Lightnews — Scholar-powered news

Reposted by Jeff Clune

Aaron Dharna @aadharna.bsky.social · Jul 10

For all the details, please give the paper a read!
Paper: arxiv.org/abs/2507.06466

Infinite thanks to @jeffclune.com and @cong-ml.bsky.social for all their guidance!

Foundation Model Self-Play: Open-Ended Strategy Innovation via Foundation Models

Multi-agent interactions have long fueled innovation, from natural predator-prey dynamics to the space race. Self-play (SP) algorithms try to harness these dynamics by pitting agents against ever-impr...

arxiv.org

2 8

Reposted by Jeff Clune

UBC Computer Science @cs.ubc.ca · Aug 7

Read this article on how AI is contributing to its own development, featuring UBC Computer Science Professor @jeffclune.com!

MIT Technology Review @technologyreview.com · Aug 6

The possibility of self-improvement distinguishes AI from other revolutionary technologies. But LLMs can optimize the computer chips they run on, train other LLMs cheaply and efficiently, and perhaps even come up with original ideas for AI research. And they’ve already made progress. trib.al/YbuLO16

Five ways that AI is learning to improve itself

From coding to hardware, LLMs are speeding up research progress in artificial intelligence. It could be the most important trend in AI today.

trib.al

1 3

Reposted by Jeff Clune

Schwartz Reisman Institute for Technology and Society @torontosri.bsky.social · 23d

💡The fall SRI Seminar Series kicks off on Wednesday with
@jeffclune.com (UBC / Vector / DeepMind):

“Open-ended and AI-generating algorithms in the era of foundation models”

Wed 12:30 ET
Free, online🔗https://srinstitute.utoronto.ca/events-archive/seminar-2025-jeff-clune-2

1 3

Jeff Clune @jeffclune.com · Jul 10

I'd post exclusively here if @bsky.app would get rid of the silly character limit

1 2

Jeff Clune @jeffclune.com · Jul 10

Thrilled to introduce Foundation Model Self-Play, led by
Aaron Dharna. FMSPs combine the intelligence & code generation of foundation models with the curriculum of self-play & principles of open-endedness to explore diverse strategies in multi-agent games. Thread x.com/jeffclune/st...

Jeff Clune on X: "Thrilled to introduce Foundation Model Self-Play, led by @_aadharna. FMSPs combine the intelligence & code generation of foundation models with the curriculum of self-play & principles of open-endedness to explore diverse strategies in multi-agent games, like the one below 🧵👇 https://t.co/GqTWM5P3CC" / X

Thrilled to introduce Foundation Model Self-Play, led by @_aadharna. FMSPs combine the intelligence & code generation of foundation models with the curriculum of self-play & principles of open-endedness to explore diverse strategies in multi-agent games, like the one below 🧵👇 https://t.co/GqTWM5P3CC

x.com

1 7

Jeff Clune @jeffclune.com · Jun 27

Nice summary, with coverage of the Darwin Gödel Machine.

www.youtube.com/watch?v=C1ku...

AI’s “Intelligence Explosion” Is Coming. Here’s What That Means.

YouTube video by Sabine Hossenfelder

www.youtube.com

2

Jeff Clune @jeffclune.com · Jun 27

A great summary of the latest in open-endedness research! The next big wave in AI? 🤔📈🚀

richardcsuwandi.github.io/blog/2025/op...

The future of AI is open-ended | Richard Cornelius Suwandi

Embracing open-endedness in the pursuit of creative AI

richardcsuwandi.github.io

3

Jeff Clune @jeffclune.com · Jun 23

How it felt to know AGI is coming soon long before the world was paying attention.

officechai.com/ai/building-...

Building AI Felt Like Watching An Alien Intelligence Arrive: Early OpenAI Researcher

OpenAI had stunned the world by releasing ChatGPT in November 2022, but it turns out that the researchers working on the technology were even more stunned at what they'd developed. Jeff Clune, a fo...

officechai.com

1 5

Jeff Clune @jeffclune.com · Jun 20

Very nice summary of the Darwin Gödel Machine in Fortune, and an interesting tie-in to Sam's excellent recent blog post.

fortune.com/2025/06/19/o...

OpenAI CEO Sam Altman says "we are past the event horizon." Is he right?

Researchers from Sakana AI and Google DeepMind have created AI that can improve its own code.

fortune.com

1

Jeff Clune @jeffclune.com · Jun 10

Thanks!

The first two talks here are a good entry point 🍿🎬

jeffclune.com/videos.html

Jeff Clune - Professor - Computer Science - University of British Columbia

Homepage for Jeff Clune, Professor of Computer Science at the University of British Columbia. I conduct research in deep learning, including deep reinforcement learning. I improve our understanding of...

jeffclune.com

1 1

Jeff Clune @jeffclune.com · Jun 3

I am excited to be a part of @Yoshua_Bengio's new non-profit focused on AI Safety and Existential Risk, joining the great team of Scientific Advisors. This is a critically important mission for humanity.

Yoshua Bengio @yoshuabengio.bsky.social · Jun 3

Today marks a big milestone for me. I'm launching @law-zero.bsky.social, a nonprofit focusing on a new safe-by-design approach to AI that could both accelerate scientific discovery and provide a safeguard against the dangers of agentic AI.

LawZero - LoiZéro @law-zero.bsky.social · Jun 3

Every frontier AI system should be grounded in a core commitment: to protect human joy and endeavour. Today, we launch LawZero, a nonprofit dedicated to advancing safe-by-design AI. lawzero.org

17

Reposted by Jeff Clune

Stand Up for Science! @standupforscience.bsky.social · May 31

It's Friday Night, which means Trump's weekly science bloodbath.

BUT WHY DOES HE DO THIS ON FRIDAY NIGHTS?

Simple: He's a coward!

Americans overwhelmingly support science! His attack on science is unpopular and he wants to bury it in the news cycle.

Not on our watch.
#StandUpForScience

5 94 360

Jeff Clune @jeffclune.com · May 30

Excited to introduce the Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents. We harness the power of open-ended algorithms to search for agentic systems that get better at coding, including improving their own code.

sakanaai.bsky.social @sakanaai.bsky.social · May 30

Introducing The Darwin Gödel Machine

sakana.ai/dgm

The Darwin Gödel Machine is a self-improving agent that can modify its own code. Inspired by evolution, we maintain an expanding lineage of agent variants, allowing for open-ended exploration of the vast design space of such self-improving agents.

6 30

Jeff Clune @jeffclune.com · May 20

Note: This is not a fundamental dismissal of deep learning. In fact, we have worked improving and leveraging deep learning for years, and are optimistic about its present and future. These results point to a way it could be substantially improved, which is exciting.

8

Jeff Clune @jeffclune.com · May 20

Is there a cancer at the heart of modern AI, lurking just beneath the surface of its dazzling performance? Our research suggests maybe, but also shows elegant solutions are possible (though how to get them at scale remains a mystery). Check out the eye-opening, riveting paper below!

Kenneth Stanley @kennethstanley.bsky.social · May 20

Could a major opportunity to improve representation in deep learning be hiding in plain sight? Check out our new position paper: Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis.

Paper: arxiv.org/abs/2505.11581

2 13

Jeff Clune @jeffclune.com · May 16

Why do we tolerate loud motorcycles? I can't walk around with a crazy loud speaker, yet we allow people to have (and companies to sell) insanely loud machines that torture everyone else. They can be made quiet or silent...we should start heavily taxing and fining noise polluting machines.

12 7 47

Jeff Clune @jeffclune.com · May 14

Great work AlphaEvolve team!

MAP-Elites co-authored with Jean-Baptiste Mouret

1

Jeff Clune @jeffclune.com · May 14

Very cool work! Fun to see an algorithm inspired by MAP-Elites help advance numerous open problems in math, computer science, AND improve Google's production infrastructure (data centers, TPU design, and AI training)! (Paraphrasing their paper and tweet).

MIT Technology Review @technologyreview.com · May 14

AlphaEvolve finds new algorithms that outperform the best human-made solutions for data center management, chip design and more.

Google DeepMind’s new AI uses large language models to crack real-world problems

AlphaEvolve finds new algorithms that outperform the best human-made solutions for data center management, chip design and more.

www.technologyreview.com

1 2 11

Jeff Clune @jeffclune.com · Apr 29

I greatly enjoyed “The Spectrum of AI Risks” panel at the Singapore Conference on AI. Thanks @teganmaharaj.bsky.social for great moderating, Max Tegmark for the invitation, and the organizers and other panelists for a great event!

PS. Do I really have sad resting panel face? 😐

2 1 8

Jeff Clune @jeffclune.com · Apr 28

I'm giving a talk at 11:30 today in the #ICLR World Models workshop. "Open-ended Agent Learning in the Era of Foundation Models and Foundation World Models." Drop by if you are interested! sites.google.com/view/worldmo...

ICLR 2025 Workshop World Models - Schedule & Speaker

Schedule (Tentative)

sites.google.com

5

Jeff Clune @jeffclune.com · Apr 27

Tom @rockt.ai did a great job in his #ICLR2025 keynote on open-endedness of explaining the ideas we are all so passionate about. A huge thanks for the kind words and for featuring our work, including Jenny Zhang's OMNI. cc @kennethstanley.bsky.social @joelbot3000.bsky.social

10

Jeff Clune @jeffclune.com · Apr 25

Very excited for this keynote by @_rockt! Awesome to see open-endedness go from a niche (😉) area to a keynote at #ICLR ! 🌱🌿🌳🌲🍀🌍✨ 📈 🧬🧪 cc @joelbot3000.bsky.social @kennethstanley.bsky.social

1 10

Jeff Clune @jeffclune.com · Apr 8

Since the dawn of my career I've heard scientists joke after every AI advance “Now we just need to get AI to write the paper. Now The AI Scientist does all the research AND writes the paper! Wild times! 🧪🔬👩‍🔬🤖

sakanaai.bsky.social @sakanaai.bsky.social · Apr 8

Introducing The AI Scientist-v2, which produced the 1st fully AI-generated paper to pass peer review at a workshop level (at #ICLR2025) ‼️

Tech Report: pub.sakana.ai/ai-scientist...
GitHub: github.com/SakanaAI/AI-...

This work is a proud collaboration between Sakana AI, UBC, and Oxford University.

2 13

Reposted by Jeff Clune

sakanaai.bsky.social @sakanaai.bsky.social · Apr 8

Introducing The AI Scientist-v2, which produced the 1st fully AI-generated paper to pass peer review at a workshop level (at #ICLR2025) ‼️

Tech Report: pub.sakana.ai/ai-scientist...
GitHub: github.com/SakanaAI/AI-...

This work is a proud collaboration between Sakana AI, UBC, and Oxford University.

1 7 18

Reposted by Jeff Clune

UBC Science @science.ubc.ca · Apr 4

Collecting a diamond in #Minecraft is “a very hard task”, says Dr. Jeff Clune (@jeffclune.com) of @cs.ubc.ca, who was part of a team that trained a program to find diamonds using videos of human play2. “This represents a major step forward for the field.”

www.nature.com/articles/d41...

AI masters Minecraft: DeepMind program finds diamonds without being taught

The Dreamer system reached the milestone by ‘imagining’ the future impact of possible decisions.

www.nature.com

3 4