Embrace Reality, Painter of Possibilities, Cryptography, Internet Insecurity, Lover of Languages, spoken and coded.
Joined November 2009
- Tweets3.6K
- Following1.9K
- Followers550
- Likes1.6K
Matthew retweeted
Lost my phone at the office and spent 30 minutes turning the place over. Find My was disabled by MDM.
Out of ideas, I asked Claude how I could find it. It suggested tracking the Bluetooth signal strength, then wrote me a meter in about a minute.
I walked around watching the number climb. Found it.
Apparently you can just make the tool you need now.
Code: github.com/ben-z/findphone
Matthew retweeted
I co-founded pfSense. For the last year I've been building its successor.
If you run pfSense today, you already know the reasons to look around: development you can't influence, a CE edition that feels like an afterthought, FreeBSD driver roulette on modern hardware, and a config workflow where one bad apply on a remote box means a drive.
nfSensei is my answer. Built from scratch in Rust, on Linux, and designed around the things pfSense users actually complain about.
Your config.xml imports. There's an importer that reads your pfSense config, shows you exactly what maps over and what needs attention, then applies it. You don't start from zero.
You can't brick it from the couch. Changes stage as a candidate, diff before apply, validate through the real engines before anything is written, and auto-roll-back if you don't confirm in time. If a config ever fails at boot, the box falls back to the last good one on its own.
The hardware works. Linux base means modern NICs and drivers just work — and the fast path compiles your rules to XDP at 40Gbps.
Automation is native, not scraped. Everything the UI does is a documented API call — about 1,140 of them, with a built-in explorer. Your Ansible finally gets a real interface.
The VPNs are current. WireGuard, IPsec, Tailscale, and self-hosted mesh — your own control plane, your keys — plus post-quantum key exchange where it counts.
The experimental stuff has its own wing. Thirty-plus Labs features behind toggles: WAN bonding that fuses multiple cheap uplinks through a $5 VPS into one resilient pipe, per-flow SLA telemetry with tamper-evident audit chains, GeoDNS that steers traffic by live RTT and load, application-aware QoS, config push to a whole fleet of remote nodes, and an AI assistant on the box that reads your actual interfaces and logs using local models. Toggles are per-browser and can't touch your running config — flip things on, break them, tell me about it. Oh, and there is much more to mention here!
Self-hosted, on your hardware, no cloud account, no subscription.
It's NOW IN ACTIVE BETA (previously alpha) with about 40 testers, and bug reports typically get fixed in days. I want more people who know what pfSense does well and can tell me exactly where nfSensei falls short.
If you are interested in testing please email me: [email protected]. Tell me about your pfSense setup and I'll get you access.
Note: Affiliates and employees of Netgate are not invited.
Matthew retweeted
Understand Assembly low-level programming in 22 minutes youtu.be/DWkImpawzhc
I remember how it clicked for me. This video shows what I saw, illustrated with examples from all the mainstream CPU architectures
ElectricityEmissions.jl: Calculating carbon intensity signals on the power grid | Noah Rhodes
youtube.com/watch?v=fmvyd0Oo…
Matthew retweeted
Hermes Desktop just removed the biggest barrier to AI agents.
No more terminal.
Here’s what changed everything:
→ You can now run Hermes in a clean Mac app
→ Edit memory + skills without SSH
→ Search every past conversation instantly
→ Run multi-agent workflows visually
→ Assign tasks like a real team board
This turns AI from “tool” → “operator.”
Save this video, you’ll finally understand how agents actually work.
Want the SOP? DM me. 💬
Matthew retweeted
You can text someone hundreds of miles away with no cell service, no wifi, and no servers in between.
All it takes is a cheap radio board and open source software like Meshtastic. Meshtastic turns the board into a node in an encrypted mesh, capable of receiving and relaying information across vast distances.
Matthew retweeted
Your Hermes Agent can now build full videos with the official HyperFrames skill by @HeyGen
HyperFrames videos are HTML-native, so your agent has total control over the final output
Video made entirely by Hermes using the HyperFrames skill
Matthew retweeted
Introducing SubQ - a major breakthrough in LLM intelligence.
It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),
And the first frontier model with a 12 million token context window which is:
- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus
Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).
Only a small fraction actually matter.
@subquadratic finds and focuses only on the ones that do.
That's nearly 1,000x less compute and a new way for LLMs to scale.
Matthew retweeted
OpenTUI Keymap is a host agnostic key/cmd engine for DOM like apps.
It allows extreme customisation of a generic core, from key stroke syntax parsing to to command resolution and dispatch behaviour. It composes layers of bindings and commands into a single adaptive dispatch model.
Demo link in the replies.
There are many such systems, most of which stop at the key binding part. Apps often are expected to implement their own command layer on top of these. This disconnects the bindings from the actual command registry.
Looking for a proper solution usable in OpenCode I couldn't find anything that fulfilled the need of being extremely extensible to allow plugins to fully control key mapping and command behaviour. I wanted it to always be able to know exactly which keys and commands are reachable at any point in time, no matter from where and how the mappings are manipulated.
The main driver was enabling a which-key like plugin and vim like bindings in a mostly declarative way. While any other plugin could extend the keymap even further with a custom config syntax for key strokes for example. So addons for it are mostly composable and it comes with a variety of addons providing common behaviour.
It can also power other discovery features directly from live state. Help views, command palettes and graph/debug UIs can all ask the engine what is active, reachable, shadowed, pending, or dispatchable instead of rebuilding that knowledge separately.
The core is intentionally very much generic: hosts adapt focus, hierarchy, input events, and lifecycle, while addons extend parsing, tokens, sequence patterns, command metadata, resolvers, interceptors, and event matching. The result is a keymap that plugins can compose with rather than work around.
This could all be built directly into OpenCode and just be app specific. Working on OpenTUI though I want applications to have an out-of-the-box solution to build keyboard-first apps easily. Well, at least agents can.
It is extremely over-engineered. And I love it.
I hope some of you will too.
Matthew retweeted
nless is a TUI tool that is like Excel for live log data.
It takes in logs, CSV or JSON and transforms them into columns in the terminal. You can then filter, sort, search, stream live data and more.
Matt Pryor (mpryor on GitHub) made nless using @textualizeio and is Terminal Tool of the Week! ⭐️
Matthew retweeted
VLMs too slow for production? Not anymore: 46ms end-to-end inference, 60+ fps on a single H100. Introducing Photon, Moondream's inference engine. Runs on everything from edge to server. moondream.ai/blog/photon-rea…
Matthew retweeted
🚨 BREAKING: Warp just dropped its entire source code. 42,000 GitHub stars. Written in Rust. And it just made the traditional terminal obsolete.
This is not a terminal with AI features bolted on.
This is an agentic development environment where coding agents operate with full autonomy. From issue to PR. No babysitting.
Here's how it works:
→ You plug in Claude Code, Codex, or Gemini CLI
→ You give it a GitHub issue
→ The agent reads the codebase, writes the spec, implements the fix
→ It opens a PR with full context
→ You review and merge
That's the entire workflow. No prompting. No copy-pasting. No hand-holding.
Here's how far ahead they are:
At build. warp. dev you can watch thousands of AI agents building Warp's own codebase. Live. Right now. Triaging real issues. Writing real code. Submitting real PRs.
Click into any session. Watch the agent work in a real terminal. In your browser.
They are building the product with the product.
OpenAI is the founding sponsor. GPT powers the agentic management layer. The entire client is Rust-native and GPU-accelerated. Works on macOS, Linux, and Windows.
Your terminal types commands.
Warp ships features.
100% Open Source. AGPL-3.0 + MIT.
(Link in the comments)
Matthew retweeted
We're open-sourcing more of our stack: iron-proxy, a default-deny egress proxy for untrusted workloads.
It's been a bad month for supply chain security. Trivy, KICS, LiteLLM, and Axios all followed the same playbook: compromise a package, harvest env vars, exfiltrate secrets to a C2 server. Your CI runner, your agents, and your sandboxes all have real keys.
iron-proxy sits between your workload and the internet. Workloads get proxy tokens. iron-proxy swaps in real credentials at the network boundary. If a compromised dependency exfiltrates your API key, it gets a token that's worthless outside the proxy. It's how we secure our VMs at iron[.]sh.
The only way out is making sure there are no real secrets to steal.
GitHub: github.com/ironsh/iron-proxy
Matthew retweeted
Here is my take on why Elixir is the best language for AI: immutability, documentation, stability, and tooling for coding agents.
It builds on the recent study in which Elixir had the highest completion rate across models among 20 different languages.
Link in the thread below.
Joy is not always equal to cohesion:
"Perl6 introduces far more TIMTOWTDI than ever before." <-- perl.fandom.com/wiki/TIMTOWT…
