Andy Konwinski's (@andykonwinski) CAIS 2026 keynote is now on YouTube.
The throughline: real breakthroughs come from convening the right people, working in the open, and actually shipping the research (not just publishing it). That's the bet @CAISconf is built on.
youtu.be/T35_LpmnP7w
CAIS 2026 session recordings are live!: LOADS of talks from San Jose, now up on ACM's YouTube channel.
Watch the full playlist: youtube.com/playlist?list=PL…
ACM Conference on AI and Agentic Systems retweeted
At @CAISconf last month, @andykonwinski sat down with researchers on the conference floor -- @matei_zaharia @istoica05 @lateinteraction @dawnsongtweets @gneubig @pgasawa @JonSaadFalcon @heathercmiller @ryan_marten @alexgshaw @profjoeyg @swyx and Ioannis Ioannidis -- to talk open vs. closed, agent evals, compound AI systems, and where the frontier is headed 👀
youtube.com/watch?v=Pbymvxpb…
Replying to @lily_gpupoor @aiDotEngineer
we are expanding into research this year btw! thanks to @dbreunig @heathercmiller @matei_zaharia and @CAISconf partnership
app.notion.com/p/Call-For-Po…
ACM Conference on AI and Agentic Systems retweeted
Missed SAO at @CAISconf?
Fear not, our deep dive on the keynotes (thanks @ceo_clickhouse @andy_pavlo @nikitabase!), panel and papers is out!
TL;DR: we are moving from “can agents query data?” to “can agents build and operate data systems?” 🧵
ACM Conference on AI and Agentic Systems retweeted
🏆 Rutgers ECE wins Best Paper at @CAISconf for TraceFix, improving reliable AI agent coordination. ortiz.rutgers.edu/blog/trace…
@jortiz81
ACM Conference on AI and Agentic Systems retweeted
Nice feature this morning about @andykonwinski , @LaudeInstitute, and the researchers we back who bet on themselves. Just doing our part to try to save democracy...
American society will certainly be in a better state if everyone has the ability to build and use AI, not just a handful of closed 800-pound gorillas. If you've got a project with breakout potential, send me a DM or apply for Laude support at laude.org/slingshots!
I talked to @andykonwinski about how he’s using his wealth to encourage open research and plug the brain drain from academia to frontier labs, during @TheOfficialACM’s CAIS conference
ACM Conference on AI and Agentic Systems retweeted
Super excited to share that 2 of our papers got the Industry Spotlight at the ACM CAIS Conference!
optimize_anything: A Universal API for Optimizing any Text Parameter
Composing Policy Gradients and Prompt Optimization (mmGRPO)
Thanks to @CAISconf and @aiDotEngineer!
ACM Conference on AI and Agentic Systems retweeted
Can an interface do more than display information.. can it actually shape how you think and what you explore?
We explored this by building Scideator, a human–LLM system for scientific ideation, presented at @CAISconf last week!
ACM Conference on AI and Agentic Systems retweeted
Check out our paper on AI for scientific ideation, with a human-LLM collaboration interface for scientific creativity. Presented last week by @its_sshahid at ACM Conference on AI and Agentic Systems — @CAISconf .
Can an interface do more than display information.. can it actually shape how you think and what you explore?
We explored this by building Scideator, a human–LLM system for scientific ideation, presented at @CAISconf last week!
ACM Conference on AI and Agentic Systems retweeted
What an awesome first @CAISconf! I'm so grateful to @heathercmiller and @matei_zaharia for letting me be part of it. It was a joy to work with such dedicated folks. Huge shout-out to @aviaviavi__, @lateinteraction, @ChenLingjiao, @deeptir18, @jaredq_ and @JoshQuicksall!
We welcome you to the first @CAISconf . We have a great lineup of keynotes, talks and demos starting today for the main conference. Come join us!
Three Best Papers at #CAIS2026. Each one takes something the field treats as safe and breaks it (or fixes it). Worth your time even if you skipped San Jose 🧵
TraceFix 🏆
Multi-agent coordination fails in interleavings no one tests. TraceFix turns a plain-language task into a TLA+-verified protocol, repairing it from model-checker counterexamples. All 48 tasks verified. At runtime it hit 89.4% task completion vs 29.3% for unstructured chat, and held up as models got weaker.
caisconf.org/program/2026/pa…
@jortiz81 (@RutgersECE)
One problem, two solutions, one through-line: agentic systems need real guarantees, not heuristics.
All three, plus 60 more papers: caisconf.org/program/2026/pa…
ACM Conference on AI and Agentic Systems retweeted
Hard to express how high the signal-to-noise ratio was at @CAISconf. So much great content, so many great chats with people getting their hands dirty working on hard problems. Incredible.
Replying to @bradenjhancock
@CAISconf was sold out but small, focused on a particular well-scoped area (compound agentic systems). Some workshop papers were submitted as late as just a couple weeks before (fresh content), and good thing, too! Agentic systems already look quite different today than even 3 months ago.
Perhaps the systems focus naturally attracts researchers who are inclined to build prototypes and libraries (not just publish), but I was still pleasantly surprised to see how many members of the Laude community had independently identified this conference as one they wanted to submit to and attend. On top of that, throwing a lounge after hours with food, drinks, demos, salon-style conversation, (and a mini podcast studio, because why not?) made it easy to get an even higher concentration of interesting people with interesting half-baked ideas in a particular niche that happens to be on fire right now.
Loved the experience. Well-done, CAIS organizers (@heathercmiller, @lateinteraction, @matei_zaharia, @deeptir18, et al). Same time, same place next year?
ACM Conference on AI and Agentic Systems retweeted
@CAISconf was sold out but small, focused on a particular well-scoped area (compound agentic systems). Some workshop papers were submitted as late as just a couple weeks before (fresh content), and good thing, too! Agentic systems already look quite different today than even 3 months ago.
Perhaps the systems focus naturally attracts researchers who are inclined to build prototypes and libraries (not just publish), but I was still pleasantly surprised to see how many members of the Laude community had independently identified this conference as one they wanted to submit to and attend. On top of that, throwing a lounge after hours with food, drinks, demos, salon-style conversation, (and a mini podcast studio, because why not?) made it easy to get an even higher concentration of interesting people with interesting half-baked ideas in a particular niche that happens to be on fire right now.
Loved the experience. Well-done, CAIS organizers (@heathercmiller, @lateinteraction, @matei_zaharia, @deeptir18, et al). Same time, same place next year?






























