Chen Sun @NeurIPS2025
@ChenSun92
Followers
2K
Following
4K
Media
75
Statuses
571
Research Scientist @ Google DeepMind Building long term memory & open-ended AI ex-neuroscientist ex-IMO team Canada Views are mine not Google's.
Joined April 2019
At #NeurIPS2025 this week ✈️ My team at @GoogleDeepMind work on long term memory & open endedness. DMs open. Let’s have coffee & brainstorm the ultimate solution to continual learning 🧙♂️
7
4
111
Insight from @zeynep 's beautiful keynote titled "are we having the wrong nightmares".
0
0
3
Current metaphors for thinking about AI will completely miss its most important consequences, just like thinking about cars as horseless carriages and other analogies to "things we know" completely missed the mark of how steam engines actually transformed society. - from
5
14
35
People have been talking about getting compositionality to work in ML for ages, but maybe getting it to emerge in solely gradient based methods is the wrong approach. In this gorgeous interview by @blaiseaguera on @MLStreetTalk , there is the intriguing proposition that you
DNA sequences are computer code! It was a pleasure to explore John Von Neumann's theory on life as a form of computation and the lessons we can learn from modern AI systems with @ecsquendor on @MLStreetTalk: https://t.co/wgl4cPtOix Thank you for the thoughtful questions!
0
3
20
The gamification of learning as it stands is probably a terrible detour / dead end in attempt to scale human learning. From the neuro point of view, there is likely a huge distinction between the cheap dopamine of gamification and the endogenous opioids (endorphins) released
there's an epidemic of fake learning. duolingo, tiktok, youtube. it's all entertainment cleverly disguised as education. real learning is hard. it's uncomfortable. if it feels 'fun', you probably aren't learning anything.
1
0
8
huge ++1 @kenneth0stanley if one considers all of AI since mcculloch pitts as taking useful things from nature and figuring out how to formalize and use them, then evolution is still that one big gaping hole that no one has figured out how to properly exploit. UFRs are super
Reflecting the gut feeling of many, Ilya says “something important” is missing from current AI models. But what is the concrete nature of this chasm? One candidate: the difference between fractured entangled representation (FER) and unified factored representation (UFR).
0
0
15
AI researchers love to think they've outsmarted biological evolution. But when things don't work, they come back to the brain. It's the only true proof of intelligence.
The @ilyasut episode 0:00:00 – Explaining model jaggedness 0:09:39 - Emotions and value functions 0:18:49 – What are we scaling? 0:25:13 – Why humans generalize better than models 0:35:45 – Straight-shotting superintelligence 0:46:47 – SSI’s model will learn from deployment
1
3
10
Super awesome opportunity from Google research!! If I am not mistaken this team is under the visionary @blaiseaguera Fun fact: I was originally hired by Blaise to Google. If you’re interested in making the big leap toward understanding intelligence and the origins of sentient
Great opportunity at Google Research for folks interested in AutoML, evolutionary methods, meta-learning, and open-endedness: London: https://t.co/clY7vfiW9S Zürich: https://t.co/r5fty4ibPf
0
0
13
Our team at @GoogleDeepMind is looking to hire a talented new Research Scientist! Our group (under @edchi) aims to push the frontier of AI-human interactions through personalization of LLMs and deeply understanding the open-ended nature of user intentions. Beneath this lies
5
28
259
Can’t recommend enough applying to @robertarail ‘s team and making a yuuuge impact in AI open endedness!
I’m building a new team at @GoogleDeepMind to work on Open-Ended Discovery! We’re looking for strong Research Scientists and Research Engineers to help us push the frontier of autonomously discovering novel artifacts such as new knowledge, capabilities, or algorithms, in an
1
1
24
Congrats to our very own @GoogleDeepMind and @lmthang for getting the first OFFICIALLY VERIFIED AI Gold at the IMO. As a former IMO student, it was heartening also to see Google respecting the brilliant students at the IMO, to follow the official rules and deferring the
Very excited to share that an advanced version of Gemini Deep Think is the first to have achieved gold-medal level in the International Mathematical Olympiad! 🏆, solving five out of six problems perfectly, as verified by the IMO organizers! It’s been a wild run to lead this
2
6
99
Having competed in the IMO, I find the generalist training this work provides to be its most incredible aspect. 🧙♂️ No more rabbithole specialization as we did in high school, thinking specifically about olympiad problem solving 🤔, eating and sleeping olympiad problems.
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
1
4
38
Are AI scientists already better than human researchers? We recruited 43 PhD students to spend 3 months executing research ideas proposed by an LLM agent vs human experts. Main finding: LLM ideas result in worse projects than human ideas.
12
196
635
video link here:
Hello friends! Uploading the video of my invited talk 👇: Transforming long-horizon experience into wisdom in Brains and AIs. Talks thru the day include Jurgen Schmidhuber, Jay McClelland [5:56], Mine [1:03:08] Ivana Kajic [1:48:20] Naomi Saphra [2:12:16] Michael Lepori
0
0
2
Hello friends! Uploading the video of my invited talk 👇: Transforming long-horizon experience into wisdom in Brains and AIs. Talks thru the day include Jurgen Schmidhuber, Jay McClelland [5:56], Mine [1:03:08] Ivana Kajic [1:48:20] Naomi Saphra [2:12:16] Michael Lepori
0
7
38
🚀Introducing “StochasTok: Improving Fine-Grained Subword Understanding in LLMs”!🚀 LLMs are incredible but still struggle disproportionately with subword tasks, e.g., for character counts, wordplay, multi-digit numbers, fixing typos… Enter StochasTok, led by @anyaasims! [1/]
1
26
77