Pinned Tweet
AI agents could transform the economy by slashing the costs of making deals. But can they actually negotiate and honor agreements?
Turns out they negotiate badly under uncertainty and break deals on purpose!
We study AI contracting in our new paper: 🧵👇 arxiv.org/abs/2608.10475
Bhavyesh Sajja retweeted
What if a swarm of AI agents could learn to work together on its own? 🐝
No leader. No roles. No turn order. Just one shared reward.
Our new paper s³: multi-agent RL lets LLM swarms self-organize, beat best-of-N, and scale past their training size!
scalingswarms.com/ 🧵
Bhavyesh Sajja retweeted
An agentic economy might greatly reduce transaction costs, leading to a "Coasean singularity" where cooperation via contracting replaces firms or even states.
But are AI agents any good at negotiating & following natural language contracts? Our paper formalizes & studies this.
AI agents could transform the economy by slashing the costs of making deals. But can they actually negotiate and honor agreements?
Turns out they negotiate badly under uncertainty and break deals on purpose!
We study AI contracting in our new paper: 🧵👇 arxiv.org/abs/2608.10475
Bhavyesh Sajja retweeted
Excited by our new work on AI contracting! Now that we each have our own agents, they need to be able to find mutual benefit with each other and honor their agreements. This is a promising route to supercooperators!
AI agents could transform the economy by slashing the costs of making deals. But can they actually negotiate and honor agreements?
Turns out they negotiate badly under uncertainty and break deals on purpose!
We study AI contracting in our new paper: 🧵👇 arxiv.org/abs/2608.10475
Bhavyesh Sajja retweeted
Very excited to share this paper led by @s_bhavyesh9313! More thoughts soon.
AI agents could transform the economy by slashing the costs of making deals. But can they actually negotiate and honor agreements?
Turns out they negotiate badly under uncertainty and break deals on purpose!
We study AI contracting in our new paper: 🧵👇 arxiv.org/abs/2608.10475
Bhavyesh Sajja retweeted
AI agents could transform the economy by slashing the costs of making deals. But can they actually negotiate and honor agreements?
Turns out they negotiate badly under uncertainty and break deals on purpose!
We study AI contracting in our new paper: 🧵👇 arxiv.org/abs/2608.10475
Some future directions we’re working on:
- Institutions and incentives that promote cooperative contracting
- Scaling laws for contracting: is current training incentivizing pure utility-maximization and lying?
- Complying w/ the spirit of incomplete contracts (w/ @PradyuPrasad)
1
6
117
Joint work with Max Kleiman-Weiner (@maxhkw), Roger Zimmermann (@RogerZimmer00), and Tan Zhi-Xuan (@xuanalogue).
Work initiated during the Cooperative AI Research Fellowship (@coop_ai) earlier this year :)
Paper: arxiv.org/abs/2608.10475
Code: github.com/cosilab/ContractS…
