We don’t have an established workflow for prompt development for agents. Heck, we don’t even have established workflows for non-agentic prompt development. A lot of it is still a mix of art and science, and iterating repeatedly.
@emollick
-
GenAI Chatbot Improves Train Quality Control in Controlled Study
By
–
This is the first (small) controlled study I have seen of GenAI on industrial quality control. Here, engineers commissioning new trains took part in an experiment using a GPT-3.5 powered troubleshooting system. Those who used the chatbot had significant increases in work quality
-
Effective Prompting Methods Organize Thoughts and Expression
By
–
Any prompting method that helps you organize your thoughts and clearly express them is likely to be a good prompting method.
-
Veo 3 generates moon landings across historical time periods
By
–
I found another fun Veo 3 prompt: "Realistic footage of the moon landing, but it took place in [year]"
— Ethan Mollick (@emollick) 25 juillet 2025
Here is 1883 AD, 1255 AD, 44 AD, 2300 BC, 30,000 BC, and 65 million years ago
Yes, there is apparently wind on the moon back then, you are just going to have suspend disbelief pic.twitter.com/7j2NvooWrqI found another fun Veo 3 prompt: "Realistic footage of the moon landing, but it took place in [year]" Here is 1883 AD, 1255 AD, 44 AD, 2300 BC, 30,000 BC, and 65 million years ago Yes, there is apparently wind on the moon back then, you are just going to have suspend disbelief
-
Evolutionary Economics and Lotka-Volterra Equations in Organizational Theory
By
–
This is a pretty explicit thing, evolutionary approaches to economics and org theory have been around for ever. Plenty of use for the Lotka–Volterra equations, etc.
-
Google’s Potential Escape from Innovators Dilemma via AI
By
–
It is now entirely possible, whether by luck or planning or both, that Google may escape the Innovators Dilemma and transition from web search to AI. (To be fair, this is not as rare as a lot of people believe: https://
researchgate.net/publication/28
3877064_How_Useful_Is_the_Theory_of_Disruptive_Innovation
…) -
Prompting alternatives to code based task execution methods
By
–
Coders want everything to be code but you can prompt these things all sorts of ways.
— Ethan Mollick (@emollick) 25 juillet 2025
"Show me the thing that Frank Sinatra mentions before he talks about jupiter and mars" pic.twitter.com/siY8emd4OICoders want everything to be code but you can prompt these things all sorts of ways. "Show me the thing that Frank Sinatra mentions before he talks about jupiter and mars"
-
Veo 3 PowerPoint Integration Enables Creative Video Generation
By
–
Tired: Prompting Veo 3 videos with JSON.
— Ethan Mollick (@emollick) 25 juillet 2025
Wired: Prompting Veo 3 videos with PowerPoint. pic.twitter.com/9HUQLXlPzzTired: Prompting Veo 3 videos with JSON. Wired: Prompting Veo 3 videos with PowerPoint.
-
ARC-AGI Benchmark: Keeping AI Labs Honest About Performance
By
–
(Since I am on a benchmark theme today) The ARC team does well keeping AI labs honest about their benchmarks, including showing that Qwen's big ARC-AGI performance doesn't replicate But ARC-AGI also has a strong philosophy of what AI should do. We need other benchmarking efforts
-
AI Benchmarks Correlation Despite Known Limitations and Issues
By
–
The mitigating factor for the problem with AI benchmarks (errors, saturation, contamination) is that, despite issues, they are all still fairly heavily correlated. So if your AI does well on GPQA or MMLU or HLE it also tends to do well on other benchmarks & on vibes & real work.