Head of Engineering @0xPredicate. Building stuff in Go & Rust.
Localhost
Joined April 2007
- Tweets29.1K
- Following1.3K
- Followers4.5K
- Likes2.1K
Pinned Tweet
Announcing "dux": a terminal agent orchestrator for all your needs!
Dux is a terminal UI that lets you run multiple AI coding agents side by side, each in its own git worktree, with full companion terminals, macros, commit generation, and a command palette that probably knows more tricks than you do!
Available and ready to use to install via Homebrew on macOS and Linux (and WSL on Windows, of course). You can run this via an SSH session, directly in your machine, and everywhere where a terminal is. Too much terminal? Mouse support is *chef's kiss*: resize panes, single and double click support, etc!
Link in the first comment!
Proud of this team! Three years of building compliance that runs in real time, right inside the transaction, and we're just getting started.
Thank you @bcap, @Paxos, @1kxnetwork and @tribecap for backing us. Back to shipping now! 😄
We are thrilled to announce our $8M Series A to connect regulated markets to onchain finance.
Predicate enables digital assets to reach every market compliantly. We do this by embedding compliance controls directly into digital assets to block high-risk and sanctioned activity at every transaction.
To date, Predicate has:
- Screened $1B+ in transaction volume
- Evaluated 6M+ transactions against custom compliance policies
- Blocked 100K+ non-compliant transactions
The round was led by @bcap with participation from @Paxos, @1kxnetwork and @tribecap.
Loving the update from "About that" by @AndrewChangCBC ... I would've loved if he would've nudged the even bigger security issue around eMudhra though as it's even crazier!
Patrick D'appollonio retweeted
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
Patrick D'appollonio retweeted
Ok all is ready for Monday now! Come and do stuff on our mainnet! We have an awesome wallet, we are connected to many chains already and maybe there is something like fomo3d
Patrick D'appollonio retweeted
BREAKING: eMudhra exits SSL/TLS certificate business, blames social media instead of incompetence for the same
Patrick D'appollonio retweeted
The important thing with incidents like @eMudhra is to proceed openly and by engaging with the communit… ah crap
Global TLS Intermediate CA incident: We’ve updated our incident page with how it happened, who is affected (one US customer ecosystem), what’s not affected, and what we’re putting in place. We will publish the root cause once our investigation is complete.
👉emudhra.com/en/post/certific…
Readers added context they thought people might want to know
eMudhra minimizes the scope by claiming only "one US customer ecosystem" is affected. In reality, 8 of its CA certificates had catastrophically weak private keys (one being simply the number 4), allowing anyone to forge trusted HTTPS certificates for any website on the internet.blog.google/security/prote…
You know you made it in life when you become a flag for mitmproxy 😂
github.com/mitmproxy/mitmpro…
Oh my God 😂😂😂
Sorry everyone... We're announcing a 0-Day vuln in QUIC that allows anyone to remotely turn off a server running these libraries. I swear we didn't plan this.
foxio.io/blog/header-accordi…
Not you too 😆
An urgent notice from the Railway team
blog.railway.com/p/security-…
Should we start coining a term for how red the day was today due to the massive number of security issues?
How about "patchpocalypse"? 😂
Patrick D'appollonio retweeted
DeepSeek is everything i thought a company named "OpenAI" would be:
Be DeepSeek:
- One of the worlds leading AI labs
- Release model weights with MIT license
- Open-source FlashMLA and DeepGEMM so others can benifit.
- Releases 3FS, your distributed file systems for AI workloads
- Release DeepSeek Harness, the best agent agent harness.
- Now releases DeepInfer, the inference framework powering production services.
Weights. Kernals. Storage. Agent harness. Inference infrastructure.
OpenAI on other hand has released gpt-oss and codex cli.
At this point, shouldn’t DeepSeek be the one called "OpenAI"?
People keep calling DeepSeek a model company. Look at what they actually opened.
The weights, under MIT.
FlashMLA, attention kernels at 580 TFLOPS on H800.
DeepEP, the communication library for MoE.
DeepGEMM, FP8 matrix multiply at 1,350+ TFLOPS.
3FS, an entire distributed file system they built for themselves.
And today DeepInfer, the engine behind V4.1 Flash, plus deepseek-recipe so your deployment behaves like their API.
Thats every layer of an AI lab. Model, kernels, communication, storage, serving. Thomas called the streak absurd, and he is right.
Most labs open source a small model when the big one is safe behind an API. DeepSeek opens the parts that make their own business work. Then they ship the next model anyway.
I learn more from their repos than from most papers. Anyone can. For me thats a bigger gift than the cheap tokens.
Patrick D'appollonio retweeted
‼️ BREAKING: The private keys of 8 certificates from Indian certificate authority eMudhra were compromised, Mozilla says. Firefox blocked them within 3 hours.
Each key should be a 116-digit number, but one was just 4.
Anyone can guess them and forge HTTPS certs for any website.
Patrick D'appollonio retweeted
DeepSeek just published the price of a token, and it should make you angry.
From their own DeepInfer docs: 1 cluster serves 155 trillion tokens a day, at $38 of revenue per GPU-hour, and thats at off-peak prices. Their API is $0.30 per million input tokens and $1.20 per million output at peak.
Yuchen Jin did the math: even if they rented B300s at $7 an hour, thats about 82% gross margin. Other estimates go higher.
So cheap AI is not charity and it is not "dumping". At these prices serving a frontier-class model is a very profitable business, if your inference stack is good enough.
Now remember every time a US lab told you that intelligence is expensive. That $200 a month is a fair price. That limits exist because compute is scarce.
Part of that is true. Part of it is a business model built on you not knowing what a token actually costs.
DeepSeek just showed everyone the numbers and gave away the engine. Every inference company can now compete on price with the same software. Thats how AI gets cheap for everyone.
“A single cluster currently serves 155 trillion tokens/day, generating $38 in revenue per GPU-hour (at off-peak pricing).”
Huge flex from DeepSeek. Not only did they open-source their inference engine, but they also open-sourced their revenue numbers.
They didn't disclose which GPUs they're using here, but DeepInfer supports both NVIDIA Blackwell and Huawei Ascend.
Even assuming they're renting B300s at $7 per GPU-hour, that's ~82% gross margin.
