I wrote about some of the early impact of AI on science last year, including for writing.
@emollick
-
Something Lost: Ethical Concerns About AI Use
By
–
But yes, something is likely lost as well. And not all use is good use.
-
AI Accelerates Scientific Research Through Abstract Writing Automation
By
–
This is, in fact, what early acceleration of science looks like with jagged AI. Writing abstracts takes a lot of time, many scientists are bad writers and clearly written abstracts are important. As long as the authors check over the results, this is a good use case for AI.
-
No-Code AI Game Creation with Iterative Prompting
By
–
This took 5 iterations, I did no coding and all my prompts were a sentence or less (eg "make it nicer and better, also add a tutorial"). Play it here: https://
glittery-raindrop-318339.netlify.app -
Gemini Designs Relativity-Based Missile Command Game
By
–
I prompted Gemini 2.5 Deep Think: "create a missile command game that incorporates relativity in realistic ways but is still playable." I then asked it to improve the design a few times.
— Ethan Mollick (@emollick) 2 août 2025
The full design & all code & calculations came from the AI, each version ran without errors. pic.twitter.com/zT8EfsXItGI prompted Gemini 2.5 Deep Think: "create a missile command game that incorporates relativity in realistic ways but is still playable." I then asked it to improve the design a few times. The full design & all code & calculations came from the AI, each version ran without errors.
-
Llama 4 failure reshapes AI landscape geopolitically
By
–
The relative failure of Llama 4 turned out to be very consequential to the AI landscape. It led to the shifting the locus of open weights development to China, a move towards closed models as companies running local Llama couldn’t continue to upgrade, & big talent wars in the US.
-
Apple’s AI Transition: Can Big Tech Navigate the Shift?
By
–
There is a very good chance that Apple navigates whatever transition AI brings, but also I think about this 2007 Forbes cover often.
-
Anthropic Research on AI Alignment and Personality Control
By
–
This is more neat research from Anthropic, providing a lot of ways for careful organizations to shape the personality and guardrails of AI in deeper ways than prompts, including measuring and reducing sycophancy. Also the idea of an "evil vector" is interesting in and of itself.
-
Prompt Engineering Tips Show Unpredictable Question-Level Performance Impacts
By
–
We keep finding that simple prompting tips and tricks don't really work overall, but, weirdly, can have significant impacts at the question level, sometimes increasing, sometimes decreasing performance in ways that you cannot predict in advance. https://
papers.ssrn.com/sol3/papers.cf
m?abstract_id=5375404
… -
Threats and Tips Don’t Improve AI Performance Study
By
–
New prompting report, from us: Don't bother with threats. Does threatening an AI really make it perform better (the way Google founder Brin claimed)? How about offering to tip the AI? We find no impact of threats or tips on average performance (but variance at question level)