Researchers at Google DeepMind used our free puzzle database and reinforcement learning to train a model to generate creative chess puzzles.
➡️ Read more on this by Tom Zahavy from the DeepMind discovery team: lichess.org/@/tomas135/b...
➡️ Read more on this by Tom Zahavy from the DeepMind discovery team: lichess.org/@/tomas135/b...
AI-Generated Chess Puzzles
A new research by the Discovery team at @GoogleDeepMind using RL and generative models to discover creative chess puzzles
lichess.org
November 11, 2025 at 2:42 PM
Researchers at Google DeepMind used our free puzzle database and reinforcement learning to train a model to generate creative chess puzzles.
➡️ Read more on this by Tom Zahavy from the DeepMind discovery team: lichess.org/@/tomas135/b...
➡️ Read more on this by Tom Zahavy from the DeepMind discovery team: lichess.org/@/tomas135/b...
Thrilled to welcome @fofrAI to Google DeepMind and the AI Studio team as a Senior Prompt Engineer!!
Our generative media models (Veo, Nano Banana, Lyria, etc) are reshaping the industry, such a special moment to make sure everyone can get the most out of them, and more!
Our generative media models (Veo, Nano Banana, Lyria, etc) are reshaping the industry, such a special moment to make sure everyone can get the most out of them, and more!
November 10, 2025 at 3:36 PM
Thrilled to welcome @fofrAI to Google DeepMind and the AI Studio team as a Senior Prompt Engineer!!
Our generative media models (Veo, Nano Banana, Lyria, etc) are reshaping the industry, such a special moment to make sure everyone can get the most out of them, and more!
Our generative media models (Veo, Nano Banana, Lyria, etc) are reshaping the industry, such a special moment to make sure everyone can get the most out of them, and more!
lmao this opener. deepmind being like "maybe it's possible AI could cure all disease! maybe ..... maybe it will happen within the next ten years??" christ could these people be serious for a second
November 11, 2025 at 9:25 AM
lmao this opener. deepmind being like "maybe it's possible AI could cure all disease! maybe ..... maybe it will happen within the next ten years??" christ could these people be serious for a second
Researchers at Google DeepMind have created an AI system capable of generating creative Chess Puzzles, some of which impressed Experts in Chess Compositions. #GenAI #Chess www.chess.com/news/view/ai...
DeepMind's AI Learns To Create Original Chess Puzzles, Praised By GMs
In a new study, researchers from Google DeepMind have created an AI system that is capable of generating creative chess puzzles, some of which impressed experts in chess compositions.
www.chess.com
November 11, 2025 at 8:10 PM
Researchers at Google DeepMind have created an AI system capable of generating creative Chess Puzzles, some of which impressed Experts in Chess Compositions. #GenAI #Chess www.chess.com/news/view/ai...
This is a collab with DeepMind & Terence Tao btw
November 6, 2025 at 10:32 AM
This is a collab with DeepMind & Terence Tao btw
Today's my final routine newsletter of the hurricane season, so I thought I'd dig into the preliminary results of our forecast models. Google's DeepMind scorched the competition, but for America's flagship weather model, it was a season to forget.
This Hurricane Season, Two Forecast Models Stand Out, but for Very Different Reasons
Google’s newest DeepMind hurricane model was the star of the 2025 hurricane season while America’s flagship weather model stumbled to new lows
michaelrlowry.substack.com
November 3, 2025 at 7:41 PM
Today's my final routine newsletter of the hurricane season, so I thought I'd dig into the preliminary results of our forecast models. Google's DeepMind scorched the competition, but for America's flagship weather model, it was a season to forget.
New Google DeepMind paper: "Consistency Training Helps Stop Sycophancy and Jailbreaks" by @alexirpan.bsky.social, me, Mark Kurzeja, David Elson, and Rohin Shah. (thread)
November 4, 2025 at 12:18 AM
New Google DeepMind paper: "Consistency Training Helps Stop Sycophancy and Jailbreaks" by @alexirpan.bsky.social, me, Mark Kurzeja, David Elson, and Rohin Shah. (thread)
HEADLINE: "Apple is planning to use a custom version of Google Gemini for Apple Intelligence"
ALT HEADLINE: "Apple is wishing they bought DeepMind before Google did"
www.theverge.com/news/814654/...
ALT HEADLINE: "Apple is wishing they bought DeepMind before Google did"
www.theverge.com/news/814654/...
Apple is planning to use a custom version of Google Gemini for Apple Intelligence
Siri AI, powered by Google on Apple servers?
www.theverge.com
November 5, 2025 at 8:28 PM
HEADLINE: "Apple is planning to use a custom version of Google Gemini for Apple Intelligence"
ALT HEADLINE: "Apple is wishing they bought DeepMind before Google did"
www.theverge.com/news/814654/...
ALT HEADLINE: "Apple is wishing they bought DeepMind before Google did"
www.theverge.com/news/814654/...
I wrote up some notes on two new papers on prompt injection: Agents Rule of Two (from Meta AI) and The Attacker Moves Second (from Anthropic + OpenAI = DeepMind + others) simonwillison.net/2025/Nov/2/n...
New prompt injection papers: Agents Rule of Two and The Attacker Moves Second
Two interesting new papers regarding LLM security and prompt injection came to my attention this weekend. Agents Rule of Two: A Practical Approach to AI Agent Security The first is …
simonwillison.net
November 2, 2025 at 11:10 PM
I wrote up some notes on two new papers on prompt injection: Agents Rule of Two (from Meta AI) and The Attacker Moves Second (from Anthropic + OpenAI = DeepMind + others) simonwillison.net/2025/Nov/2/n...
🐍 Python Term of the Day: Gemini (AI Coding Tools)
A family of multimodal AI models developed by Google DeepMind.
realpython.com/ref/ai-codin...
A family of multimodal AI models developed by Google DeepMind.
realpython.com/ref/ai-codin...
Gemini | AI Coding Tools – Real Python
A family of multimodal AI models developed by Google DeepMind.
realpython.com
November 5, 2025 at 3:41 PM
🐍 Python Term of the Day: Gemini (AI Coding Tools)
A family of multimodal AI models developed by Google DeepMind.
realpython.com/ref/ai-codin...
A family of multimodal AI models developed by Google DeepMind.
realpython.com/ref/ai-codin...
DeepMind was very impressive. Thank you for sharing your knowledge with us through the season.
November 3, 2025 at 7:46 PM
DeepMind was very impressive. Thank you for sharing your knowledge with us through the season.
Instead of wasting time on another movie during the flight (I was too tired to build at night), I chose a documentary featuring Demis Hassabis, Co-Founder and CEO of #Google #DeepMind.
🧵 My honest review, in case you’re considering watching "The Thinking Game":
#buildinpublic #startuplife #AI #AGI
🧵 My honest review, in case you’re considering watching "The Thinking Game":
#buildinpublic #startuplife #AI #AGI
November 3, 2025 at 8:33 AM
Instead of wasting time on another movie during the flight (I was too tired to build at night), I chose a documentary featuring Demis Hassabis, Co-Founder and CEO of #Google #DeepMind.
🧵 My honest review, in case you’re considering watching "The Thinking Game":
#buildinpublic #startuplife #AI #AGI
🧵 My honest review, in case you’re considering watching "The Thinking Game":
#buildinpublic #startuplife #AI #AGI
Great blog post from hurricane researcher Andy Hazelton about the busy western Pacific, including talking about how Google DeepMind was too aggressive in intensifying #Kalmaegi, apparently due to easterly shear.
www.brightband.com/blog/kalmaeg...
www.brightband.com/blog/kalmaeg...
Kalmaegi Moves Through the Philippines, 90W on the Horizon: AI Models Discussion 11-3-2025
AI models forecast discussion on Typhoon Kalmaegi and Invest 90W in the West Pacific.
www.brightband.com
November 3, 2025 at 7:56 PM
Great blog post from hurricane researcher Andy Hazelton about the busy western Pacific, including talking about how Google DeepMind was too aggressive in intensifying #Kalmaegi, apparently due to easterly shear.
www.brightband.com/blog/kalmaeg...
www.brightband.com/blog/kalmaeg...
Google's DeepMind model has been exceptionally good 🌀
And here is the same, but for all 13 storms in the entire season so far. "OFCL" is the NHC's forecast, and "GDMI" is the Google #DeepMind ensemble mean and is the leading model for both track and intensity this season.
[2/3]
[2/3]
October 31, 2025 at 4:23 PM
Google's DeepMind model has been exceptionally good 🌀
I wrote up some notes on two new papers on prompt injection: Agents Rule of Two (from Meta AI) and The Attacker Moves Second (from Anthropic + OpenAI = DeepMind + others) https://simonwillison.net/2025/Nov/2/new-prompt-injection-papers/
New prompt injection papers: Agents Rule of Two and The Attacker Moves Second
Two interesting new papers regarding LLM security and prompt injection came to my attention this weekend. Agents Rule of Two: A Practical Approach to AI Agent Security The first is …
simonwillison.net
November 2, 2025 at 11:11 PM
I wrote up some notes on two new papers on prompt injection: Agents Rule of Two (from Meta AI) and The Attacker Moves Second (from Anthropic + OpenAI = DeepMind + others) https://simonwillison.net/2025/Nov/2/new-prompt-injection-papers/
They need to harvest other people's creativity and originality. The truly generative systems (like DeepMind is experimenting with) are very narrowly specialized. But GPTs are like a smart rolodex, it's ok for archiving related tasks
November 3, 2025 at 12:24 PM
They need to harvest other people's creativity and originality. The truly generative systems (like DeepMind is experimenting with) are very narrowly specialized. But GPTs are like a smart rolodex, it's ok for archiving related tasks
Meta AI published 'Agents Rule of Two', and an Oct 10, 2025 arXiv paper by 14 authors from OpenAI, Anthropic and DeepMind found adaptive attacks defeated most of 12 defenses (>90% success; 100% in a 500-person red-team contest).
New prompt injection papers: Agents Rule of Two and The Attacker Moves Second
Two interesting new papers regarding LLM security and prompt injection came to my attention this weekend. Agents Rule of Two: A Practical Approach to AI Agent Security The first is …
simonwillison.net
November 3, 2025 at 1:05 PM
Meta AI published 'Agents Rule of Two', and an Oct 10, 2025 arXiv paper by 14 authors from OpenAI, Anthropic and DeepMind found adaptive attacks defeated most of 12 defenses (>90% success; 100% in a 500-person red-team contest).
Our team at Google DeepMind is hiring a software engineer to help us build the future of AI-driven scientific discovery. 🚀
If you want to work with us on scaling scientific simulations, AI agents, and our latest training runs, come join us! 1/2
If you want to work with us on scaling scientific simulations, AI agents, and our latest training runs, come join us! 1/2
November 1, 2025 at 3:14 PM
Our team at Google DeepMind is hiring a software engineer to help us build the future of AI-driven scientific discovery. 🚀
If you want to work with us on scaling scientific simulations, AI agents, and our latest training runs, come join us! 1/2
If you want to work with us on scaling scientific simulations, AI agents, and our latest training runs, come join us! 1/2
And here is the same, but for all 13 storms in the entire season so far. "OFCL" is the NHC's forecast, and "GDMI" is the Google #DeepMind ensemble mean and is the leading model for both track and intensity this season.
[2/3]
[2/3]
October 31, 2025 at 4:04 PM
And here is the same, but for all 13 storms in the entire season so far. "OFCL" is the NHC's forecast, and "GDMI" is the Google #DeepMind ensemble mean and is the leading model for both track and intensity this season.
[2/3]
[2/3]
Google Labs & DeepMind Launch Pomelli AI Marketing Tool www.searchenginejournal.com/google-labs-...
Google Labs & DeepMind Launch Pomelli AI Marketing Tool
Pomelli, a Google Labs & DeepMind AI experiment, builds a "Business DNA" from your site and generates editable branded campaign assets for small businesses.
www.searchenginejournal.com
November 1, 2025 at 3:34 PM
Google Labs & DeepMind Launch Pomelli AI Marketing Tool www.searchenginejournal.com/google-labs-...
I don't know how people can still make this kind of blanket anti-AI of any kind argument after what we just saw Google's DeepMind AI weather modeling do for the Hurricane Melissa forecast. That AI model was what gave the NHC sufficient confidence to start warning about a Cat 5 as early as they did.
I got added to this blocklist, and you know what, fair. bsky.app/profile/did:...
October 30, 2025 at 6:27 AM
I don't know how people can still make this kind of blanket anti-AI of any kind argument after what we just saw Google's DeepMind AI weather modeling do for the Hurricane Melissa forecast. That AI model was what gave the NHC sufficient confidence to start warning about a Cat 5 as early as they did.
"The NHC began using the #DeepMind model at the start of the Atlantic hurricane season in June. For the 13 named storms that have formed so far, the DeepMind model has generally performed well in predicting both their tracks and intensities."
with @franklinjamesl.bsky.social
[1/2]
with @franklinjamesl.bsky.social
[1/2]
This ‘impressive’ AI model predicted Hurricane Melissa’s perilous growth
An artificial-intelligence tool is helping to crack the challenge of forecasting hurricane intensity.
www.nature.com
October 29, 2025 at 9:07 PM
"The NHC began using the #DeepMind model at the start of the Atlantic hurricane season in June. For the 13 named storms that have formed so far, the DeepMind model has generally performed well in predicting both their tracks and intensities."
with @franklinjamesl.bsky.social
[1/2]
with @franklinjamesl.bsky.social
[1/2]
“As early Oct 21 long before it was obvious that Melissa would be a monster storm the model predicted a 50–60% chance that it would reach cat 5, the Google DeepMind team says. On Oct 23 the model estimated the storm had an 80% chance of reaching a cat 5”
www.nature.com/articles/d41...
www.nature.com/articles/d41...
This ‘impressive’ AI model predicted Hurricane Melissa’s perilous growth
An artificial-intelligence tool is helping to crack the challenge of forecasting hurricane intensity.
www.nature.com
October 30, 2025 at 3:14 AM
“As early Oct 21 long before it was obvious that Melissa would be a monster storm the model predicted a 50–60% chance that it would reach cat 5, the Google DeepMind team says. On Oct 23 the model estimated the storm had an 80% chance of reaching a cat 5”
www.nature.com/articles/d41...
www.nature.com/articles/d41...
We wrote about the Google DeepMind AI forecast model that has been doing quite well in predicting Atlantic #hurricane tracks & intensities this year. Thanks @bmcnoldy.bsky.social @franklinjamesl.bsky.social for insights, @NOAA for working without pay:
www.nature.com/articles/d41...
🧪🌀
www.nature.com/articles/d41...
🧪🌀
This ‘impressive’ AI model predicted Hurricane Melissa’s perilous growth
An artificial-intelligence tool is helping to crack the challenge of forecasting hurricane intensity.
www.nature.com
October 29, 2025 at 8:42 PM
We wrote about the Google DeepMind AI forecast model that has been doing quite well in predicting Atlantic #hurricane tracks & intensities this year. Thanks @bmcnoldy.bsky.social @franklinjamesl.bsky.social for insights, @NOAA for working without pay:
www.nature.com/articles/d41...
🧪🌀
www.nature.com/articles/d41...
🧪🌀
DeepMind team: “We are happy to have been able to contribute useful guidance to NHC, but we caution against representing the model’s capabilities based on a single case or metric." @ferranalet.bsky.social
Fair, but DeepMind's track & intensity verification for the entire season looks *VERY* good!
Fair, but DeepMind's track & intensity verification for the entire season looks *VERY* good!
October 29, 2025 at 9:07 PM
DeepMind team: “We are happy to have been able to contribute useful guidance to NHC, but we caution against representing the model’s capabilities based on a single case or metric." @ferranalet.bsky.social
Fair, but DeepMind's track & intensity verification for the entire season looks *VERY* good!
Fair, but DeepMind's track & intensity verification for the entire season looks *VERY* good!