🎉 Just released TinyTroupe v0.5.1-alpha AND a preprint paper about TinyTroupe in general! The paper is (mostly) in sync with this latest release, so I’d start there :-) (Proper peer-review will come later too) Paper: https://t.co/yk4cp3B7ec Release:

Easily one of the most, if not the most, interesting places anywhere. The largest dig in a lifetime is under way in Pompeii https://t.co/9ai5rEa3g3 From The Economist

Read “"It's just predicting the next token"“ by Paulo Salem on Medium:

Sharing our recent preprint paper about TinyTroupe, our multiagent persona simulation Python library (open source): https://t.co/yk4cp3BF3K Please let me know if you have any feedback, we're working to submit it to an upcoming conference too.

Most "AI people" now (including many who really should know better) seem to think "agents" and "multiagent systems" are new ideas. They are not! People have been researching this for 40 years, at least. Below, a selection of decades-old books from my shelves. Ignore that at your

@hammer_mt , might be of interest, maybe the paper can give you some ideas for your company. BTW, I cited it in one of the footnotes as related :-)

There seems to be a widespread misunderstanding about what "PhD-level" is. Solving difficult problems is only incidentally part of that, many other fine people do it. What truly distinguishes a PhD is posing new interesting problems - and then solving them. Or so it should.

🎉 We've just released TinyTroupe v0.4.0-alpha. This brings various improvements, such as more complex personas, JSON specifications, fragments, propositions and interventions! GitHub repository: https://t.co/z8GyyPR1kq GitHub page: https://t.co/4SKOei7sJc Have fun!

I knew it! Giving children the wrong (or not enough) toys may doom a society https://t.co/Vx2B1x1ZiV From The Economist

A recent paper by the same author of the famous Generative Agents one. They synthesize behavior based on interviews with real people, use *that* as agent definition, and later compares both. I do think we could distill these into (complex) personas though. https://t.co/yTrul4E90C

Wow, TinyTroupe is the 6th top-trending repo on GitHub today! https://t.co/z8GyyPRz9Y https://t.co/kWoPgdynyq

Thanks for making the first TinyTroupe video review that I heard of, @OlenaKutsenko ! I'm glad you had some fun :-) And I appreciate your careful evaluation and feedback.

We got our first independent TinyTroupe video review, and a rather detailed one at that. She even shows the source code with the inner-workings of the agents, which is a clever way to push us to further improve and tidy it up 😂 https://t.co/RyDXxVXKu8

🚀 Now Open-Source: TinyTroupe, a new Python library for LLM-powered multiagent simulation I'm glad to report I'm leading this research project, with the collaboration and support of amazing colleagues. We’ve made it public yesterday (I'm late on X!):

https://t.co/wNnABdstGd

I knew this fridge magnet would come handy one day. Thanks, some @texasgov gift shop.

Beware of those preaching virtue. After all, who wouldn’t be tempted to sacrifice a few if that meant clearing up the way to save many? And who would give up that power, once acquired? Behind Davos, Claims of a Toxic Workplace

This is what a toy store is supposed to look like.

"The mind is its own place, and in it self Can make a Heav'n of Hell, a Hell of Heav'n. What matter where, if I be still the same, And what I should be, all but less then he Whom Thunder hath made greater?" - Paradise Lost, Book I

I for one welcome our new Clippy overlord.