Out of interest I tried running my Primes benchmark [1] on both the x86_64 and x86 Alpine and the riscv64 Buildroot, both in Chrome on M1 Mac Mini. Both are 2nd run so that all needed code is already cached locally.
x86_64:
localhost:~# time gcc -O primes.c -o primes
real 0m 3.18s
user 0m 1.30s
sys 0m 1.47s
localhost:~# time ./primes
Starting run
3713160 primes found in 456995 ms
245 bytes of code in countPrimes()
real 7m 37.97s
user 7m 36.98s
sys 0m 0.00s
localhost:~# uname -a
Linux localhost 6.19.3 #17 PREEMPT_DYNAMIC Mon Mar 9 17:12:35 CET 2026 x86_64 Linux
x86 (i.e. 32 bit):
localhost:~# time gcc -O primes.c -o primes
real 0m 2.08s
user 0m 1.43s
sys 0m 0.64s
localhost:~# time ./primes
Starting run
3713160 primes found in 348424 ms
301 bytes of code in countPrimes()
real 5m 48.46s
user 5m 37.55s
sys 0m 10.86s
localhost:~# uname -a
Linux localhost 4.12.0-rc6-g48ec1f0-dirty #21 Fri Aug 4 21:02:28 CEST 2017 i586 Linux
riscv64:
[root@localhost ~]# time gcc -O primes.c -o primes
real 0m 2.08s
user 0m 1.13s
sys 0m 0.93s
[root@localhost ~]# time ./primes
Starting run
3713160 primes found in 180893 ms
216 bytes of code in countPrimes()
real 3m 0.90s
user 3m 0.89s
sys 0m 0.00s
[root@localhost ~]# uname -a
Linux localhost 4.15.0-00049-ga3b1e7a-dirty #11 Thu Nov 8 20:30:26 CET 2018 riscv64 GNU/Linux
Conclusion: as seen also in QEMU (also started by Bellard!), RISC-V is a *lot* easier to emulate than x86. If you're building code specifically to run in emulation, use RISC-V: builds faster, smaller code, runs faster.
Note: quite different gcc versions, with x86_64 being 15.2.0, x86 9.3.0, and riscv64 7.3.0.
Interesting to see the gcc version gap between the targets. The x86_64 image shipping gcc 15.2.0 vs 7.3.0 on riscv64 makes the performance comparison less apples-to-apples than it looks - newer gcc versions have significantly better optimization passes, especially for register allocation.
MIPS (the arch of which RISCV is mostly a copy) is even easier to emulate, unlike RV it does not scatter immediate bits al over the instruction word, making it easier for an emulator to get immediates. If you need emulated perf, MIPS is the easiest of all
The thing I most want to use this (or some other WASM Linux engine) for is running a coding agent against a virtual operating system directly in my browser.
Claude Code / Codex CLI / etc are all great because they know how to drive Bash and other Linux tools.
The browser is probably the best sandbox we have. Being able to run an agent loop against a WebAssembly Linux would be a very cool trick.
I had a play with v86 a few months ago but didn't quite get to the point where I hooked up the agent to it - here's my WIP: https://tools.simonwillison.net/v86 - it has a text input you can use to send commands to the Linux machine, which is pretty much what you'd need to wire in an agent too.
In that demo try running "cat test.lua" and then "lua test.lua".
> The thing I most want to use this (or some other WASM Linux engine) for is running a coding agent against a virtual operating system directly in my browser.
Has there ever been any other topic that was not only the subject of the majority of submissions, but also had a subset of users repeatedly butting into completely unrelated discussions to go "b-but what about <thing>? we need to talk about <thing> here too! how can I relate this to <thing>? look at my <thing> product!"?
You can't just roll in to a random post to tell people about your revolutionary new AI agent for the 50th time this week and expect them not to be at least mildly annoyed.
I'm with you, but he wasn't telling us about his agent, he was saying "this is a cool technology and I've been wanting to use it to make a thing". The thing just happened to be LLM-adjacent.
Almost all of his comments "just happen" to be LLM-adjacent. At some point it stops "just happening" and it becomes clear that certain people (or their AI bots) are frequenting discussion spaces for the sole purpose of seeking out opportunities to bring up AI and self-promote.
Simon has been here since way before LLMs were a thing, and it's fairly obvious (to me, at least) that he's genuinely excited about LLMs, he's not just spamming sales or anything.
The entire thing is just quotes and a retelling of events. The closest thing to a "take" I could find is this:
> I have no idea how this one is going to play out. I’m personally leaning towards the idea that the rewrite is legitimate, but the arguments on both sides of this are entirely credible.
Which effectively says nothing. It doesn't add anything the discussion around the topic, informed or not, and the post doesn't seem to serve any purpose beyond existing as an excuse to be linked to and siphon attention away from the original discussion (I wonder if the sponsor banner at the top of the blog could have something to do with that...?)
Literally just a quote from his fellow member of the "never stops talking about AI" club, Karpathy. No substance, no elaboration, just something someone else said or did pasted on his blog followed by a short agreement. Again, doesn't add anything or serve any real purpose, but was for some reason submitted to HN[1], and I may be misremembering but I believe it had more upvotes/comments than the original[2] at one point.
I run agents as a separate Linux user. So they can blow up their own home directory, but not mine. I think that's what most people are actually trying to solve with sandboxing.
(I assume this works on Macs too, both being Unixes, roughly speaking :)
Apptron uses v86 because its fast. Would love it for somebody to add 64-bit support to v86. However, Apptron is not tied to v86. We could add Bochs like c2w or even JSLinux for 64-bit, I just don't think it will be fast enough to be useful for most.
Apptron is built on Wanix, which is sort of like a Plan9-inspired ... micro hypervisor? Looking forward to a future where it ties different environments/OS's together.
https://www.youtube.com/watch?v=kGBeT8lwbo0
While this may be a better sandbox, actually having a separate computer dedicated to the task seems like a better solution still and you will get better performance.
Besides, prompt injection or simpler exploits should be addressed first than making a virtual computer in a browser and if you are simulating a whole computer you have a huge performance hit as another trade off.
On the other hand using the browser sandbox that also offers a UI / UX that the foundation models have in their apps would ease their own development time and be an easy win for them.
tldr; devcontainers let you completely containerize your development environment. You can run them on Linux natively, or you can run them on rented computers (there are some providers, such as GitHub Codespaces) or you can also run them in a VM (which is what you will be stuck with on a Mac anyways - but reportedly performance is still great).
All CLI dev tools (including things like Neovim) work out of the box, but also many/most GUI IDEs support working with devcontainers (in this case, the GUI is usually not containerized, or at least does not live in the same container. Although on Linux you can do that also with Flatpak. And for instance GitHub Codespaces runs a VsCode fully in the browser for you which is another way to sandbox it on both ends).
This is interesting (and I've seen it mentioned in some editors), but how do I use it? It would be great if it had bubblewrap support, so I don't have to use Docker.
Do you know if there's a cli or something that would make this easier? The GitHub org seems to be more focused on the spec.
Maybe not, but in the past some here see that the blog is the product that is being promoted here.
Even in this thread alone https://news.ycombinator.com/item?id=47314929 some commenters here are clearly annoyed with the way AI is being shoved in each place where they do not want it.
I don't care, but I can see why many here are getting tired of it.
Sorry for the off-topic, but what a bliss to see Windows 2000 interface. And what an absolute abomination from hell pretty much all the modern UIs are.
It looks like container2wasm uses a forked version of Bochs to get the x86-64 kernel emulation to work. If one pulled that out separately and patched it a bit more to have the remaining feature support it'd probably be the closest overall. Of course one could say the same about patching anything with enough enthusiasm :).
> Access to Internet is possible inside the emulator. It uses the websocket VPN offered by Benjamin Burns (see his blog). The bandwidth is capped to 40 kB/s and at most two connections are allowed per public IP address. Please don't abuse the service.
I use bellard.org/jslinux to test compilation of strange code sometimes[1], since it came with compilers that are different versions from what I have installed locally, and it's easier to open up a browser than starting a VM.
I use a similar emulator (v86) as a way to share my hobby OS. Approximately zero people, even my friends, are going to boot my hobby OS on real hardware; I did manage to convince some of them to run it in qemu, but it's difficult. A browser environment shows the thing quite well; and easy networking is cool too.
My hobby OS itself is not very useful, but it's fun if you're in the right mood.
We are a playful species. People enjoy play. If we didn't have to work for a living but still enjoyed food security that is all most of us would do. But we are also a very exploitative species, some more than others. Companies have made billions of dollars on top of Fabrice Bellard's works, qemu, ffmpeg etc.
These companies don't have any imagination. Their management has no vision. They could not create anything new and wonderful if they tried. People like Fabrice do and we are all richer for it. If your asking about the practical use you are likely in the exploitative mindset which is understandable on HN. The hacker/geek mindset enjoys this for what it is.
The tech docs at bellard.org/jslinux/tech.html describe the TEMU config format. Each VM is just a .cfg file - you can see the existing ones linked on the main page. The VFsync integration (vfsync.org) is probably your best bet for getting files into the VM without rebuilding the disk image.
For a classroom with Windows PCs this is close to ideal - zero install, no admin rights, works in any browser. Students get a real gcc toolchain and shell without touching the host OS.
Working with VMs always felt difficult because of this. So authoring was built-in to Docker. Now you can use Apptron to author and embed a Linux system on a web page. This aspect is usable, but it's only going to get better.
Most such emulators have Internet access on the IP level. Therefore, this is a very cheap way to test anything on the Internet.
apk add nmap
nmap your.domain.com
However, the speed is heavily throttled. You can even use ssh and login to your own server.
It can also be used as a very cheap way to provide a complete build environment on a single website, for example to teach C/C++. Or to learn the shell. You don't have to install anything.
Agentic workloads create and then run code. You don't want to just run that code in a "normal" environment like a container, or even a very well protected VM. There are other options, ofc - eg. gvisor, crossvm, firecracker, etc, but this one is uncommon enough to have a small number of attackers trying to hack it.
What's wrong with a well protected VM? Especially compared to something where the security selling point is "no one uses it" (according to your argument; I don't know how secure this actually is)
Yeah but GP was answering to a comment saying "you don't want to run code in a well protected VM". Which is of course complete non sense to say and GP was right to question it.
Maybe if you’ve got some ancient software that’s missing source code and only runs with X Y and Z conditions, you could continue to offer it on the web and build around it like that? Not sure if that would be practical at all, but could be interesting
x86_64:
x86 (i.e. 32 bit): riscv64: Conclusion: as seen also in QEMU (also started by Bellard!), RISC-V is a *lot* easier to emulate than x86. If you're building code specifically to run in emulation, use RISC-V: builds faster, smaller code, runs faster.Note: quite different gcc versions, with x86_64 being 15.2.0, x86 9.3.0, and riscv64 7.3.0.
[1] http://hoult..rg/primes.txt
Also MIPS code is much larger.
Claude Code / Codex CLI / etc are all great because they know how to drive Bash and other Linux tools.
The browser is probably the best sandbox we have. Being able to run an agent loop against a WebAssembly Linux would be a very cool trick.
I had a play with v86 a few months ago but didn't quite get to the point where I hooked up the agent to it - here's my WIP: https://tools.simonwillison.net/v86 - it has a text input you can use to send commands to the Linux machine, which is pretty much what you'd need to wire in an agent too.
In that demo try running "cat test.lua" and then "lua test.lua".
That exists: https://github.com/container2wasm/container2wasm
Unfortunately I found the performance to be enough of an issue that I did not look much further into it.
This thing is really inescapable those days.
I should have replied there instead, my mistake.
I'm excited about them and I think discussion on how to combine two exciting technologies are exactly what I'd like to see here.
You can't just roll in to a random post to tell people about your revolutionary new AI agent for the 50th time this week and expect them not to be at least mildly annoyed.
The entire thing is just quotes and a retelling of events. The closest thing to a "take" I could find is this:
> I have no idea how this one is going to play out. I’m personally leaning towards the idea that the rewrite is legitimate, but the arguments on both sides of this are entirely credible.
Which effectively says nothing. It doesn't add anything the discussion around the topic, informed or not, and the post doesn't seem to serve any purpose beyond existing as an excuse to be linked to and siphon attention away from the original discussion (I wonder if the sponsor banner at the top of the blog could have something to do with that...?)
This seems to be a pattern, at least in recent times. Here's another egregious example: https://simonwillison.net/2026/Feb/21/claws/
Literally just a quote from his fellow member of the "never stops talking about AI" club, Karpathy. No substance, no elaboration, just something someone else said or did pasted on his blog followed by a short agreement. Again, doesn't add anything or serve any real purpose, but was for some reason submitted to HN[1], and I may be misremembering but I believe it had more upvotes/comments than the original[2] at one point.
[1] https://news.ycombinator.com/item?id=47099160
[2] https://news.ycombinator.com/item?id=47096253
(I assume this works on Macs too, both being Unixes, roughly speaking :)
Apptron uses v86 because its fast. Would love it for somebody to add 64-bit support to v86. However, Apptron is not tied to v86. We could add Bochs like c2w or even JSLinux for 64-bit, I just don't think it will be fast enough to be useful for most.
Apptron is built on Wanix, which is sort of like a Plan9-inspired ... micro hypervisor? Looking forward to a future where it ties different environments/OS's together. https://www.youtube.com/watch?v=kGBeT8lwbo0
For a full-stack demo see: https://vitedemo.browserpod.io/
To get an idea of our previous work: https://webvm.io
Besides, prompt injection or simpler exploits should be addressed first than making a virtual computer in a browser and if you are simulating a whole computer you have a huge performance hit as another trade off.
On the other hand using the browser sandbox that also offers a UI / UX that the foundation models have in their apps would ease their own development time and be an easy win for them.
tldr; devcontainers let you completely containerize your development environment. You can run them on Linux natively, or you can run them on rented computers (there are some providers, such as GitHub Codespaces) or you can also run them in a VM (which is what you will be stuck with on a Mac anyways - but reportedly performance is still great).
All CLI dev tools (including things like Neovim) work out of the box, but also many/most GUI IDEs support working with devcontainers (in this case, the GUI is usually not containerized, or at least does not live in the same container. Although on Linux you can do that also with Flatpak. And for instance GitHub Codespaces runs a VsCode fully in the browser for you which is another way to sandbox it on both ends).
Do you know if there's a cli or something that would make this easier? The GitHub org seems to be more focused on the spec.
Please? I'm begging here.
Even in this thread alone https://news.ycombinator.com/item?id=47314929 some commenters here are clearly annoyed with the way AI is being shoved in each place where they do not want it.
I don't care, but I can see why many here are getting tired of it.
For a more open-source version, check out container2wasm (which supports x86_64, riscv64, and AArch64 architectures): https://github.com/container2wasm/container2wasm
It looks like container2wasm uses a forked version of Bochs to get the x86-64 kernel emulation to work. If one pulled that out separately and patched it a bit more to have the remaining feature support it'd probably be the closest overall. Of course one could say the same about patching anything with enough enthusiasm :).
Even though it has no JIT. Truly magic :)
> Access to Internet is possible inside the emulator. It uses the websocket VPN offered by Benjamin Burns (see his blog). The bandwidth is capped to 40 kB/s and at most two connections are allowed per public IP address. Please don't abuse the service.
https://bellard.org/jslinux/tech.html
[1] For example:
https://www.ioccc.org/2020/yang/index.html#:~:text=tcc%200.9...
https://www.ioccc.org/2018/yang/index.html#:~:text=tcc%200.9...
My hobby OS itself is not very useful, but it's fun if you're in the right mood.
These companies don't have any imagination. Their management has no vision. They could not create anything new and wonderful if they tried. People like Fabrice do and we are all richer for it. If your asking about the practical use you are likely in the exploitative mindset which is understandable on HN. The hacker/geek mindset enjoys this for what it is.
Any advice on how to create a JSLinux clone with a specific file pre-installed and auto-launching would be much appreciated!
For a classroom with Windows PCs this is close to ideal - zero install, no admin rights, works in any browser. Students get a real gcc toolchain and shell without touching the host OS.
It can also be used as a very cheap way to provide a complete build environment on a single website, for example to teach C/C++. Or to learn the shell. You don't have to install anything.
We have Windows PCs in the classroom.
From "Show HN: Amla Sandbox – WASM bash shell sandbox for AI agents" (2026) https://news.ycombinator.com/item?id=46825119 :
>>> How to run vscode-container-wasm-gcc-example with c2w, with joelseverin/linux-wasm?
>> linux-wasm is apparently faster than c2w
From "Ghostty compiled to WASM with xterm.js API compatibility" https://news.ycombinator.com/item?id=46118267 :
> From joelseverin/linux-wasm: https://github.com/joelseverin/linux-wasm :
>> Hint: Wasm lacks an MMU, meaning that Linux needs to be built in a NOMMU configuration
From https://news.ycombinator.com/item?id=46229385 :
>> There's a pypi:SystemdUnitParser.