...id_Silver_(computer_scientist)
6. ^1 ^2David Silver and Richard S. Sutton, "Welcome to the Era of Experience," preprint of a chapter in Designing an Intelligence (MIT Press), DeepMind, 2025. storage.googleapis.com/...20Experience%20Paper.pdf
7. ^1 ^2 ^3David Cendon Garcia, "Ineffable Intelligence lands historic $1.1 billion Seed round at $5.1 billion valuation," EU-Startups, 27 April 2026. eu-startups.com/...-round-at-5-1-billion-valuation
8. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8"NVIDIA, Ineffable Intelligence Team Up to Build the Future of Reinforcement Learning Infrastructure," NVIDIA Blog, 13 May 2026. blogs.nvidia.com/...cement-learning-infrastructure
9. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10Jeremy Kahn, "Exclusive: Ineffable Intelligence adds six 'cofounders,' hiring veterans from Google DeepMind, InstaDeep and venture firm Flying Fish," Fortune, 7 September 2026. fortune.com/...ogle-deepmind-instadeep-flying-fish
10. ^1 ^2 ^3Companies House, "INEFFABLE INTELLIGENCE LTD (16865241): company overview," accessed October 2026. find-and-update.company-information.service.gov.uk/...16865241
11. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8Companies House, "INEFFABLE INTELLIGENCE LTD (16865241): officers," accessed October 2026. find-and-update.company-information.service.gov.uk/...officers
12. ^1 ^2 ^3 ^4Companies House, "INEFFABLE INTELLIGENCE LTD (16865241): persons with significant control," accessed October 2026. find-and-update.company-information.service.gov.uk/...-control
13. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9Ineffable Intelligence, company website (Mission, Beliefs, and David Silver, "A Note from Dave," 15 January 2026), accessed October 2026. ineffable.ai
14. ^1 ^2 ^3 ^4 ^5 ^6 ^7"Ineffable Intelligence Selects Google Cloud To Power Its Superintelligence Mission," PR Newswire, 16 June 2026. prnewswire.com/...erintelligence-mission-302801139
15. ^Kirstie Pickering, "Ineffable Intelligence and Google Cloud partner on new frontier AI lab," UKTN, 16 June 2026. uktech.news/...ner-on-new-frontier-ai-lab-20260616
16. ^1 ^2 ^3 ^4Uttara Kumar, "Google Cloud and NVIDIA Announce General Availability of NVIDIA Vera Rubin Infrastructure," NVIDIA Blog, 21 October 2026. blogs.nvidia.com/...loud-vera-rubin-infrastructure
17. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14 ^15 ^16 ^17 ^18 ^19Sonya Huang (host), "David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience," Training Data podcast, Sequoia Capital, 13 October 2026. sequoiacap.com/...earner-for-the-age-of-experience
18. ^Sequoia Capital, "Ineffable Intelligence" company page, accessed October 2026. sequoiacap.com/...ineffable-intelligence
19. ^Sonya Huang and Alfred Lin, "Partnering with Ineffable Intelligence: A Superlearner for the Era of Experience," Sequoia Capital, 27 April 2026. sequoiacap.com/...earner-for-the-era-of-experience
20. ^1 ^2 ^3 ^4Cate Lawrence, "Ineffable Intelligence launches with record-breaking $1.1B Seed round," Tech.eu, 27 April 2026. tech.eu/...hes-with-record-breaking-11b-seed-round
21. ^"British Business Bank and Sovereign AI invest into AI superintelligence company Ineffable Intelligence," IFA Magazine (British Business Bank announcement), 27 April 2026. ifamagazine.com/...-company-ineffable-intelligence
22. ^Sovereign AI, "Sovereign AI invests in Ineffable Intelligence," accessed October 2026. sovereignai.gov.uk/...ts-in-ineffable-intelligence
23. ^1 ^2James Titcomb, "UK backs 'self-learning' AI start-up in effort to catch up," The Telegraph (via Yahoo Finance), 27 April 2026. finance.yahoo.com/...ks-self-learning-ai-193621676
24. ^Jason Nelson, "Google DeepMind Veteran Raises $1.1 Billion to Build AI That Isn't Trained With Human Data," Decrypt, 28 April 2026. decrypt.co/...1-billion-ai-isnt-trained-human-data
25. ^1 ^2Kai Nicol-Schwarz, "Nvidia's Jensen Huang bets on this British startup to build 'next frontier' of AI," CNBC, 13 May 2026. cnbc.com/...rtup-partners-nvidia-superintelligence
26. ^1 ^2Ana Maria Constantin, "Ineffable Intelligence names six cofounders, Fortune reports," The Next Web, 7 September 2026. thenextweb.com/...s-deepmind-instadeep-flying-fish
27. ^Alex Laterre, post on X, 27 April 2026. x.com/...2048785535376773526
28. ^1 ^2Junhyuk Oh, Gregory Farquhar, Iurii Kemaev, Dan A. Calian, Matteo Hessel, Luisa Zintgraf, Satinder Singh, Hado van Hasselt and David Silver, "Discovering state-of-the-art reinforcement learning algorithms," Nature 648, 312-319 (published online 22 October 2025). nature.com/...s41586-025-09761-x
29. ^1 ^2Oriol Vinyals et al., "Grandmaster level in StarCraft II using multi-agent reinforcement learning," Nature 575, 350-354 (2019). nature.com/...s41586-019-1724-z
30. ^1 ^2Lasse Espeholt et al., "IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures," arXiv:1802.01561, 2018. arxiv.org/...1802.01561
31. ^1 ^2Lasse Espeholt et al., "SEED RL: Scalable and Efficient Deep-RL with Accelerated Central Inference," arXiv:1910.06591, 2019. arxiv.org/...1910.06591
32. ^1 ^2Ineffable Intelligence, careers page (job listings), accessed October 2026. jobs.ashbyhq.com/ineffable
33. ^Founders Pledge, "Dave Silver" member profile, accessed October 2026. founderspledge.com/...3692
34. ^Danny Rimer, "The Superlearner: Investing in David Silver and Ineffable Intelligence," Index Ventures, 27 April 2026. indexventures.com/...er-and-ineffable-intelligence