I'm really excited to release 2 major Holochain apps today.
Both free, open source (MIT), and out now on Linux, Mac and Windows.
Flowsta Vault — your identity, on your machine.
Your keys and recovery phrase never leave your device; we never see them. It runs a local Holochain conductor, so just by running it, you're not only using the network — you're strengthening it. Your data stays on your machine and exports in full. You own it, not us. Never locked in.
ProofPoll - verified polls, no central control.
It's the first Holochain app built on Flowsta Vault. Real voting where one vote = one real person (sybil-resistant via your @WeAreFlowsta identity). No server to seize, no admin who can rewrite the count. It even keeps private data on a public network — encrypted client-side so only you can read it.
But here's the part I'm most excited about: ProofPoll is built to be forked.
A review site, a task tracker, a social feed — the voting bits swap out easily. The genuinely hard parts of a desktop Holochain app — conductor lifecycle, DNA migration, identity linking, and encrypted data - are all solved, commented and documented inside it. We've spent 8 months building on Holochain nearly every day, and everything we've learned is baked into ProofPoll. You don't start from a toy example — you start from real, production know-how.
So if you've ever wanted to build on Holochain, my advice is, don't start from a blank page. Clone ProofPoll, open it in Claude Code, point it at the repo + our docs, and tell it what you want to build. We've watched a working app come together this way in an afternoon.
Full write-up: https://t.co/nfi6DwIACF
Major Holochain app launches: two desktop apps, out now — free, open source, on Mac/Windows/Linux.
🔐 Flowsta Vault — your identity, on your machine
🗳️ ProofPoll — verified voting, the first app on Vault, and built to be forked
Full story 👇 https://t.co/GG3OU70UUQ
Spot on. Knowledge was built collectively. We’re building the alternative: Flowsta Sign It.
Holochain cryptographic proof of authorship + machine-readable consent for your art and creations.
So people can actually recognize and reward your value.
do you understand what just happened to your computer..
Google Chrome secretly downloaded a 4GB AI model onto your device. Without asking.. Without telling you..
It's called weights.bin. It lives deep in your system folders. It powers Gemini Nano - Google's on-device AI.
And if you delete it? Chrome re-downloads it automatically. Like nothing happened.
Just Google deciding your hard drive is their storage unit.
At 1 billion Chrome users - that's 4 BILLION gigabytes of data pushed silently across the internet.
The carbon footprint alone equals tens of thousands of cars running for a year.
Check your disk right now:
📁 %LOCALAPPDATA%\Google\Chrome\User Data\OptGuideOnDeviceModel
To stop it: chrome://flags → disable Optimization Guide On Device Model → restart Chrome → delete the folder.
Reshare so people know what's sitting on their computers.
I watched this World ID 4.0 presentation Sam Altman & co put on the other day. A few people have asked what I think, given the work I do on identity and proof of humanity. Honestly, my reaction is more layered than I thought it might be.
At face value, it actually does a lot right. Open protocol. Zero-Knowledge proofs. Biometric data that stays on the user's device. These are genuinely good design choices.
There are scenarios where what they've built could useful. Iris biometrics are strong. Hardware attested verification has anti-Sybil properties that pure software can't easily match.
But proof of humanity is too important to leave to any single scheme. There are some structural concerns I keep coming back to.
The trust chain has unverifiable links. The orb hardware and firmware are open source in design, but you can't verify the orb in front of you actually runs that source. Same shape problem as voting machines.
The app is its own version of this. The thing actually handling your biometric data on your phone isn't open source, as far as I can tell. It could be — and if it were, that would be a real step forward. Without it, you're trusting a closed binary the same way you're trusting a closed orb. And even where biometric algorithms are published openly somewhere, that doesn't help on its own. What matters is the binary you're running. Without reproducible builds and published checksums, there's no way to verify the code you can read on GitHub is the code actually executing on your device. "The algorithm is open source" is not the same as "I can verify what's running."
Issuance is centralised. Only World ID orbs issue credentials. One organisation gates humanity-as-a-credential globally. For something this load-bearing, that's a very big single point of trust - even if the team holding it has the best intentions today.
The proof is device-bound, which makes recovery brittle. Lose your phone, lose your humanity. There's no graceful path back - you have to find another orb and start over. For a system meant to be permanent identity infrastructure, that's a serious failure mode for normal users.
And one global ID creates correlation pressure even with zero-knowledge proofs. The fact that two pseudonyms share the same World ID is itself revealing. A single credential presented across services is structurally different from per-context identifiers backed by the same underlying verifications.
What I think matters more than any specific scheme is the principles. No single party should gate humanity globally, including any company, including any government, including what I do with @Holochain and @WeAreFlowsta . The user, not the issuer, should control what's revealed to whom. Recovery should be graceful, not "find another orb." And hardware attestation is a useful input, but it shouldn't be the answer.
So, where does this point? For me, proof of humanity probably shouldn't be one credential at all. It's a portfolio of signed attestations a person assembles over time from sources they trust. Peers who know them, groups they're part of, third-party verifiers, even hardware attested proofs where those make sense, and present selectively to whoever is asking. No single issuer. No centralised global ID. Recovery happens through your social graph if you lose everything.
That shape solves things a single issuer system can't. It also lets schemes like World ID slot in as one valid input among many, rather than positioning any one of them as the answer.
"Who is a human?" is one of the most important infrastructure questions of the next decade. It's too important to be answered by any one source.
There's a point where there's no guard rails on AI.
Whether it's a nation state, or a kid in their bedroom. You never know where something unexpected could come from.
Anthropic drip feeding Claude Mythos is an important step in transitioning to that future.
All the software we use and depend on everyday is full of security issues.
The issues have been found by Anthropics new Claude Mythos AI.
A very responsible move by Anthropic to make their new AI available in stages. Although I wish I had access now.
The government can buy your personal data, track where you go, and access your financial records, no warrant required.
Not long ago, that was illegal.
So what changed? Our new video breaks it down.
This actually protects children and empowers the parents to say what is and isn't good for their kids.
We need these kinds of solutions instead of Orwellian age verification.
Let the parents choose.
Your Digital Life. Your Choice.
✅Sovereign Identity.
✅Your Data is Yours, as is the Choice in how it's used.
✅Made possible by Holochain.
➡️Welcome to the New Internet.
Own Your
Own AI
Or It Will
Own You.
New York wants to hold AI companies to NOT allow “protected professions” questions and results.
“For safety”.
Yes safety of persevering those professions.
The slippery slope in the 5000 Day Interregnum.
Soon 100000000 professions…