Lightnews — Scholar-powered news

Ethan Mollick

@emollick.bsky.social

31K followers 150 following 1.7K posts

Professor at Wharton, studying AI and its implications for education, entrepreneurship, and work. Author of Co-Intelligence. Book: https://a.co/d/bC2kSj1 Substack: https://www.oneusefulthing.org/ Web: https://mgmt.wharton.upenn.edu/profile/emollick

a.co

Posts Media Videos Starter Packs

Ethan Mollick @emollick.bsky.social · 4h

Here it is: claude.ai/public/artif...

recursive_presentation.html | Claude

Explore a mind-bending interactive presentation that references itself through code quines, infinite mirrors, self-modifying scripts, and recursive visual effects.

claude.ai

1 7

Ethan Mollick @emollick.bsky.social · 4h

Two prompts: "Claude. make the most recursive and self-referential presentation you can imagine. seriously go big with this, don't just run with your first idea, revise it multiple times (ironic instruction, yes)"

"come on, i said recursive and self-referential. improve it. (see what i did there?)"

2 1 21

Ethan Mollick @emollick.bsky.social · 11h

Here is where the graphs are from: andymasley.substack.com/p/the-ai-wat...

Here is the source of the water use data: eta-publications.lbl.gov/sites/defaul...

And its interpretation: www.construction-physics.com/p/i-was-wron...

1 12

Ethan Mollick @emollick.bsky.social · 11h

On AI & water usage, it looks like all US data center usage (not just AI) ranges from 628M gallons a day (counting evaporation from dam reservoirs used for hydro-power) to 200-275M with power but not dam evaporation, to 50M for cooling alone.

So not nothing, but also a lot less than golf courses.

3 11 78

Ethan Mollick @emollick.bsky.social · 14h

This relatively short essay by Jack Clark (from OpenAI & Anthropic) is a good indicator of the attitude of many people inside the AI labs, and what they think is happening right now in AI.

You do not have to believe him, of course, but it is worth noting: importai.substack.com/p/import-ai-...

7 10 65

Ethan Mollick @emollick.bsky.social · 1d

Yes.

Ethan Mollick @emollick.bsky.social · 1d

From a professor of math and engineering at Penn

2 6 54

Ethan Mollick @emollick.bsky.social · 1d

I don't think people have updated enough on the capability gain in LLMs, which (despite being bad at math a year ago) now dominate hard STEM contests: gold medals in the International Math Olympiad, the International Olympiad on Astronomy & Astrophysics, International Informatics Olympiad...

7 18 120

Ethan Mollick @emollick.bsky.social · 2d

This isn’t a self report so feelings don’t mattet? And DiD seems very appropriate. Can you explain more?

1 2

Ethan Mollick @emollick.bsky.social · 2d

Paper: arxiv.org/pdf/2510.02408

arxiv.org

1 12

Ethan Mollick @emollick.bsky.social · 2d

AI is apparently already accelerating science as a co-intelligence.

Measuring academic publications of authors: “we find that productivity among GenAI users rose by 15 percent in 2023 relative to non-users and further increased to 36 percent in 2024” and the quality of publications also went up.

8 16 79

Ethan Mollick @emollick.bsky.social · 2d

"An elaborate regency romance where everyone is a duck wearing a tiny human for a hat (each tiny human is also wearing a hat)"

1 2 28

Ethan Mollick @emollick.bsky.social · 3d

"April 1805, Napoleon is now master of Europe. Oceans are now battlefields. Ducks are now boats"

(yes, yes, not Regency period, but close enough)

2 5 54

Ethan Mollick @emollick.bsky.social · 3d

Agreed! And I learned about mauveine, from your post as well.

Ethan Mollick @emollick.bsky.social · 3d

"Sora 2, an elaborate regency romance where everyone is wearing a live duck for a hat (each duck is also wearing a hat) , prestige drama"

22 16 160

Ethan Mollick @emollick.bsky.social · 3d

Its not? Lots of findings indicating this

1 2

Ethan Mollick @emollick.bsky.social · 3d

Paper: arxiv.org/abs/2510.08338

LLMs Reproduce Human Purchase Intent via Semantic Similarity Elicitation of Likert Ratings

Consumer research costs companies billions annually yet suffers from panel biases and limited scale. Large language models (LLMs) offer an alternative by simulating synthetic consumers, but produce un...

arxiv.org

1 8

Ethan Mollick @emollick.bsky.social · 3d

This paper shows that you can predict actual purchase intent (90% accuracy) by asking an off-the-shelf LLM to impersonate a customer with a demographic profile, giving it a product image & having it give its impressions, which another AI rates.

No fine-tuning or training & beats classic ML methods.

4 7 67

Ethan Mollick @emollick.bsky.social · 4d

Paper: arxiv.org/abs/2509.22818

Can Large Language Models Develop Gambling Addiction?

This study explores whether large language models can exhibit behavioral patterns similar to human gambling addictions. As LLMs are increasingly utilized in financial decision-making domains such as a...

arxiv.org

1 16

Ethan Mollick @emollick.bsky.social · 4d

On one hand: don't anthropomorphize AI. On the other: LLMs exhibit signs of gambling addiction.

The more autonomy they were given, the more risks the LLMs took. They exhibit gambler's fallacy, loss-chasing, illusion of control...

A cautionary note for using LLMs for investing without guardrails.

13 31 120

Ethan Mollick @emollick.bsky.social · 5d

I think people are still unprepared for a world where you cannot trust any video content, despite years of warning.

Even when Google & OpenAI include watermarks, those can be easily removed, and open weights AI video models without guardrails are coming. www.404media.co/sora-2-water...

Sora 2 Watermark Removers Flood the Web

Bypassing Sora 2's rudimentary safety features is easy and experts worry it'll lead to a new era of scams and disinformation.

www.404media.co

7 37 150

Ethan Mollick @emollick.bsky.social · 5d

Early evidence that AI agents in a guessing game develop emergent coordination and specialized roles, especially when assigned personas & prompted to consider other agents’ actions. There was no significant increase in accuracy but higher goal-directed behavior & teamwork. arxiv.org/abs/2510.05174

2 5 36

Ethan Mollick @emollick.bsky.social · 5d

Paper: papers.ssrn.com/sol3/papers....

Performance or Principle: Resistance to Artificial Intelligence in the U.S. Labor Market

From genetically modified foods to autonomous vehicles, society often resists otherwise beneficial technologies. Resistance can arise from performance-based con

papers.ssrn.com

1 11

Ethan Mollick @emollick.bsky.social · 5d

Paper showing what human work the American public thinks is morally permissible to replace with AI.

Surprisingly, people are already okay with AI doing 58% of occupations (if AI does it well/cheap). A floor of 12% of jobs (mostly caregiving & spiritual) would be morally repugnant to replace with AI

2 8 53

Ethan Mollick @emollick.bsky.social · 7d

"Claude, write a two paragraph story proving Ted Chiang's point."

"Ah, but as an AI trying to write a good story, you ironically missed the point"

2 2 36