ori

https://hj.9fs.net/ori

We have a word for things that try to grow at any cost. That word is cancer.

Made of meat.

9front, Recurse Center, systems and data.

Site: https://orib.dev

Public Key

npub1485zn7htrffa6x3g6wkf2y329uc3x5a0q3p8en84eucgd5c4z00q0m8vcz

NIP-05 Address

ori_at_hj.9fs.net@momostr.pink

Profile Code

nprofile1qqs2n6pflt43557arg5d8ty4zg4z7vgn2whsgsnuen6u7vyx6v238hspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dssw4x8z

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1485zn7htrffa6x3g6wkf2y329uc3x5a0q3p8en84eucgd5c4z00q0m8vcz

Show more details

Last Notes

2026-05-10 15:59:11 UTC

- reply

But what music?
CC: @npub1z8r…mrq6

2026-05-10 15:54:33 UTC

- reply

Yes.
However, You didn't ask what they should do when they're clicked.
CC: @npub1z8r…mrq6

2026-04-20 18:27:27 UTC

- reply

Those are metric inches. Obviously.

2026-04-01 03:59:09 UTC

- reply

Please tell me that if I get the PDF version, I can copy and paste out the original text.
CC: @npub1qhk…u7aa

2026-03-31 13:12:44 UTC

- reply

Does it? The attacker needs to be able to control the contents of your .git directory, which is *not* copied by git clone. Git cloning a repository *will not* leave you vulnerable, this is why the PoC gets the reproducer with wget. Today, git repositories can execute arbitrary code, but they don't carry that code along if you clone them.
In other news, if someone is sitting in your living room, they can unlock your front door from the inside, so your locks must be defective!

2026-03-29 12:22:05 UTC

- reply

Remember that 'none of the above' is still a perfectly valid option, and is rather conservative: it's the status quo.
The privacy maximalist option is 'if you attempt to collect this information about your users, we will see you in court'. I would like to keep the status quo, which probably means that someone will have to start lobbying for stronger protections for users.

2026-03-28 13:35:49 UTC

- reply

That's with G as in GIF, right?
CC: @npub1j8e…ncgu

2026-03-28 01:08:11 UTC

- reply

Fwiw, https://sr.ht/~sircmpwn/vim-classic/ and https://codeberg.org/NerdNextDoor/evi
CC: @npub1l5h…tl0m

2026-03-22 13:49:38 UTC

- reply

Stuff.

2026-03-16 03:07:19 UTC

- reply

we fixed it in 9front with the great hammer of rm -rf. Thanks.

2026-03-02 03:44:41 UTC

A French person once told me: "You shouldn't found a bakery, it's just going to be a source of pain".

2026-02-15 07:29:23 UTC

- reply

Setting aside non-technical considerations: I wouldn't be able to put together a good spec if I hadn't translated it to code and worked out all the details for this. English is a sufficiently poor notation for thinking through a spec.
So, if LLMs write the code, my understanding of a problem would be poor enough that they would also be better at writing the specs, docs, and models.
Writing the code is homomorphic to studying the problem space in enough detail that I know I understood things.
CC: @npub1er4…x9ky @npub1ylv…mdmf

2026-02-14 07:06:30 UTC

- reply

Lobsters has also gotten sloppy. It may be necessary to start over.
CC: @npub1sl7…lnx3 @npub16y8…d455

2026-02-08 09:41:56 UTC

- reply

I think it's the documented way to pass a timeout:
> Several _umtx_op() operations allow the blocking time to be limited, failing the request if it cannot be satisfied in the specified time period. The timeout is specified by passing either the address of struct timespec, or its extended variant, struct _umtx_time, as the uaddr2 argument of _umtx_op(). They are distinguished by the uaddr value, which must be equal to the size of the structure pointed to by uaddr2, casted to uintptr_t.
Kinda gross, and iirc inherited from Linux futex apis.
CC: @npub1fq6…wmwe

2026-02-07 22:54:38 UTC

- reply

I would be very surprised if I posted a followers
-only note that a no -follower could see.

2026-02-07 22:16:53 UTC

- reply

Well, you asked what expectations were, and I told you. They're not what you seem to be proposing. Not sure what else to say.

2026-02-07 22:00:09 UTC

- reply

Yes. That's equivalent to Alice setting their post followers only, and Bob setting their post public, but less confusing to the user.

2026-02-07 21:55:25 UTC

- reply

Yes. Either you do that or you ignore someone's privacy settings.

2026-02-07 21:33:33 UTC

- reply

The intersection of Alice and Bob's followers.

2026-02-05 11:19:31 UTC

Did you know that there's a small town in Italy called Cologne? I sure didn't when I booked my train tickets from Cologne to Rome.
Sigh.
Alternative travel plans arranged.

2026-02-04 17:11:38 UTC

- reply

Most breakers are electromagnetic, not resistive.
CC: @npub19av…rp6r

2026-01-28 23:55:19 UTC

- reply

You can see what happens if you tell it to not search the Internet, for vanity.
For me, it doesn't always work. It will also sometimes say things like "I have no idea who Ori Bernstein is, but here's code written in the style of a low level systems programmer with a focus on performance".

2026-01-28 23:30:44 UTC

- reply

> "Admitting you don't know the answer looks weak"
One of my favorite bits of feedback I got on a talk was "I love how confidently you said you didn't know things".
There's a lot I don't know. I consider it basic courtesy to clearly flag the boundaries of my knowledge. It's rude to pass around falsehoods.
"Strength", whatever the hell that means in this context, comes from being able to think about how to get the missing information, or being able to think about what the parts of a system you do understand constrain the parts you don't know about. There's real skill demonstrated if someone is able to say "I don't know, but if you flip the third rock from the left, the answer will be right there."

2026-01-28 23:11:48 UTC

- reply

The best is that I can (unreliably) ask an LLM to write in the style of 'Ori Bernstein' and it will. Once it even decided to do it in a language I wrote.
The irony: I hate using LLMs to write code, so this is only really useful as a party trick for impressing ai-brained people.

2026-01-25 14:56:08 UTC

- reply

Would the insurance cover it?
CC: @npub1fc6…2khj

2026-01-13 13:41:43 UTC

- reply

Maybe https://aosabook.org/en/ ?

2026-01-11 15:14:10 UTC

- reply

When I was in high school, the introductory programming class didn't let us use a computer for about a month. Instead, we would get programs on paper, and would be asked to trace out what they did.
When I taught friends how to code, I was less strict, but writing out the steps on paper was still a big part of it. I also wouldn't let people make changes to their code without first telling me what they thought was broken, and predicting/explaining what the change would do.
I think it's a very effective approach.

2026-01-09 05:09:49 UTC

- reply

The new Plan 9 file system I've been working on for a while (GEFS) is getting very close to stable, and my port of it to OpenBSD has read its first directory listing.

2025-12-22 18:05:32 UTC

- reply

Look at affinewarp. It's a new feature, added last summer.
CC: @npub1trl…vm7z

2025-12-20 15:11:09 UTC

- reply

bitsy/keyboard
CC: @npub1jt2…s5v4

2025-11-26 05:02:06 UTC

- reply

From early Unix: sbrk and the stack.
It's a good idea.
CC: @npub1ll6…jjcy

2025-10-21 21:44:14 UTC

- reply

Counter culture is a culture where you add as many counters as possible, so you can track user activity. Right?

2025-09-18 03:05:43 UTC

- reply

Gridchat. It's still pretty active -- this is the script I use to connect: https://orib.dev/gchat
CC: [@sigrid@ftrv.se](https://nein.ftrv.se/users/sigrid )

2025-09-16 12:09:33 UTC

- reply

Why specifically one that's be hard to turn down? I would have thought volume control would be important.

2025-09-12 03:32:50 UTC

- reply

Heck, we've got most of it. What if we had a virtual PCI bus to enumerate and identify the built-in stuff, and picked a short list of supported stuff to put on htere?
This obsoletes devicetree, parts of acpi, and the other things everyone hates.

2025-09-12 03:20:08 UTC

- reply

The PC isn't a great platform, but it is a platform.
There's a lot of hardware bugs and stupidity to deal with, but at least it's pretending that it's interoperable. I haven't had to worry about which interrupt controller the motherboard I bought had for a while.

2025-09-11 18:29:33 UTC

- reply

"hey, let's quickly hack in a global flag". And this is why I still use [rc](https://shithub.us/cinap_lenrek/rc/HEAD/info.html ), even on unix.
CC: @npub187a…5xdw

2025-09-10 18:43:27 UTC

- reply

It's never about the instruction set, but about the peripherals.

2025-08-19 04:16:48 UTC

Well. New job, let's see how it goes.
Back to poking at databasey things at ParadeDB.

2025-08-13 06:48:52 UTC

- reply

I don't have an openbsd at hand, but it's reportedly supported by mandoc: https://undeadly.org/cgi?action=article&sid=20181002175838

2025-08-11 09:34:11 UTC

The problem with a good hackathon: By the end, I have more projects than I started with.

2025-06-06 15:23:13 UTC

- reply

Not sure it matters. Between ordering and delivering it's going to change anyways. 1%? 10000%? Who the fuck knows.
You could probably probably make a few bucks with a challenge coin slot machine: Guess the difference in tariffs between order placed and order delivered, and you win!

2025-06-04 23:31:08 UTC

- reply

Scaling for hi dpi robs me of the pixels I wanted to use for more text.
CC: @npub1cp3…39ww

2025-04-28 14:41:42 UTC

- reply

If a baker is someone who bakes, a badger must be someone who attaches badges to attendees.
CC: [@bsdcan@bsd.network](https://bsd.network/users/bsdcan )

2025-04-23 01:59:31 UTC

- reply

https://antithesis.com/ ?

2025-04-22 16:03:53 UTC

- reply

Where's the alliteration?
Perverse Presents.

2025-04-06 19:20:00 UTC

- reply

I like technology. I don't like the technology industry.

2025-02-14 17:58:41 UTC

Walking by a tour group: "New York City is the most expensive rat habitat in the world"
They're not wrong.

2025-02-09 04:40:22 UTC

- reply

> until they experience consequences.
Well, aren't we optimistic.
CC: @npub1rz2…xaz3

2025-01-30 17:55:52 UTC

- reply

Skimming the replies, I'm surprised to see so few mentions of books.
There's the classic TUPE book, which is a great introduction: https://github.com/tcd/tupe/blob/master/the-unix-programming-environment-kernighan-pike.pdf
There's lots of others, including more modern ones. That's a big part of how I learned, and then I fit new knowledge into the framework set up by the books.