#AI#Google#DeepMind#Science#Hype: "In a perspective paper published in Chemical Materials this week, Anthony Cheetham and Ram Seshadri of the University of California, Santa Barbara selected a random sample of the 380,000 proposed structures released by DeepMind and say that none of them meet a three-part test of whether the proposed material is “credible,” “useful,” and “novel.” They believe that what DeepMind found are “crystalline inorganic compounds and should be described as such, rather than using the more generic label ‘material,’” which they say is a term that should be reserved for things that “demonstrate some utility.”
In the analysis, they write “we have yet to find any strikingly novel compounds in the GNoME and Stable Structure listings, although we anticipate that there must be some among the 384,870 compositions. We also note that, while many of the new compositions are trivial adaptations of known materials, the computational approach delivers credible overall compositions, which gives us confidence that the underlying approach is sound.”
In a phone interview, Cheetham told me “the Google paper falls way short in terms of it being a useful, practical contribution to the experimental materials scientists.” Seshadri said “we actually think that Google has missed the mark here.”"
#AI#GenerativeAI#DeepMind#Inflection: "Microsoft has hired Mustafa Suleyman, the co-founder of Google’s DeepMind and chief executive of artificial intelligence start-up Inflection, to run a new consumer AI unit.
Suleyman, a British entrepreneur who co-founded DeepMind in London in 2010, will report to Microsoft chief executive Satya Nadella, the company announced on Tuesday. He will launch a division of Microsoft that brings consumer-facing products including Microsoft’s Copilot, Bing, Edge and GenAI under one team called Microsoft AI.
It is the latest move by Microsoft to capitalise on the boom in generative AI. It has invested $13bn in OpenAI, the maker of ChatGPT, and rapidly integrated its technology into Microsoft products." https://www.ft.com/content/5feedf3a-ff7a-4c89-9b1d-f9b48834ff4c
En 2018, #AlphaFold, l’IA créée par #DeepMind, faisait une percée spectaculaire. [...] Une IA capable de prédire à quoi ressemblera une #protéine à partir de sa seule séquence. Sur les 200 millions de protéines connues, nous ne connaissons la structure 3D que de 20 % d'entre elles. Le #deeplearning fait donc cette promesse : combler ces lacunes pour de grandes avancées médicales et la création de nouveaux #médicaments
Aleph Alpha ist ja die große Hoffnung der deutschen KI Industrie:
Hat jemand von denen schon mal irgendein Paper gesehen, irgendein Lebenszeichen, dass ein Alleinstellungsmerkmal wäre? Also: Sind die mehr als "der LIDL Gründer gibt nem Startup Geld für NVIDIA Hardware und Scraping um sich dann an KI für die Verwaltung gesundzustoßen"?
I used to look at these kinds of statements as deceptive PR, but increasingly I see them more through the lens of faith.
The tech billionaires are true believers and don’t accept they’re misunderstanding things like intelligence because they believe themselves to be geniuses.
To them, everything is reduced to computation: the brain is a computer; climate change is a technological problem. But none of that is true, and we’re setting ourselves up for chaos if we keep believing these men who assert tech will save us from the crises we face.
Solving problems by invoking upcoming technology breakthroughs is not solving problems at all. It's fantasizing. No technology will ever allow us to sustain our existence without having to prioritize sustainability.
This is exactly what #AI is about for me
“What’s most exciting to me is modelling new modes of human–machine collaboration,” Ellenberg adds. “I don’t look to use these as a replacement for human mathematicians, but as a force multiplier.” https://www.nature.com/articles/d41586-023-04043-w #deepmind
#AI#GenerativeAI#Google#Gemini#DeepMind: "Experts say it’s unclear whether the benchmarks Google is using to measure Gemini’s performance offer that much insight, and without transparency, it’s hard to check Google’s claims.
“Google is advertising Gemini as an everything machine—a general-purpose model that can be used in many different ways,” says Emily Bender, a professor of computational linguistics at the University of Washington. But the company is using narrow benchmarks to evaluate models that it expects to be used for these diverse purposes. “This means it effectively can’t be thoroughly evaluated,” she says.
@RyunoKi just reminded me of this piece I wrote two years ago. It seems rather timely what with the news about Harvard and Meta, and, beyond tech, Cop28, etc.
I'm here for this infighting. It reveals quite a bit.
#DeepMind CEO Demis Hassabis pushes back on claims by Meta's Yann LeCun that he, Sam Altman, and Dario Amodei are fearmongering to achieve #AI regulatory capture.
Deepmind releases AlphaMissense: an AI model that can classify the effects of 71 million possible genetic mutations on human proteins and help pinpoint the causes of diseases. (x.com)