Patrick (@patrickwoodhead) header
Patrick (@patrickwoodhead) profile picture

Patrick

@patrickwoodhead

Head of Product @confidentialdot. Ex @protocollabs. Ex IBM. Rational Flâneur.

 London, England
geomatrick.io
Joined October 2011
  • Tweets587
  • Following3K
  • Followers3.1K
End to end private inference on the most popular open weight model. No one can see your data through the whole flow.
Confidential AI (@confidentialdot) profile pictureConfidential AI
@confidentialdotNov 2
DeepSeek V4.1 Flash now live on Confidential Inference.
1
1
48
It would feel good to be a B300 right now.
1
1
105
Looking for between 8 and 16 B300 nodes, available ASAP, US preferred.
10
5
278
This is why cryptographic attestation the end user can verify is the only way. A retention policy is a promise - trusting the infrastructure provider not to look is not good enough.
Apparent Order (@apparentorder) profile pictureApparent Order
@apparentorderOct 1
AWS Bedrock – data retention mode "none": Docs: "Zero data retention. No request or response data is written to durable storage by AWS" Blog: "We may store and review flagged content ... even when mode is none" Only one of these can be correct.
1
76
If confidentiality comes with the same performance, it’s strictly better. confidential > not confidential KV cache offloading now gets the same gain with CC on as off. Great work @b0xtch.
Confidential AI (@confidentialdot) profile pictureConfidential AI
@confidentialdotOct 15
1/ Breakthrough: First production KV cache offloading implementation under confidential computing (CC). Same gain as without CC. NVIDIA B300s, CC on: throughput: +141% time to first token: 39 s → 1.5 s GPU to host copies: 4 GB/s → 50 GB/s
Image from Confidential AI (@confidentialdot)
2
77
I'm glad I studied mathematics before AI.
2
63
love inference? love scale? confidential AI is hiring an AI inference engineer to scale confidential computing to 8 billion people. and 800 billion agents. confidential.ai/careers/ai-i…
Confidential AI ･ AI Infrastructure Engineer

Confidential AI ･ AI Infrastructure Engineer

Confidential AI builds the confidential computing stack for AI.

confidential.ai
2
2
7
297
The SF billboard was a success. So we upped our marketing budget.
Image from Patrick (@patrickwoodhead)
Patrick (@patrickwoodhead) profile picturePatrick
@patrickwoodheadSep 29
Our SF billboard
Image from Patrick (@patrickwoodhead)
1
1
5
163
Our SF billboard
Image from Patrick (@patrickwoodhead)
4
1
2
12
808
Confidential Inference. Don't let anyone see your data.
1
2
1
7
420
I like this clarification of names re privacy and AI
aljaz (@aaaljaz) profile picturealjaz
@aaaljazSep 28
The spectrum of privacy in AI
Image from aljaz (@aaaljaz)
81
I had the good fortune to get a tour of the facility last month. Came away super impressed. The team at C5R are geniuses.
C5R CORP (@c5rcorp) profile pictureC5R CORP
@c5rcorpSep 24
In 12 weeks, we built a research facility that is run entirely by AI. AI designs, executes, and observes experiments end-to-end across biology, chemistry, and materials science. We’re introducing SciUniverse: a benchmark that measures AI’s ability to do real-world scientific research.
2
406
POV: you're a prompt going through confidential dot ai private inference.
1
1
1
9
44,847
Confidential Kubernetes (c8s): v0.12 to v0.33 in three weeks. Some highlights: - env and mounts pinned at admission - GPUs fail closed unless in CC mode - post-quantum attested channels - one CVM per k8s node - sealed nodes, no shell into running workloads github.com/confidential-dot-…
13
1
1
17
447
We open-sourced confidential-inference. The code behind api.confidential.ai: gateway, serving images, allowlist. Every production deploy ships as a signed release bundle. Read the code, check the digests, verify the cluster. github.com/confidential-dot-…
1
3
10
31,048
Patrick retweeted
aljaz (@aaaljaz) profile picture
aljaz
@aaaljazSep 22
new provider (@confidentialdot) added to confidentialinference. net
Image from aljaz (@aaaljaz)
1
2
1
11
556
Open weight models are having their time - and they are a much better fit for confidential computing and inference too.
Guillermo Rauch (@rauchg) profile pictureGuillermo Rauch
@rauchgSep 19
Looks like today may be a record day for token volume % of open models on Vercel AI Gateway: 🟦 Open 78.4% 🟨 Closed 21.6% While spend 💲 usually tells a different story, #3 and #4 today are Moonshot AI & DeepSeek. Adding Z⁠.ai, their combined spend surpasses OpenAI (#2). (Do note that's the spend for inference of the model across providers (mostly in the US), not revenue going directly to the open weight labs.)
Image from Guillermo Rauch (@rauchg)
1
3
130
We're hiring at Confidential AI: 🏗️ AI Infrastructure Engineer ⚡ AI Inference Engineer 🔐 Low-Level Systems & Security Engineer We're building end-to-end privacy for inference. confidential.ai/careers

Confidential AI ･ Careers

Open roles at Confidential AI, building confidential computing infrastructure for AI.

confidential.ai
3
1
1
7
379
You can now verify a live Confidential Kubernetes (c8s) cluster from your browser. Every check runs in your browser tab. Attestation chains to Intel's TDX root. Firmware, kernel and rootfs match the audited image. A pinned CA proves it is the expected cluster. Nvidia GPU attestation included too. confidential-inference-clust…
Image from Patrick (@patrickwoodhead)
1
3
250
Mathematicians: you can get fully private inference and agents at confidential dot ai
4
49
Load more