geomatrick 🌐 👷 💻

Head of Product @Confi_AI. Founder @space_meridian. Ex @protocollabs. Ex IBM. Rational Flâneur.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

Our team at @Confi_AI shipped Confidential Agents today. e2e private agents backed by hardware encryption. Ideal for enterprises in regulated industries.

[Confidential AI @Confi_AI]

Teams in finance, healthcare, and working on privacy-sensitive tasks with AI get stuck with clunky on-prem workarounds.

Why? Because the only thing protecting their data in a cloud is a legal agreement. That’s not enough.

We're fixing that today with Confidential Agents. 🧵

[geomatrick 🌐 👷 💻 @patrickwoodhead]

@iamnotnicola https://t.co/GPRfL4rcMh

[Confidential AI @Confi_AI]

Teams in finance, healthcare, and working on privacy-sensitive tasks with AI get stuck with clunky on-prem workarounds.

Why? Because the only thing protecting their data in a cloud is a legal agreement. That’s not enough.

We're fixing that today with Confidential Agents. 🧵

[geomatrick 🌐 👷 💻 @patrickwoodhead]

What's really cool about confidential inference: you actually get stronger security guarantees than on-prem deployments, while enjoying off-prem convenience.

[Confidential AI @Confi_AI]

This is what actually-private AI inference looks like.

Prompt encrypted on your device, decrypted only inside a hardware-attested enclave. Cloud, hypervisor, host OS, vendor: all see ciphertext.

Two places hold plaintext: your device and the enclave. Everything else is locked out by hardware.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

@iamnotnicola We have shipped this at @Confi_AI

[geomatrick 🌐 👷 💻 @patrickwoodhead]

New paper from the team at @Confi_AI: Kettle, Attested Builds.

Run the build inside a TEE. Staple a hardware-signed attestation of the pipeline, inputs, and outputs to the artifact.

Provenance down to the git commit.

https://t.co/PwrzwnwhSU

[Confidential AI @Confi_AI]

Accepted into NVIDIA Inception.

Confidential runs AI inside TEEs. Private inference. Private weights. Private training. Verifiable end to end.

Now with closer ties to the GPU stack underneath it all.

#NVIDIAInception

[geomatrick 🌐 👷 💻 @patrickwoodhead]

https://t.co/XwHsedyG3E

[geomatrick 🌐 👷 💻 @patrickwoodhead]

AI workloads on shared infrastructure have a three-body problem.

Artifact owners, compute providers, and end users all need confidentiality from each other.

Today my team at @Confi_AI published C8s, a confidential Kubernetes architecture that solves it.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

If your agents touch sensitive customer data, internal tools, prod systems, or proprietary prompts, this matters.

Anyone building agents for finance, healthcare, government or legal should take PrivateClaw for a spin.

[PrivateClaw @PrivateClaw]

Your AI agents are running on someone else's computer.

And that someone can read everything.

Today we're changing that. PrivateClaw is live.

Privacy-first AI agent hosting. Hardware-verified.

https://t.co/xPgDz9HUbu

[geomatrick 🌐 👷 💻 @patrickwoodhead]

Thrilled to get to work closely with the great @AnomalRoil

[ansgar grunseid @ansgargg]

✨ im thrilled to share that the swiss ritter of randomness, @AnomalRoil, joins our merry band of misfits at @Confi_AI 🎉

he and batman have never been seen in the same room at the same time. just sayin'.

[Juan Benet @juanbenet]

Excited to launch a new podcast dedicated to conversations on the future of neurotech, computing, intelligence, and more.

First guest: @maxhodak_ founder & CEO of @ScienceCorp_, which is building PRIMA, a retinal prosthetic that’s restoring meaningful vision for patients with blindness caused by age-related macular degeneration.

Science is also developing a biohybrid brain implant that grows living neurons directly onto a silicon chip, then interfaces that system with the cortex. The possibility space here is vast and new. Imagine growing new areas of the brain.

Sections
00:00 What counts as neurotech?
01:45 History of brain-computer interfaces and the smartphone dividend
07:25 PRIMA - How Science is restoring vision in blind patients
10:10 Why stimulating bipolar cells works when the optic nerve doesn't
30:30 Are we bottlenecked by biology or engineering?
32:40 Expanding the brain's bandwidth beyond 10 bits per second
37:00 Can we add new areas to the brain?
37:46 Biohybrid BCIs: neurons growing on a chip
39:20 What could neural augmentation look like?
01:13:20 How Science drives Fast R&D
01:44:00 How founders learn and level up

This is the kind of discussion I’m excited to explore on this podcast. Enjoy!

Full Episode 1 here and in links below.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

I'm in SF this week around HumanX. If you're thinking about private inference, secure model deployment, or confidential computing, I'd love to talk about what we are up to at @Confi_AI

[geomatrick 🌐 👷 💻 @patrickwoodhead]

@LunalDotDev @NVIDIAGTC Nvidia move into open source agentic with Nemoclaw

[geomatrick 🌐 👷 💻 @patrickwoodhead]

TEEs are the answer. Lunal makes deploying AI onto Trusted Execution Environments trivial. The first lab to offer private inference as a service tier will trigger a domino effect. If you're building in this space or need this now - reach out. https://t.co/43H2Cr35uH

[geomatrick 🌐 👷 💻 @patrickwoodhead]

AI is the most powerful technology of our lifetimes. Right now it's being built in a way that concentrates power, data and trust in a handful of companies. I've spent my career pushing against that. Today I joined @LunalDotDev as part of the founding team. 🧵

[geomatrick 🌐 👷 💻 @patrickwoodhead]

Three parties all have the same problem: companies won't send proprietary data to LLMs. Labs need their model weights secure. Hyperscalers need both to trust them. Three parties, one shared need: end-to-end AI privacy.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

Full post here 👇
https://t.co/3PbOvmbkT8

[geomatrick 🌐 👷 💻 @patrickwoodhead]

Web3 gave us the right vision for a sovereign internet. But AI agents might finally make it a reality.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

But there is room for improvement. The opportunity now: build AI agents that take sovereignty seriously.

Sovereignty + convenience as co-equal goals. Private inference seems to be the next big unlock.