TO: Demis Hassabis, Google DeepMind, CEO, London and Mountain View, California and all the other Alphabet and Google enterprises.
FM: Bruce E. Camber
RE: Your articles especially within Harvard Business Review (May 2023), Time (Jan., 2023), arXiv (40): Memory-based Parameter Adaptation(20180; your homepage(s): AoA, CBMM, Google Scholar, Twitter, YouTube, and Wikipedia.
This page: 81018.com/hassabis/ Within: 81018.com/alphabetical/ and 81018.com/ai/
Second communication: 20 November 2025
Dear Demis,
I’m exploring a deliberately minimal geometric construction for early-universe modeling that may lend itself to AI-based simulation and falsification. The setup is intentionally elementary: start with a single Planck-scale sphere (using only Planck units and basic irrational-number geometry) and iteratively stack-pack forward through 202 discrete steps (“notations”), with the present lying near the final step.
The model has only a few tunable assumptions: one Planck-scale sphere per Planck time/length unit, continuous-to-discrete bridging through π’s symmetry, and polyhedral packing constraints that introduce early structural variation. Using these assumptions, a derived H₀ ≈ 71 km/s/Mpc functions purely as a computational consistency check.
My question is simply whether this construction is coherent and sparse enough to be a viable candidate for AI-driven exploration—i.e., something that could be rapidly simulated, stress-tested, or ruled out using DeepMind’s scientific frameworks.
If it seems potentially useful, I can send a short technical summary; if not, no need to respond.
With appreciation,
Bruce
First communication: June 21, 2023
@demishassabis It bodes well that you are trying to understand the fundamental nature of reality. If you are studying dimensionless constants and spheres and infinitesimal composites, you may want to take a quick glance at these early pages: https://81018.com/ai/ and https://81018.com/petition/ and https://81018.com/most-simple/
Editor’s note: Although in my first few weeks thinking about how our models might fit in with AI, I have read enough to know that they must, so your comments would be highly valued. Thank you. -Bruce
_________________
More to come.
Lionhead Studios.[22] Peter Molyneux, Bullfrog Productions; DH lead AI programmer on the 2001 “god” game Black & White.[18] 1998 Elixir Studios (London-based) Eidos Interactive, Vivendi Universal and Microsoft.[23] BAFTA, Republic: The Revolution and Evil Genius.[18]
Republic: The Revolution, a highly ambitious and unusual political simulation game,[24] was delayed due to its huge scope, which involved an AI simulation of the workings of an entire fictional country. The final game was reduced from its original vision and greeted with lukewarm reviews, receiving a Metacritic score of 62/100.[25]
Evil Genius, a tongue-in-cheek Bond villain simulator, fared much better with a score of 75/100.[26] [27][28]
Blaise Agüera y Arcas in 2014 at the Wired conference in London
Elixir Studios returned to academia to obtain his PhD in cognitive neuroscience from University College London (UCL) in 2009 supervised by Eleanor Maguire.[4] Inspiration in the human brain for new AI algorithms.[29]
He continued his neuroscience and artificial intelligence research as a visiting scientist jointly at Massachusetts Institute of Technology (MIT), in the lab of Tomaso Poggio, and Harvard University,[7] before earning a Henry Wellcome postdoctoral research fellowship to the Gatsby Computational Neuroscience Unit at UCL in 2009 working with Peter Dayan.[30]
Working in the field of imagination, memory and amnesia, he co-authored several influential papers published in Nature, Science, Neuron and PNAS. His very first academic work, published in PNAS,[31] was a landmark paper that showed systematically for the first time that patients with damage to their hippocampus, known to cause amnesia, were also unable to imagine themselves in new experiences. The finding established a link between the constructive process of imagination and the reconstructive process of episodic memory recall. Based on this work and a follow-up functional magnetic resonance imaging (fMRI) study,[32] Hassabis developed a new theoretical account of the episodic memory system identifying scene construction, the generation and online maintenance of a complex and coherent scene, as a key process underlying both memory recall and imagination.[33] This work received widespread coverage in the mainstream media[34] and was listed in the top 10 scientific breakthroughs of the year by the journal Science.[35] He later generalised these ideas to advance the notion of a ‘simulation engine of the mind’ whose role it was to imagine events and scenarios to aid with better planning.[36][37]
CEO and co-founder of DeepMind, a machine learning AI startup, founded in London in 2010 with Shane Legg and Mustafa Suleyman. Hassabis met Legg when both were postdocs at the Gatsby Computational Neuroscience Unit, and he and Suleyman had been friends through family.[38] Hassabis also recruited his university friend and Elixir partner David Silver.[39]
DeepMind’s mission is to “solve intelligence” and then use intelligence “to solve everything else”.[40] More concretely, DeepMind aims to combine insights from systems neuroscience with new developments in machine learning and computing hardware to unlock increasingly powerful general-purpose learning algorithms that will work towards the creation of an artificial general intelligence (AGI). The company has focused on training learning algorithms to master games, and in December 2013 it announced that it had made a pioneering breakthrough by training an algorithm called a Deep Q-Network (DQN) to play Atari games at a superhuman level by only using the raw pixels on the screen as inputs.[41]
DeepMind’s early investors included several high-profile tech entrepreneurs.[42][43] In 2014, Google purchased DeepMind for £400 million. Although most of the company has remained an independent entity based in London,[44] DeepMind Health has since been directly incorporated into Google Health.[45]
Since the Google acquisition, the company has notched up a number of significant achievements, perhaps the most notable being the creation of AlphaGo, a program that defeated world champion Lee Sedol at the complex game of Go. Go had been considered a holy grail of AI, for its high number of possible board positions and resistance to existing programming techniques.[46][47] However, AlphaGo beat European champion Fan Hui 5-0 in October 2015 before winning 4-1 against former world champion Lee Sedol in March 2016.[48][49] Additional DeepMind accomplishments include creating a Neural Turing Machine,[50] reducing the energy used by the cooling systems in Google’s data centers by 40%,[51] advancing research on AI safety,[52][53] and the creation of a partnership with the National Health Service (NHS) of the United Kingdom and Moorfields Eye Hospital to improve medical service and identify the onset of degenerative eye conditions.[54]
More recently, DeepMind turned its artificial intelligence to protein folding, a 50-year grand challenge in science, to predict the 3D structure of a protein from its 1D amino acid sequence. This is an important problem in biology, as proteins are essential to life, almost every biological function depends on them, and the function of a protein is thought to be related to its structure. In December 2018, DeepMind’s tool AlphaFold won the 13th Critical Assessment of Techniques for Protein Structure Prediction (CASP) by successfully predicting the most accurate structure for 25 out of 43 proteins. “This is a lighthouse project, our first major investment in terms of people and resources into a fundamental, very important, real-world scientific problem”, Hassabis said to The Guardian.[55] In November 2020, DeepMind again announced world-beating results in the CASP14 edition of the competition, with a median global distance test (GDT) score of 87.0 across protein targets in the challenging free-modeling category, much higher than the same 2018 results with a median GDT < 60, and an overall error of less than the width of an atom, making it competitive with experimental methods.[56][57]
DeepMind has also been responsible for technical advancements in machine learning, having produced a number of award-winning papers. In particular, the company has made significant advances in deep learning and reinforcement learning, and pioneered the field of deep reinforcement learning which combines these two methods.[58] Hassabis has predicted that Artificial Intelligence will be “one of the most beneficial technologies of mankind ever” but that significant ethical issues remain.[59]
Ulrich Paquet <ulrich@cantab.net> Assessing Game Balance with AlphaZero, Sept. 2020