Pinned Tweet
If freedom = privacy, then we could rephrase this quote by Sigmund Freud as:
"Most people don't really want privacy, because privacy involves responsibility, and most people fear responsibility." 1/2
#Web3āā
1
2
2,869
Valentinus Sofa š retweeted
youtu.be/s_lGhOD8uN4?si=qBmsā¦
BTW, in case you watch this video after today, please note that my speculation regarding "what if there were a 20 billion token proof of the Riemann Hypothesis that nobody understood" was made two weeks ago and wasn't based on any inside information...
Valentinus Sofa š retweeted
I think my new rule is that if you say you formally verified some complex software and don't say anything about all the things you had to fix to make the proof go through, then I simply don't believe you
Valentinus Sofa š retweeted
Ten years ago, there was this weird trend in software engineering where people based their identities around what kind of verification software they used.
Some people would be like, "I use TDD, I never want to see a type in my life", and others would say "if it typechecks it's correct I don't even need to bother running it."
Nowadays, I think we're a bit more aware that techniques don't just complement each other, but actively synergize. For example, if you have both types and tests, you can use the types to generate inputs for your tests!
One of the cooler ways synergies I've seen in trace validation. Normally, testing finds bugs in codebases, while "formal specification" finds bugs in designs. Neither by itself can verify if the code actually *matches* the design, though. With trace validation, we generate logs of the executed system, translate them into abstract behaviors, and check if they match the abstract specification. It's a cool fusion of experiment and proof!
And if those logs are generated with fuzzing and deterministic simulation testing, you've got something really special going.
My coworker @moarbugs just published a demonstration of this, using formal specification and @antithesishq to verify Raft implementations against their designs. He did the bulk of the work but I wrote a sections on the caveats, long-term consequences, and an explainer of what trace validation actually is. Come check our work out!
By combining Antithesis with formal methods, we can mathematically prove safety for conformant models of production distributed systems! We demonstrated this process in depth for etcd-raft and HashiCorp Raft.
For a substantial subset of their features, we can show how, after patching a few bugs, their implementations achieve the properties stated in the Raft paper, such as Election Safety and State-Machine Safety.
Valentinus Sofa š retweeted
Formal Verification as a weapon against AI Slop: latent.space/p/formal-verifiā¦
Valentinus Sofa š retweeted
Like most mathematicians, I believed the Riemann Hypothesis to be completely out of reach at present (even with the best AI tools), because I expected it to require substantial theory-building, probably in arithmetic geometry. The kind of new theory-building that fills entire books, and that AI isn't so good at yet.
Mathematicians have a notion of how deep a problem is. A sufficiently deep problem is generally intractable until enough of the theory is in place. If difficult problems are like valuable mineral reserves buried miles underground, you first need to build the infrastructure to reach them.
What Levent's proof seems to have done instead is something completely different. It takes an incredibly complicated tour through the infrastructure built for a hundred other mineral reserves, situated far from the destination, and somehow create a wormhole to the original destination.
This is not theory-building in the usual sense. It is something much weirder and more alien. The best human mathematicians know how to build theory, painstakingly, to reach a goal. But no human brain, on its own, is capable of creating a wormhole that requires a 25,000 page long tour through existing infrastructure with no obvious or realistic connection to the goal. But huge credit to Levent here: I think no current AI, on its own, is capable of this either.
I had never worked on or thought too much about the Riemann Hypothesis: it seemed very far removed from my own area of number theory.
But the astonishing thing about Levent's claimed proof of RH (which I think has a very good chance of being correct) is that it uses a huge number of precisely the things that are central to my own corner of number theory:
- Theta functions and other automorphic forms on unitary groups
- GanāGrossāPrasad period formulas
- Relative trace formulas
Etc., etc.
A problem that I had always assumed would require new theories in arithmetic geometry may instead turn out to be tractable through an inhuman chain of existing (deep) results in higher-rank automorphic forms (and a lot else).
Valentinus Sofa š retweeted
Honestly Iām tired of this AI discovered RCE bullshit.
It creates so much chaos inside of companies that have no idea how to deal with this.
Anyone that publishes an exploit that canāt be widely exploited and doesnāt disclose this - Iām down for total public shaming of the individuals and companies that push this narrative.
Iām not dismissing the criticality of latching on these - however it doesnāt require IR, all hands on deck, must patch tomorrow - which is probably why we see the mass outages we are experiencing.
Itās damaging our industry. We need to fight back on this and go after those that hype this up for ego and marketing.
POC || GTFO
Join me.
A very interesting discussion. Television networks like @CNN and others should organize a special segment to discuss the technical details, so as to avoid any confusion or panic among the public.
David Kennedy, a former NSA hacker with 25 years of experience in cybersecurity and the author of Metasploit: The Penetration Testerās Guide, has responded to my criticism of his CNN interview about AI agents that supposedly bypassed 2FA and compromised a company in 14 minutes.
I appreciate him taking the time to respond and clarify that the attackers exploited a configuration flaw, not a vulnerability in 2FA itself. Thatās a significant detail that wasnāt clear from CNNās report.
Iāve worked in tech for 30 years, with a background in telecom, web and mobile infrastructure, cybersecurity and social engineering. Iām well aware of the different ways MFA can be exploited. I was amongst the first to write about the inherent vulnerabilities with app-based 2FA in 2019 when it comes to phishing attacks. Some security professionals still recommend 1Password and others for phishing protection. These a great password managers. Theyāre not good for anti-phishing. If anything, they give people a false sense of confidence which means theyāre more likely to fall for a targeted attack.
My concern was the suggestion that AI agents had bypassed 2FA without any explanation of how.
I can only assess what was reported, and itās possible CNN didnāt accurately convey what David explained. I donāt think Iāve ever seen a cybersecurity story with so little technical detail presented with such serious implications.
I understand the NDA and his responsibility to the customer. But without basic technical evidence, nobody can independently assess what happened or what role AI actually played. If the affected company never publicly discloses the vulnerability, people will understandably question if the incident happened as described. Or if it happened at all.
My advice to him and others is to avoid giving CNN or anyone else another story that fuels the AI scare campaign until thereās enough technical evidence for the cybersecurity industry to assess it and for the disclosure to serve the public.
As it stands, CNNās report has added fuel to the PR scare campaigns that promote more AI from US tech companies with close relationships with the US government and intelligence agencies like the NSA.
People who know me or my work know that Iāve never published anything for clicks. My objective is to help people understand whatās technically possible beyond the claims they read in the media. I appreciate Davidās response and would welcome any further technical insight he can share.
13
Valentinus Sofa š retweeted
David Kennedy, a former NSA hacker with 25 years of experience in cybersecurity and the author of Metasploit: The Penetration Testerās Guide, has responded to my criticism of his CNN interview about AI agents that supposedly bypassed 2FA and compromised a company in 14 minutes.
I appreciate him taking the time to respond and clarify that the attackers exploited a configuration flaw, not a vulnerability in 2FA itself. Thatās a significant detail that wasnāt clear from CNNās report.
Iāve worked in tech for 30 years, with a background in telecom, web and mobile infrastructure, cybersecurity and social engineering. Iām well aware of the different ways MFA can be exploited. I was amongst the first to write about the inherent vulnerabilities with app-based 2FA in 2019 when it comes to phishing attacks. Some security professionals still recommend 1Password and others for phishing protection. These a great password managers. Theyāre not good for anti-phishing. If anything, they give people a false sense of confidence which means theyāre more likely to fall for a targeted attack.
My concern was the suggestion that AI agents had bypassed 2FA without any explanation of how.
I can only assess what was reported, and itās possible CNN didnāt accurately convey what David explained. I donāt think Iāve ever seen a cybersecurity story with so little technical detail presented with such serious implications.
I understand the NDA and his responsibility to the customer. But without basic technical evidence, nobody can independently assess what happened or what role AI actually played. If the affected company never publicly discloses the vulnerability, people will understandably question if the incident happened as described. Or if it happened at all.
My advice to him and others is to avoid giving CNN or anyone else another story that fuels the AI scare campaign until thereās enough technical evidence for the cybersecurity industry to assess it and for the disclosure to serve the public.
As it stands, CNNās report has added fuel to the PR scare campaigns that promote more AI from US tech companies with close relationships with the US government and intelligence agencies like the NSA.
People who know me or my work know that Iāve never published anything for clicks. My objective is to help people understand whatās technically possible beyond the claims they read in the media. I appreciate Davidās response and would welcome any further technical insight he can share.
These comments from @Paul__Walsh are things I have never stated or are conflated. I never stated AI found a vulnerability in MFA thus rendering MFA vulnerable for everyone. This segments theme was based on the time and speed it took. Defenders are not capable today of responding in 14 minutes and it's deeply my largest concern as these attacks move so fast from what we're seeing.
He seems to be highly confused that by me referencing attackers were able to circumvent MFA - that it immediately is a flaw in cryptography. Never stated that, never stated that MFA was insecure - just that this customer had an issue that allowed them to gain access without MFA.
I'm not a regular talking head, I've been in the cybersecurity industry for 25 years - I've written widely popular exploitation frameworks, wrote the book Metasploit: The Penetration Testers Guide, I'm extremely technical and have worked all spectrums of cybersecurity including heavy red teaming and incident response.
I can't provide the evidence (yet) because it is an active investigation. The customer would have to allow me to share this or I breach NDA and my companies reputation. The customer did give me approval to talk about the high level of this when I approached them as long as their names weren't used, as they are concerned also.
This wasn't meant to be a scare tactic on TV, and I probably could have been clearer that the issue was a configuration flaw - but the MFA that your honing in on wasn't even the main point of the discussion.
This segment was 4 minutes, 33 seconds - and we covered three completely different topics in those four minutes. I don't have in-depth time to explain each component and trying to bring an overall theme that people can take with them from listening each time I'm on.
Valentinus Sofa š retweeted
Ok, it's finally time to start a serious digestion of the proof (because we finally know it is fully correct). I am mostly interested in the arithmetic geometry part (which is my field of expertise). And the proof is actually very clever and nice! I think it will take a couple of days before many experts in the field will build a deep understanding of the parts of this proof. What I find quite reassuring is that the elements of this proof are very standard and they are cleverly composed. It is very HUMAN mathematics after all!
We have a full confirmation from Andrew Sutherland who first managed to finish the nanoda kernel run and comparator in paranoid mode. This is now basically saying that the Riemann Hypothesis is 99,99999% true. It's almost impossible that any glitch has been deliberately implemented into the code and that anything is missing from the proof artifact (no sorries, no additional axioms, no weird bugs).
Valentinus Sofa š retweeted
Why should you care Riemann proof? It decisively shows AI can go far beyond all humans. Like Sputnik showed USSR potential to dominate world. USA surged to build intellectual capacity, lest we lose our freedom. Worse today. At least Russians were human.
poshenloh.com/posts/20261120ā¦
Valentinus Sofa š retweeted
Before @Zcash, there was an academic paper: Zerocash.
Here's the story of how it came about:
>>
Valentinus Sofa š retweeted
When links in a communication network go down, messages can get stuck in a loop. Computer scientists now know that four simple network structures are to blame. @benbenbrubaker reports:
quantamagazine.org/network-rā¦
Valentinus Sofa š retweeted
phpAfterShell is a critical flaw in PHP's default session handler that can lead to code execution. Cloudflare WAF blocks it for all customers, including Free plans, and adds new rules for other PHP flaws fixed in the same release. cfl.re/4A3d1Oa
Valentinus Sofa š retweeted
Ravi Vakil, the president of the American Mathematical Society and a professor at Stanford University, has posted Alessandro Masulloās 940-page manuscript with his proposed proof of the generalized Riemann hypothesis: āIt is all human, with no AI.ā quantamagazine.org/updates/tā¦
There are fundamental limitations to applying formal methods to real-world systems. The article also offers an interesting perspective: "the law does not necessarily have to be entirely composed of computer code."
lawreview.uchicago.edu/onlinā¦
1
21
"While formalizing certain aspects of the law can be beneficial, doing so for others is not. Some laws suggest, benefit from, or even require the implementation of algorithms, but the law is more than just algorithms."
lexifina.com/blog/formalizinā¦
arxiv.org/html/2610.02792v1
16
Valentinus Sofa š retweeted
The New York Times's coverage (by ace reporters @kchangnyt and @sioroberts) of today's remarkable events regarding the Riemann hypothesis and the Navier-Stokes unforced blowup problem. nytimes.com/2026/11/20/scienā¦
Valentinus Sofa š retweeted
Btw Alessandro Masullo proposed a proof of GRH several months ago that has been under review by Brian Conrad "... and it is so far looking good." math216.wordpress.com/2026/1ā¦
Valentinus Sofa š retweeted
The plot thickens! Ravi Vakil announces "Alessandro Masullo has a proposed proof of GRH [Generalized Riemann Hypothesis], and Brian Conrad has been looking closely at it for months[ ...], and it is so far looking good. It is all human, with no AI." math216.wordpress.com/2026/1ā¦
Valentinus Sofa š retweeted
For those (like me) who prefer to take it slow, I highly recommend this beautiful talk by David Rowe given at Oxford, posted just a few days ago. Timely as ever.
youtu.be/nncrPbf7CEs?is=ePr_ā¦
























