|
It reminds me of a machine-learning startup I worked at for a while. It was a company formed around a single scientist and their research, with the intent of productizing it. It wasn't a scam, people were operating in good faith, but it struggled with the basic problem that turning research into a product, and then selling that product, is difficult and time-consuming, especially for a startup. Yes, in theory there is a large market for more intelligent prediction and analytics across a lot of industries, but actually establishing yourself in say, insurance underwriting, is a difficult slog. The hype around OpenAI and LLMs has slightly obscured the fact that, traditionally, AI has been very difficult to productize. DeepMind were operating for years, doing cool research and solving problems like playing Go, without actually building any usable products. OpenAI have succeeded so far by having massive funding, and by generating enough excitement around the capabilities of their models to produce an ecosystem of people trying to figure out how to build profitable products from it. But most AI platform startups don't have their level of funding or visibility. Now, perhaps everything this company is saying is BS, but if we give the the benefit of the doubt, it sound like they have had some success in a specific area, namely training to play Atari games on more limited data that existing models. If true, that's pretty cool, but ultimately there is no market for an AI that plays Atari games, even at superhuman levels. |
"We are the next DeepMind" is easy to say ... The DeepMind founders had a stellar predigee in computer games, AI and neuroscience, the Verses founders have a cryptocurrency background. Verses also released [1] last month. What both the Atari and the Mastermind announcements have in common is the lack of details, including code. Why do they not show their code? How do we know their figures are real? We've just had the OpenAI vs FrontierMath discussion [2, 3]. Presumably, being able to play Pong, a 1972 computer game, is unlikely to be their moat ...
Interesting also their 2024 MLST presentation [4]. Does that inspire confidence? It was that video that made my priors on Friston having had a breakthrough in ML change downwards dramatically ... But do not take my word for it, please make up your own mind.
[1] https://www.verses.ai/blog/genius-outperforms-openai-model-i...
[2] https://techcrunch.com/2025/01/19/ai-benchmarking-organizati...
[3] https://news.ycombinator.com/item?id=42763231
[4] https://www.youtube.com/watch?v=bL00-jtRrMA