Polyhedra
@PolyhedraZK
Followers
902K
Following
1K
Media
428
Statuses
1K
The Future of Intelligence • Inventor of zkBridge and Expander • EXPchain: The everything chain for AI
Joined January 2023
Hey @DeFiIgnas this is exactly the problem we’re building 𝗢𝗰𝗮𝘀𝗵 to solve. Private by default, legally compliant, no Wallet → CEX → Wallet detours. Would love your thoughts and to connect on this.
If I want to avoid doxxing my wallets sending USDC I do this: Wallet1 → CEX → Wallet2 I can’t use mixers (straight to jail!) Who's building something legally compliant without KYC?
10
4
24
Noticed privacy is the first thing listed by @a16zcrypto. If crypto is going to be real internet-scale money by 2026, privacy can’t be an add-on; it has to be the default. With @opencash_system, we’re building toward that future: private by default, compliant when needed, and
It's time for our annual big ideas. Here are 17 things that various a16z crypto partners (plus a few guest contributors) are excited about for what’s ahead in 2026. On topics ranging from agents and AI; stablecoins, tokenization, and finance; privacy and security; to prediction
5
7
26
4/ What’s next — and how the community can get involved: Sam also outlined how the community can participate as Ocash evolves: • An active core group on Discord • Ongoing shows each week and more sessions planned for the new year • An “OChristmas” campaign coming soon to
0
0
9
3/ What makes Ocash different: Ocash is not a mixer. Unlike Tornado-style systems, Ocash hides execution traces using ZK proofs while keeping a UTXO-style structure underneath. Rather than mixing funds, it conceals how transactions are executed. That design enables: • Selective
1
0
12
2/ Privacy at scale isn’t trivial: One of the first scaling points TC highlighted is transaction flow in a privacy setting. In private systems, users can’t simply broadcast transactions themselves without leaking information. Instead, transactions move through relayers, which
1
0
3
1/ What shipped: Ocash Mainnet Alpha went live this week, shaped directly by real testnet usage. TC shared that testnet usage surfaced a large number of bug reports. In response, the team chose to rewrite the codebase ahead of Mainnet Alpha. After cleaning up the code and
2
1
3
This week on The Polyhedra Show, @sampearton sat down with @Tiancheng_Xie (TC) to unpack the launch of Ocash Mainnet Alpha, and what it takes to bring private transaction infrastructure into real-world use. Some key takeaways ⬇️
4
3
21
5/ As AI becomes the layer users rely on for facts, accuracy is only the baseline. Institutions need AI systems where every conclusion can show its work. That’s exactly what @PolyhedraZK is building: AI that must verify its own reasoning before its answers ever reach the public.
0
0
6
4/ zkML gives platforms a way to enforce accountable outputs: ✔️ Confirms the model version responsible for each statement ✔️ Confirms which evidence contributed to each claim, without exposing underlying files ✔️ Confirms whether reasoning steps followed required constraints ✔️
1
0
6
3/ Most AI-generated search results provide zero transparency into how they were assembled: - Which documents influenced the claim - Whether the model merged unrelated cases - Whether unsupported leaps occurred - Whether autocomplete reinforced a false narrative For businesses,
1
0
2
2/ The issue wasn’t presentation. The outputs looked authoritative, complete with citations, summaries, and references to news articles and government press releases. But reviewers found that some “sources” never mentioned the company at all. When AI shapes real-world decisions,
1
0
2
Why zkML? Because Google’s AI-generated search summaries invented a lawsuit against a U.S. electrical contractor, and attributed it to a state attorney general despite no such case ever existing. The company discovered the fabricated claims back in 2024, filed suit in March
4
4
16
6/ As hiring, work, and collaboration move fully online, identity must be verifiable, human, and secure by design — not dependent on fragile checks and vulnerable databases. With Polyhedra i-D, you can authenticate truth without revealing yourself. Try
id.polyhedra.network
Data-protective identity verification using zkML technology
0
0
8
5/ That’s why we built Polyhedra i-D. Our zk-powered face verification proves a real human is in front of the camera — without storing biometrics, and without exposing identity. AI can fake faces, but it can’t fake cryptographic proof.
1
0
8
4/ These systems fail for a simple reason: Traditional verification relies on stored biometrics, weak liveness checks, or simple video presence. Deepfakes can now bypass all three. We’re entering a world where seeing someone on video doesn’t mean they’re real.
2
0
3