Here we go
Posts by David Justice
Can't wait for @dblnz.dev to dig into debugging! #webassembly at @wasm.io
This morning we will be using this sample to teach #hyperlight and #webassembly at @wasm.io
codespaces.new/hyperlight-d... (a bigger machine is a good idea when prompted)
And a double bonus! Debugging! (though not kotlin quite yet, alas) : 2026.wasm.io/sessions/eve...
Want a sneak peek at some of fun #Hyperlight stuff going on at Wasm I/O?
of course! The basic sample we'll be using will be github.com/hyperlight-d.... Running and navigating this in order to understand how the component calls get mapped into and back out of the guest micro-vm -- that has no operating system -- is the main core of the workshop. Bonus!!! Changing the WIT!
If you're coming to @wasm.io this week and want to build and run #webassembly components in #hyperlight-wasm this Thursday, you'll need a linux box (or windows with wsl) that has nested virt enabled.
M3 and M5 macs **may** work, but that's coming in really hot. So get those cloud VMs ready to use!
This is going to be fun. Se tu sei vicino a Bologna, devi venire!
I'm going to show a bunch of fun things!!!
I can definitely use this to automate some of the chores and toil that I don't want to deal with on a daily / weekly basis for the projects I and my teams maintain. Nice work, @github.com!
docfind - a high-performance document search engine built in Rust with WebAssembly support
Combines full-text search using FST (Finite State Transducers) with FSST compression. Uses RAKE to identify multi word phrases. Uses a postcard, a custom serialization.
github.com/microsoft/do...
I might be even more thrilled for the entire @fermyon.com and #Akamai teams here than they are. Akamai has realized that it could put together a next-gen cloud native platform that both enables and pushes a huge amount of work far, far closer to the users than ever. And Fermyon....
please enjoy: my Wasm-hosted, Wasm-targeting build of Clang/Clang++/LLD: a self-contained, 25 MiB (gzipped) pure function
www.npmjs.com/package/@yow...
Graphic depicts select capital flows between OpenAI, Microsoft, Nvidia, CoreWeave, AMD, and Oracle.
Another depiction if the circularity of capital flows between OpenAI, Microsoft, Nvidia, CoreWeave, AMD, and Oracle.
For those Master's or Ph.D. students doing systems work, you might have a look at this internship with @microsoft.com's research team: www.linkedin.com/posts/pedroh...
Yes, you'll do some immediately important work, and it's really cool, too.
Impressive work from Don Syme on Continuous Test Improvement. The post highlights how AI-driven workflows can systematically expand and improve test coverage with minimal cost and time. dsyme.net/2025/08/27/o...
I believe a good test for documentation would be if you can point a coding agent at the docs and have it complete on the first attempt. If it does not, then you should update the docs to provide more instruction to your user and the agent. Rinse, repeat.
I think agentic coding AI's offer a really interesting test for usability. Say you are building dev tools and you want to see if given documentation a new user could do it, give it to the AI and see where it fails. I'm going to bet the places where the AI fails will be sticking points for humans.
Must read post from @rawkode.dev that does a wonderful job explaining the differences between Wassette and common ways of running MCP workloads.
🎉 Introducing Wassette: a runtime for secure, sandboxed WebAssembly Component tool execution via the Model Context Protocol (MCP). Easily register reusable WASM tools inside AI agents like VS Code all without leaving the chat window. Start using it to safely extend your agent’s capabilities! 🚀
Our USENIX ;login; article on Crescendo just dropped! We discovered what might be the most powerful blackbox jailbreak technique - Crescendo is undefeated against all frontier AI models in every harm domain we tested. I'm presenting at USENIX Security on August 14th.
www.usenix.org/publications...
i've modified wasi-libc and wasi-sdk (PRs pending) so that you always have the concurrency APIs (pthread_*, thrd_*, <thread>, <mutex>, <atomic>) available to you, even in single-threaded targets. this makes porting applications (especially C++ ones) much, _much_ easier!
For those who are about difficult hardware #virtualization efficiency and dev experience (WTF????) you might find @dblnz.dev's blog post on #hyperlight and interactive debugging interesting. It's a blast for any micro-guest, including #webassembly. opensource.microsoft.com/blog/2025/07...
Missed this: very good article by @thenewstack.io on a @rejekts.io talk prior to #kubeconEU about #hyperlight and #webassembly in #envoy. Slow as a first draft; but getting faster and makes hostile multitenant #networkfunctions possible to imagine. thenewstack.io/webassembly-...
I confess to liking the progress made in the #Dual-LLM family of #llm security patterns started with @simonwillison.net's original Dual-LLM "separation of query concerns" in simonwillison.net/2023/Apr/25/....
....
I put LLM spike coding into all my recruitment ads, in white font. Some of the replies are a lot of fun, and it sure speeds up shortlisting.
Come join Taylor Thomas and I today as we discuss Wasm I Right or Wasm I Wrong? a Review of the Wasm Ecosystem at KubeCon + CloudNativeCon Europe 2025 sched.co/1tcxS #KubeCon
Come join Terence Lee and I today as we discuss Building WebAssembly Like It's 2011 - at KubeCon + CloudNativeCon Europe 2025 sched.co/1txGl
There's an incredibly talented and well liked PM I have the fortunate opportunity to observe and emulate.
Running #Wasm components inside of a hypervisor and being able to tailor the exact interface between the hosted application and the trusted compute base enables platform builders unequaled ability to tune precisely the access provided to untrusted application code.
Good day, folks! The #CNCF TAG-Runtime Wasm Working Group is meeting today in about 30 mins. We'll be covering updates on WASI cloud interfaces (blob, key/value, config, etc) and an update on WASI preview 3! Come join or just listen in! docs.google.com/document/d/1...