#AI4Innovation I use AI to study the drivers of innovation and discovery @KnowLab @DSI_UChicago
Joined March 2020
- Tweets 1,660
- Following 464
- Followers 561
- Likes 3,713
82 Photos and videos
Pinned Tweet
Excited to share our @DSI_UChicago new work accepted by ICLR 2026, with the remarkable Siyang Wu, Sida Li, Ari Holtzman @universeinanegg, and @profjamesevans James Evans! Link: lnkd.in/gd9YZNBh A thread (1/n)
1
8
26
4,515
Honglin (虹霖) Bao retweeted
"AI's imagination is bounded by what has already been written in the training corpora. Non-reasoning language models collapse into a narrow "hivemind" of similar ideas."
arxiv.org/abs/2606.08251
I've reviewed a couple AI-written papers now, and this was exactly my thoughts. 1/2
1
2
12
1,175
Honglin (虹霖) Bao retweeted
Jun 9
Things look good for AI as a “normal technology” so far.
LLMs have rapidly impacted coding, but few other professions; and even in coding it’s hard to see value generation.
Bottlenecks, regulation, other frictions keeping humans complementary and relevant
26
17
149
38,794
Honglin (虹霖) Bao retweeted
A reporting checklist for large language models in behavioural science dlvr.it/TSxtMM
10
39
7,566
“AI agents will outperform humans at almost all jobs by 2026–2027.” - The forecast is everywhere.
So we built the exam to test that claim, on real labor-market aligned work. On the hardest tier, top agents pass 2.6%.
Meet Agents' Last Exam (ALE), a rolling benchmark measuring whether agents can actually do real jobs. 🧵👇
14
102
331
84,853
Honglin (虹霖) Bao retweeted
Important to keep in mind that even in what is arguably the most rigorous part of social science -- econ studies with preregistration -- the papers hide most of hypotheses tested
3
6
31
5,530
Honglin (虹霖) Bao retweeted
Students already extensively use AI. This lets everyone have access to the same models and (hopefully) allows new kinds of data protections and in-house uses. AI is not being held at bay by the university not providing it. Not an AI booster, but don't understand this reaction.
Genuinely, can someone give me the steel man version of the rationale behind the new “give everyone AI” university strategy? What is the theory of the case here? Do universities think it’s sustainable to ask students to pay over $90k per year to cheat their way through college?
8
4
43
7,197
genuinely no clue why people are this mad. because it's Chicago? the last bastion of the humanities or whatever?
UChicago announced today that it had partnered with AI company Anthropic to give students, faculty, and staff access to Claude Enterprise services on a rolling basis starting in July. All University community members will have access by fall quarter. Story to come.
3
702
Honglin (虹霖) Bao retweeted
Let's compare which social science topics general science journals (Nature, Science, PNAS...) publish on VS disciplinary journals (AER, APSR, ...)
Econ: general sci overrepresents environment energy
Poli sci: general sci way overrepresents media misinfo
3
5
30
9,988
Honglin (虹霖) Bao retweeted
Yes, it is really Brown university.
2
2
56
18,297
Honglin (虹霖) Bao retweeted
I guess I'm no longer assisting and can finally ASSOCIATE!!!
(And get ready for the posts / takes to get spicier! 🌶️🌶️🌶️)
Huge thanks to my *students*, colleagues, advisors!
And frankly huge thanks to @umsi for supporting junior faculty, and having reasonable standards, and good leadership. I hear at other places pre-tenure period can be super stressful and sleepless, but here it's been (almost entirely) fun! <3
9
1
96
4,338
Honglin (虹霖) Bao retweeted
Setting up a rival to Arxiv where AI slop is not just acceptable but actively encouraged: no censorship, no one to tell you your claims are obviously fake. a whole ecosystem of people outdoing each other with no quality bar to hold them back. it’s called LinkedIn
16
156
2,091
31,730
📄 Excited to share our latest preprint: the first cross-field audit of LLM-hallucinated citations in science
⚠️ Across arXiv, bioRxiv, SSRN & PMC, we estimate 147K fake citations in 2025 alone — threatening both the quality and equity of scientific work.
5
57
164
21,636
Honglin (虹霖) Bao retweeted
May 8
Thoughtful coverage from Nature magazine on our recent article in Science about how aging narrows innovation in science, individually (scientist by scientist) and collectively (field by field, nation by nation). Thanks again, Haochuan Cui, Yiling Lin, & @LingfeiWu!
nature.com/articles/d41586-0…
15
41
3,636
Honglin (虹霖) Bao retweeted
May 8
Dashun Wang talking 'airplanes for the mind'
🛩️ are heavier than air, but can fly faster than sound—remind you of another technology that "shouldn't" be able to work, but does?
Apr 1
Excited to announce the 2026 iteration of the Communication & Intelligence Symposium at UChicago!
We have an amazing lineup of speakers @Diyi_Yang @johnhewtt @dashunwang @TomerUllman
We have a simple call for abstract that is due on Apr 15 (links 👇). Please come and share your research!
Co-organized with the awesome @universeinanegg and @divingwithorcas
3
12
2,807
Honglin (虹霖) Bao retweeted
May 8
Out today in @ScienceMagazine: with the amazing Haochuan Cui, Yiling Lin, & @LingfeiWu, we analyzed 3.6 million scientists publishing 1960–2020. The findings reshape a century-old debate about age and scientific creativity.
7
95
284
80,564
Honglin (虹霖) Bao retweeted
May 6
HAI Faculty Affiliate @james_y_zou is exploring how AI can accelerate scientific research and peer review. At the AI Science conference, he highlighted that AI excels at spotting gaps, but judgment calls still need humans. Read more about his insights: hai.stanford.edu/news/ais-gr…
2
8
49
5,850
Honglin (虹霖) Bao retweeted
May 5
Happening on Friday! Super excited to have the event as a post-neurips unwind!
Apr 1
Excited to announce the 2026 iteration of the Communication & Intelligence Symposium at UChicago!
We have an amazing lineup of speakers @Diyi_Yang @johnhewtt @dashunwang @TomerUllman
We have a simple call for abstract that is due on Apr 15 (links 👇). Please come and share your research!
Co-organized with the awesome @universeinanegg and @divingwithorcas
4
20
7,331
Honglin (虹霖) Bao retweeted
May 3
If you produce 100 papers per year you're probably doing more harm than good for the community
if you are able to get 14 papers accepted to ICML, maybe you do not actually need to post about getting 14 papers accepted to ICML
3
22
331
33,933
Honglin (虹霖) Bao retweeted
May 3
Again, this is bad for science, please don't do it
May 3
Such a great evening to start a brand new research for NeurIPS in 3.5 days.🧘♂️
Day 1: planning.
Night 1: running experiments and sending the abstract.
Day 2: reading results fighting with Claude, and sending again.
Night 2: sleep (optional).
Day 3: opening Codex, and finally, write the pape in parallel.
Night 3: resolving the “beef” with Claude (temporary peace) and going to sleep.
Day 4: final reading, last-minute fixes, submission then some relaxation, maybe a beach walk.
I’ll keep you posted on the results.
This will be my only single-author paper, so I can’t hide behind other submissions if it gets rejected 😅
10
19
539
65,391
Honglin (虹霖) Bao retweeted
Can you boost your AI review scores by asking an LLM to rewrite your paper?
Yes! We call it paper laundering
Our @icmlconf spotlight paper argues current AI reviewers aren't ready to automate peer review, and outlines what a science of peer review automation should look like🧵👇
ALT First page of the ICML 2026 spotlight paper "Stop Automating Peer Review Without Rigorous Evaluation" by Joachim Baumann, Jiaxin Pei, Sanmi Koyejo, and Dirk Hovy (Stanford University and Bocconi University). The abstract argues that today's AI systems should not be used to produce paper reviews, grounded in two empirical findings: a "hivemind effect" where AI reviewers show excessive agreement and reduce perspective diversity, and "paper laundering," where prompting an LLM to rewrite a paper trivially increases AI reviewer scores through stylistic changes rather than scientific improvements. The paper calls for a science of peer review automation rather than wholesale deployment of general-purpose LLMs.
14
75
458
53,012