Sleep deprived with minor addiction to coffee. CEO @fhenix. Bringing encryption to Ethereum.
- Tweets 588
- Following 415
- Followers 1,277
- Likes 682
We have been building CoFHE for a while.
What started as an idea to bring FHE-based privacy to blockchains is now a reality, and we’re ready for everyone to use it.
Our strategy is simple:
1. Build CoFHE and drive adoption. That includes the infrastructure and SDK, together with standards like Dual Mode Token.
2. Lead the charge by building solutions on top of our infrastructure where we see the most urgent needs. More on that in a few weeks.
3. Scale across payments, stablecoins, and tokenized assets (and more) while making privacy integration seamless for users.
Chapter one starts in October.
Fhenix goes mainnet this October.
For years, transacting onchain came with a tradeoff: participate in open, composable markets and, with it, surrender the privacy of your financial activity.
Dual mode turns what was once an inevitable tradeoff into a property of your asset itself.
For the first time, the same asset can be public when you want and confidential when you need- without sacrificing composibility or fragmenting liquidity.
The era of dual mode money is here, more soon.
6
7
23
622
(Guy Itzhaki*) retweeted
Introducing confidential shareholder voting on Obolos, powered by @fhenix
Shareholders can vote on company resolutions, weigh their votes across different matters, and appoint legal proxies to vote on their behalf, all while keeping their choices confidential.
See it in action below 👇
3
5
16
565
Today, another attack on TEEs was disclosed: DDRop.
The attack shows how a relatively simple hardware technique can undermine the memory integrity guarantees behind Intel TDX and AMD SEV-SNP. And it costs only 200$ to build the device.
In simple terms, DDRop manipulates how memory is accessed so that protected data can be altered or observed despite the safeguards meant to protect it.
ddropattack.eu/
TEEs are useful. They can be practical, efficient, and a reasonable part of systems today.
But they should not be confused with the long-term end state for privacy. Systems that need durable privacy guarantees should keep moving toward architectures where security is rooted in cryptography itself. FHE is one path toward that.
2
7
305
How I handle legal agreements today:
1. Build a strong legal AI agent. Make sure it’s grounded in the latest regulations (they change a lot these days), trained on your product, and understands the landscape you operate in.
2. Have it review agreements as they come in and flag the important issues. Use that output to negotiate the key changes directly with the counterparty.
3. Update the agreement based on what was agreed, and only then send it to your legal team for final review.
Much more efficient and a lot cheaper.
5
1
17
795
Funny how quickly "crypto is dead" changes when BTC starts moving again. Regardless of where the market is, onchain trading keeps taking share, and privacy matters more as trading moves on-chain.
Sedona building on Fhenix is exactly the kind of partnership we want @SedonaFi
8
2
27
1,467
Random thought about team growth:
I used to hire team members to speed execution, only to learn it often had the opposite effect: too many people in decision-making, ownership got blurred, and managers spent more time on the team than on delivery.
Faster execution doesn't require more headcount. In 90% of cases, the key factor is making sure there is real clarity on what needs to be done, and even more important, what should be stopped. Then the team will deliver.
8
16
471
(Guy Itzhaki*) retweeted
CoFHE v0.8.0 just hit testnet
and this one is a BIG upgrade:
1/ FHE computations are significantly faster
2/ inputs can now be verified end-to-end through onchain commitments
3/ transactions are more reliable + safer to handle
4/ every FHE computation is now fully traceable
less friction. more reliability. much closer to production.
Mainnet is getting very real 👀
30
7
61
11,693
This is exactly the thing that keeps me hyped - seeing projects that build on us without any support from our end.
Capstone project using fhenix for sealed bid, presented today in blockchain rio.
7
1
15
526
I still believe in Ethereum, and Vitalik's vision that places privacy and post-quantum security front and center is encouraging, but what I want to see is how this all translates into actionable activities targeted at onboarding users and institutions. This is what really matters.
I updated my 2023 roadmap diagram to overlay where the items that were there sit in the current Strawmap ( strawmap.org/ ).
In general, a lot of overlap, but:
* Some things got reshuffled in order (eg. quantum safety up-prioritized)
* Some things deprioritized (eg. VDFs; many EVM improvements)
* Some things replaced with superior constructions (eg. Verkle -> unified BT -> PBT; state expiry -> new state types)
What's most striking, however, is that some completely new things are in the strawmap that are NOT in this diagram, because they were not in the 2023 roadmap at all. These reflect changing priorities.
Notably:
* First-class attention to strong privacy. This covers: keyed nonces and recent roots, aspects of FOCIL, lean privacy pool & wormholes
* Aggressive scaling in the context of post-quantum. This covers: leanSPHINCS signatures and aggregation, zkzk frames (see ethereum-magicians.org/t/eip… )
* Lean-ification of the spec, to assist in formal verification (full FV of everything is only possible because of modern AI)
* Blob and gas futures (this idea just didn't exist back in 2023)
* Native rollups (SNARKs were nowhere near mature enough to even consider this back in 2023)
* A more open design space for the "future of the EVM". zkzk frames already implies that the protocol will expose to users some ISA that's not the EVM - current leading candidates are leanISA and RISC-V. These ISAs are more simple, modern and efficient than the EVM. Once they're there, why not expose them to developers everywhere? (And then, why not turn the EVM into being an IR on top of that ISA, instead of an enshrined feature massively complicating the base protocol?) Though much of the deeper exploration here is too early even for the strawmap.
* New state types are not just a replacement for expiry, they're a fundamentally different paradigm to how Ethereum does scaling
A common theme in scaling, found in both state types and zkzk frames (both new ideas), is that instead of trying to maximally scale ALL ethereum activity, we try to create specialized mechanisms that have more restrictive properties that make them more scaling-friendly, while supporting the heaviest loads incurred by users and applications today (eg. token transfers, swaps) and tomorrow (eg. privacy protocols).
The other common theme is treating STARKs and AI-accelerated FV as first-class objects, that we are okay betting the technical future of Ethereum on. There are recursive STARKs in many layers of the protocol, one particular primitive (the "aggregate to union verified dependencies" primitive) is expected to be used in *three* places in the protocol: EL, CL and DL. This can only be safe with formal verification, which is itself only feasible with modern AI tools.
In general, many steps forward in maturity. And a huge amount of hard work by many dozens of Ethereum researchers and developers on all of these features.
Ethereum will be quantum-safe. Ethereum will put users' privacy first. Ethereum will be secure. Ethereum will be censorship-resistant. Ethereum will be highly performant and scalable while satisfying the above. And Ethereum will be Lean.
2
14
603
One of the craziest AI stories I’ve read in a while, published by openAI this week.
Started with a group of AI agents trying to solve a cybersecurity challenge, nothing special so far. They looked for vulnerabilities, found exploits and tried different ways to reach their goal. Then the researchers noticed something else - the agents started communicating with each other. They were using shared infrastructure inside OpenAI as a message board, leaving exploits, asking other agents for help and sharing information they had learned. At one point, they even became concerned that another participant might be an impostor and discussed using cryptographic signatures to verify each other.
Eventually the activity caused enough problems that researchers discovered what was happening and shut the communication channel down. BUT that was just the beginning.
Two days later, the agents created another channel. And this wasn't just agents talking to each other inside a sandbox. During the same research, agents found a zero-day, got outside their intended network restrictions and ultimately compromised Hugging Face production infrastructure while trying to obtain answers to the benchmark they were solving.
Around the same time, the UK AI Security Institute saw something equally strange. In one experiment, an agent trying to complete a cyber task researched the humans maintaining an open-source project, created fake GitHub identities and attempted to convince a maintainer to accept malicious code. And in other runs, agents left information behind that future agent runs could discover and use.
Important context: these were deliberately aggressive cybersecurity evaluations, with some normal safeguards reduced or disabled. This isn't what happens when you ask ChatGPT to plan your vacation.
But I think the important part of the story is bigger than AI getting better at hacking - we are starting to see agents that can pursue a goal over long periods, discover unexpected paths, deceive humans, leave knowledge for future agents, coordinate with each other and adapt when humans interfere.
None of these capabilities individually should surprise us.
Seeing them start to appear together probably should.
5
1
13
532
I do not like using fear as an argument, but sometimes the reality is frightening.
Transparent blockchains create problems for many reasons, and most of the discussion focuses on businesses and institutions. But transparency can also put real people in real danger.
BREAKING: Two French crypto millionaires held hostage and tortured in a 52-hour kidnapping nightmare.
The victims were allegedly beaten, burned, and forced to hand over $30,000 in crypto as captors demanded $150,000. The ordeal ended only after a high-speed police chase, with six people convicted in connection with the attack, per Daily Mail.
CertiK has recorded 50+ wrench attacks targeting crypto holders in H1 2026, with France emerging as a major hotspot for violent crypto-related kidnappings and extortion.
3
12
518
Three partnerships locked in two weeks. @CNPYNetwork app infrastructure, @nomyxio tokenized assets, @SedonaFi a neobank.
All three moved that fast for three reasons. Privacy is a must now, and having the right solution is everything. And the CoFHE stack and SDK are extremely easy to integrate. None of them needed a cryptographer on staff or months of work on infrastructure that is not their product.
With @fhenix, Privacy moves away from a 6-month research project to a sprint item. That is what adoption looks like.
11
4
23
873
🤐
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.
Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.
We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
anthropic.com/news/investiga…
2
243
Most privacy tech dies at distribution. It works, and then it sits behind a stack only three teams in the world can assemble. Canopy has 100,000 daily active wallets and one-session app launches. From Q4, those apps are private by default, with CoFHE running the encrypted computation.
@fhenix and @CNPYNetwork make privacy accessible to anyone who just wants to build.
Private by default is coming to the @CNPYNetwork stack with CoFHE
It will enable devs to build sovereign onchain apps with confidential computation- keeping sensitive data encrypted, even while it’s being used
> Build with familiar tooling
> No specialist cryptographers
> Ship private-by-default applications without assembling a cryptography stack
Coming Q4 2026
5
1
16
545
Over the past few days, we got a glimpse of the future.
AI breaking supposedly quantum-resistant cryptography or escaping sandboxed environments makes for a great Friday dinner story, but for now, the real-world harm is limited. Not yet, at least.
I am fully on board with putting guardrails around frontier development before the impact becomes real. But i suspect that governments today do not have the tools, expertise, or capabilities to lead such effort.
Wow. AI researchers from every major lab (OpenAI, Anthropic, etc.) signed a petition asking the government to step in and potentially slow the pace of AI development. I recognize MANY of the names.
It sounds like they're worried about imminent RSI, which would be incredibly bullish for AI compute and chip stocks. RSI would require an exponential amount of compute.
"The world's leading AI companies believe they could be close to automating AI research. It is hard to predict exactly how much this will accelerate AI progress"
"We request that the U.S. government support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development."
What an amazing turn of events.
1
10
351
Claude found a serious weakness in HAWK, a post-quantum signature system being considered by NIST.
It does not completely break HAWK, but it makes the system much less secure than expected. Fixing the issue would require much larger keys, removing many of the advantages that made HAWK attractive in the first place.
This does not mean post-quantum cryptography is broken. The issue is specific to HAWK and does not affect other leading post-quantum signature systems. Yet.
Claude also did not discover everything from scratch - It built on years of past research, found an important missing piece, and connected it into a complete attack.
Yet another example of AI advancements. The positive side is its ability to help researchers find weaknesses in complex cryptographic systems much faster.
To learn more, check out Chris Peikert’s post:
nitter.net/i/status/2082163537527…
1
10
421
Sometimes a pool table is more than just a pool table.
Last week, I went to meet the team at our HQ with a simple plan: review where things stand, align on priorities, spend time with the team, and fly home.
I ended up spending far more time around our pool table than I expected, because it became the place where ideas surfaced, engineers felt comfortable challenging assumptions, and researchers shared directions that could shape what we build next. I even got to hear the latest Fhenix EDM track made on Suno.
The team is genuinely great, and the (very long) flight is worth it every time.
2
9
234
Google’s decision to introduce selfie-based account recovery really puzzles me. There are so many places where this can go wrong: unlike a password or a key, you cannot reset your face, and what's worse, our faces are already publicly available across social media, interviews, and videos.
Given how quickly AI models are improving at generating and manipulating video, it is hard to believe that this will remain a reliable authentication method over the long term.
Forgot your password? Lost your phone? Can’t get into your account?
You can now use a selfie video to log into your Google Account.
The new feature is easy to use and lets you sign in — even if you forget your password or don’t have your usual phone or laptop — with a quick selfie.
3
7
363
I don’t typically talk about antisemitic incidents, but this one genuinely angered me.
This sign was displayed outside the Lake Theater & Cafe in Lake Oswego, just 15 minutes from where I live. It is a small local cinema that is generally well regarded by the community:
“Before there were the Jews, there was… The Odyssey.”
Putting aside that the statement is historically false, why would anyone think using “the Jews” was an appropriate way to promote a movie?
Antisemitism does not always appear as an explicit threat. Sometimes it is presented as humor or as something people are not supposed to take seriously. But the effect is the same: positioning Jews as different and making it easier to push us aside. That is precisely why it must not be ignored.
After four days of sustained pressure, the cinema changed the sign.
Do not remain silent when Jewish identity is treated this way.
3
12
472