In case this sounds like I'm anti-AI: I'm not, I love it. I also think the SAT Competition 2026's SAT track should have been won by the competitor using AI, Harrison Green.
Paper of the SAT solver using AI: arxiv.org/pdf/2609.14836v1
Results of the comp: satcompetition.github.io/202…
My take on OpenAI’s ETH paper, as someone who spent nearly 20 yers on SAT: I don’t care about worst case analysis. That’s not what SAT solvers do. They do average case.
Doing analysis of average case is HARD. But it’s what matters. Rest is ivory tower stuff. Not for me.
My take on OpenAI’s ETH paper, as someone who spent nearly 20 yers on SAT: I don’t care about worst case analysis. That’s not what SAT solvers do. They do average case.
Doing analysis of average case is HARD. But it’s what matters. Rest is ivory tower stuff. Not for me.
Mate Soos retweeted
Spend a perfectly nice Saturday indoors with beep boop robo friends, find and validate 30 new attacks, casually publish on a web site ngcc.dev/reports/index.html
Some of these are pretty good. There's a couple of full key recovery scripts coming up ("laptop breaks").
Mate Soos retweeted
Happy to be a speaker at the DeFi Security Summit @summit_defi during @EFDevcon time in Mumbai!!
Will introduce our approach against a long-standing fuzzing problem, "After we fuzz/search against a smart contract, how do we know if it's safe or not?
Replying to @andymooselee
@andymooselee joins DeFi Security Summit 2026.
“No Longer a Dark Forest: measuring What Fuzzing Actually Proves About Smart Contract Security”
See the session and add it to your DSS agenda:
defisecuritysummit.org/2026/…
Tickets: luma.com/9qtxqqxx
Mate Soos retweeted
We published a status update on Core Solidity: what the language can do today, what is still missing, what the community told us, and what comes next.
It's true unfortunately. I am also getting lazier and lazier :S
news.berkeley.edu/2026/10/09…
Mate Soos retweeted
Terence Tao needs to be given a copy of Antifragile.
Birds take off, then some academic branch lectures birds how to fly & takes the credit for it.
My thoughts on incentives in scientific research
msoos.org/2026/10/incentives…
It's a sad post, but it had to be written.
Mate Soos retweeted
Replying to @gloupin
I think STOC made the right call. acm-stoc.org/stoc2027/stoc20…
Mate Soos retweeted
We're happy to share that the @powdr_labs team has joined @axiommathai!
As we begin this next chapter, here's a look at some of the work that brought us here:
New blog post, all hand-written, on XNF solving and CryptoMiniSat:
msoos.org/2026/09/cryptomini…
Turns out that XNF solvers were not as high-performance as they seemed. With a few tweaks, CryptoMiniSat is much faster. I hope someone will write a tool to challenge CryptoMiniSat soon!



















