arxiv.org/abs/2205.11327 Sketching algorithms people are at it again, this time with logloglog and loglogloglog (!?) terms 😔

Excited to share our paper on generative social choice, which "fuses the rigor of social choice theory with the flexibility and power of generative AI." We plan to put this approach into action soon as part of OpenAI's "Democratic Inputs to AI" program. arxiv.org/pdf/2309.01291…

🚨New paper🚨 Algorithmic Collusion by Large Language Models Joint w/Sara Fish& Ran Shorrer LLM use is automating many business decisions. Pricing might be next (or is already). What if multiple firms decide in good faith to use off-the-shelf-LLMs for pricing? 1/3 #EconTwitter

Here are some brief experiments providing evidence that text in the system message has a stronger effect on gpt-4o's behavior than text in the rest of the prompt. Similar patterns emerge for all other models in the gpt-4 series, except gpt-4-0314

🚨Major new version🚨 Algorithmic Collusion by Large Language Models Joint w/Sara Fish & Ran Shorrer LLMs are automating many business decisions. Pricing might be next (or is already). What if multiple firms, in good faith, use off-the-shelf-LLMs for pricing? 1/3 #EconTwitter

Come join at 3:30 Eastern today to hear me talk about algorithmic collusion by large language models, as well as a sneak peak into a new, not-yet-online, paper! Sara Fish Ran Shorrer #econtwitter

Amazing workshop at UCL Faculty of Laws yesterday. I presented my joint work w/ Sara Fish Yannai A. Gonczarowski in panel w/speakers from European Commission Competition & Markets Authority Autorité de la Concurrence HELLENIC COMPETITION COMMISSION. So gratifying to see our research makes positive impact in the world. Thanks much Ioannis Lianos for inviting us!