Karl Fogel

Home page: http://red-bean.com/kfogel/
Fediverse:
- <@[email protected]> (I normally tweetidentitoot from here)
- <@[email protected]> (my blog, which is also Fediverse-enabled)

Public Key

npub1y8jvd0xyzjjtwty3wahdkrkfaqndjahy87d2eamg6dj3d2qj7s4sm2kc54

NIP-05 Address

[email protected]

Profile Code

nprofile1qqszrexxhnzpff9h9jghwmkmpmy7sfkewmjrlx4v7a5dxegk4qf0g2cpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsy07ejd

Publishing to

offchain.pub

relay.damus.io

nos.lol

Author Public Key

npub1y8jvd0xyzjjtwty3wahdkrkfaqndjahy87d2eamg6dj3d2qj7s4sm2kc54

Show more details

Last Notes

2026-08-11 20:20:50 GMT

Really interesting Principal Tech Engineer job opening from a good organization (OpenET) that @nprofile…p7yv and I have worked with:
https://www.linkedin.com/posts/openetdata_openet-principal-technology-engineer-activity-7490861737311723521-LBb2
This position needs someone who has human skills and project management experience, as well as some tech chops -- but you don't have to be an extraordinary programmer: you just need to be solid, have cloud computing experience, and be able to be a contributing member on an already existing and very capable cross-organizational team.
Feel free to ask me questions, or just apply if you think you are the right person!

2026-05-18 05:14:04 GMT

- reply

@nprofile…50mt Another take: An industrial regulation created by publishers to curtail competition, descended from and modeled on the state censorship law that it replaced, shouldn't be our framework for figuring out the ethics of LLMs nor, that matter, the ethics of anything else.
There are plenty of critical things one can say about LLMs, but the fact that training them involves copyright violation just puts them in the same company as, say, jazz musicians. It's not a point against them; it might even be a point in their favor.

2026-03-10 15:54:29 GMT

Don't look a gift horse in the mouth while changing it in mid-stream after the barn door has closed.
Just trust me on this one.

2026-03-03 20:44:07 GMT

I'm trying to better understand the non-FOSS source-available licenses, and who uses them and why. If you have opinions and feel like sharing them, I'm all ears!
I'm aware of:
Commons Clause
Business Source License (BUSL)
Functional Source License
Server Side Public License (SSPL) (arguably in some ways a kind of super-strict free software license, but that's just my opinion and I know most of you don't agree with me it's okay we can be friends)
Fair Core License (see fcl.dev)
Microsoft's three:
Limited Public License
Limited Reciprocal License
Reference Source License
Elastic License (did anyone other than Elastic use it? Does even Elastic still use it?)
There are a few others listed here but I think I've hit the major ones above.
Did I miss anything big? And does anyone know of any notable uses of any of these (other than the uses listed on the Wikipedia page above)? Especially, if there's some flourishing software project that gets incoming contributions using one of them, that would be interesting to know about.

2026-01-26 16:20:02 GMT

So: governance diagrams for open source projects? Does anyone know of any particularly good ones?
I'm not a particularly visual person myself, for better or worse, so I tend not to turn to such diagrams first. But others have said that they're very helpful, so I'm trying to have some good examples to draw on!

2026-01-19 20:26:50 GMT

- reply

@nprofile…4tva That's what I can't stop thinking about.

2026-01-19 16:02:43 GMT

Real life is analog.

2026-01-08 05:17:54 GMT

- reply

@nprofile…p7yv "Rosamund" or "Rosamond"?

2025-12-18 20:14:09 GMT

It's another one of those "Ok, what bodily subsystem do I have to sell to be granted the ability to write that well?" moments, this time upon reading John Lanchester's multi-book review about finance in LRB:
https://www.lrb.co.uk/the-paper/v46/n17/john-lanchester/for-every-winner-a-loser
Astonishingly good. ^/ to @nprofile…998y for the link.

2025-11-10 21:27:34 GMT

- reply

@nprofile…j0re "In Case It Matters", but this may not be as widely-used as I guessed! Hard to keep track of which acronyms are prevalent in which communities, sometimes... :-)

2025-11-10 18:49:05 GMT

Dear Fediverse: Does anyone have any great examples of a open source projects undergoing successful governance transitions? Or for that matter unsuccessful transitions?
I'm interested in all types of projects that have done this, but in particular in those that started from a single founding organization and then later made a transition to effective governance by a broader set of stakeholders. The founding organization may well have encouraged this transition -- I'm not trying to focus on hostile or scandalous forks here, though if your example happens to fall in that category I wouldn't turn it down!
(ICIM: Yes, I have read "Governance Matters: Lessons from Restructuring the data.table OSS Project" (Oliviera, Amoakohene, Hocking, Gerosa, and Steinmacher) and seen tickets #5676 and #5656.)

2025-10-24 19:56:26 GMT

This is wild: https://museum.unesco.org/
Did you know about this? I did not; now I do, and I can't get it out of my head.
Make sure you go in far enough to get to the "physical" manifestation of the museum itself, and then explore from there.
There are a things one could say about some specifics of the user experience sure, but what a great concept, and not shoddy on the production values either.

2025-08-22 15:47:15 GMT

Just for next week, everyone switch to pronouncing "logic" with a hard "g" and "login" with a soft "g".

2025-06-12 17:14:01 GMT

In an email from the company that runs our co-working space:
"We've noticed you haven’t been opening our emails recently. ..."
ZOMG where to start? No, what you've noticed is that I'm using a mailreader that doesn't betray me. Sorry you're accustomed to users who run pro-betrayal software. That's a pity, and has apparently led to unrealistic expectations that we're probably not ever going to discuss. Oh well.

2025-06-11 22:18:37 GMT

Me, four and a half years ago, toward the end of Trump's first term:
It's time for my Increasingly Frequent Reminder:
He will put the military in a bind and force them to choose. He will go there. Why not try? No cost to him if he fails. The only question is what "national emergency" to concoct in order to maximize his chances of success.
(on birdsite: https://x.com/kfogel/status/1330011144795336706)

2025-04-24 03:02:18 GMT

Even good organizations make mistakes sometimes. The OSI should do exactly what this petition says and release the original full results of its 2025 election. The petition explains why very clearly; I've signed it, and I urge you to do the same.
I support the OSI's mission wholeheartedly. I served a three-year term on its Board of Directors, about a decade ago -- that was the Board that transformed the OSI into a membership organization that holds member-driven elections, so I take deviations from that very seriously. (More recently -- full disclosure -- my company has done research consulting for OSI.)
The reason the OSI should publish the unaltered election results is that is the only way such elections should ever be run is with full transparency. That's what voters and candidates both expect and deserve. There can be no qualifiers, conditionals, or hedging when it comes to how an election is done: you either do it right or, in the rare cases where there are truly exceptional circumstances, you disclose them and explain precisely why they required a change in procedure. In this case, there were no circumstances that really required a change in procedure, and the OSI then changed the procedure post facto anyway without explaining why.
One may agree or disagree with the particular platform of two of the disqualified candidates, but that platform is not the issue here. The issue is that the OSI held an election and then neither honored nor published the actual results.
The OSI is clearly aware of this: their updated blog post about the results makes some reference to those in-flight procedural changes. You can read it yourself and compare it with the factual allegations in the petition. I do believe you should take the latter as firm findings of fact: I've known candidate @nprofile…yat9 for many years and I cannot imagine him ever knowingly misrepresenting or misreporting a particular sequence of events that he was a witness to (and note that the OSI's post does not contradict the petition -- it merely presents things in a different light).
As I said at the start, even good organizations make mistakes sometimes. This was such a mistake, and I hope they'll quickly fix it in the most direct way possible.
I believe it is still the case that these elections are technically only advisory anyway. Under the bylaws, the Board ultimately decides who serves. Thus what's been happening in OSI elections all along is the Board has simply chosen to rubber-stamp, and thus officialize, the results of the member-driven elections. This is a fine arrangement; if the Board wants to decline to rubber-stamp a particular election result, it should announce that decision and explain why. But by refusing to acknowledge that that's what they're doing, and by not publishing the actual results, they're just temporarily silencing a signal that is unlikely to go away.
My opinion is that it would be much better for them to let the results stand, get those candidates onto the Board (if they indeed won), and have a healthy debate over the proposed reforms and all the other matters that that signal was signaling so loudly about. But that's just my opinion; if the Board feels that the presence of those candidates would be bad for the organization, then they don't have to seat them. But they should own that decision publicly and not hide behind false proceduralism.

2025-04-17 19:54:07 GMT

Can we all agree that the "I agree to use electronic records and signatures." checkbox on DocuSign and other similar services represents an obvious bootstrapping problem?

2025-03-12 03:28:18 GMT

- reply

@nprofile…yat9 Some of the reactions in there are... I don't even know where to start. Bernie Sanders ran for the Democratic Party's Presidential nomination without joining the Democratic Party. (IIRC, on a second run later he did join the party -- because <drum roll> they'd instituted a new rule requiring it.)
I mean, what you're doing isn't conceptually hard. If someone wants to oppose your candidacy on substantive grounds, then fine, but this weird, incantatory, nominalist officialism-that-isn't-even-official does not fly.

2025-01-16 04:29:51 GMT

- reply

@nprofile…s06p The big question is deliverability. When I've tried this, I've ended up spending lots of time trying to convince -- sometimes successfully, sometimes not -- large providers that the emails coming from our service were not spam (which they 100% were not, by the way). The burden of those interactions was very high, but the price of not engaging in them was that our users could not send email to anyone with, e.g., a Hotmail address.
How did you solve this? Or do you not have this problem? Or do you have it but tolerate it?
We knew about SPF/DKIM/DMARC, and AFAIK implemented everything correctly. I'm sure that helped -- our delivery rates would have been much worse without that. But still, some providers are just eager to reject, and yet sometimes people you want to reach are using those providers.

2024-12-31 17:48:15 GMT

These days, whatever I'm doing, I just automatically ask myself whether I should be using SQLite. Opening a jar of pickles with a tough-to-turn lid? We've all been there -- try SQLite! Retrieving a ball that rolled under the couch? Sounds like a job for SQLite. Car engine running a little low on coolant? It wouldn't be doing that if the manufacturer had used SQLite...

2024-12-27 02:55:36 GMT

- reply

@nprofile…w8ng Ok, but I still didn't know what memo I missed. (Of course, given that I've pretty clearly signaled that I don't have an open mind regarding what that memo is apparently saying, perhaps it doesn't matter that I remain otherwise unfamiliar with it!)

2024-12-27 02:32:07 GMT

- reply

@nprofile…w8ng Wha-hut? Did I miss a memo? Is someone proposing that Wikipedia (to which I donate, and intend to continue donating, monthly) is not a good cause?

2024-12-20 21:50:43 GMT

- reply

@nprofile…qtsx I happened to see this post of yours and the one immediately before it together, the one right after the other, in my timeline, so I treated them as threaded. But in my fedi client, they weren't actually shown as threaded, which makes me wonder if you had explicitly threaded them or if you just posted one and then the other? (I guess I'm asking for a technical detail so I can't decided whether I have a bug, basically.)

2024-12-19 18:12:12 GMT

These are tests. Suing the pollster who (incorrectly, it turned out) showed him being down in Iowa right before the election. Getting the GOP to agree to push for an FBI probe of Liz Cheney for the second impeachment. All of it is really just Trump probing to find out the boundaries of what he can get away with. Will a court dismiss the lawsuit in 5 milliseconds, as it should? Well, maybe, but maybe not, and hey, why not try? It'd be great news for him if he were to discover otherwise -- truly important information to have, if you're trying to know all the levers that are available to you to punish your enemies. Will the FBI investigate a U.S. Representative for conducting her official duties in an above-board manner? You and I may hope not, but for Trump the best news would be that the FBI is willing to go there -- not so much because he wants to punish Liz Cheney (though he wants that too) but because then he'd know he has a very powerful, versatile tool at his disposal for threatening and punishing more people.
Expect to see similar tests involving the military very soon after Trump's second inauguration. Days or at most weeks, not months.

2024-10-30 13:56:52 GMT

Rick Perlstein is right about the stakes here:
https://prospect.org/politics/2024-10-30-what-will-you-do-trump-wins/

2024-09-03 14:39:00 GMT

Y'all know there are two kinds of database programmers: you got row people, and you got column people.
You know how you can tell them apart immediately?
Watch how they eat corn on the cob.

2024-07-30 20:51:29 GMT

“I used to understand computers. Why’d we have to make computers more like people?”
– Frank Duncan, programmer, after a session of prompt-engineering in his first LLM project

2024-07-09 04:35:31 GMT

I changed /etc to /etsy on my computer and now the whole filesystem has a much homier, more handmade feel. I highly recommend everyone try it!

2023-10-09 18:18:35 GMT

Why doesn't Shake Shack sell shakshuka?

2023-10-05 22:16:14 GMT

- reply

@npub1hyk…3p33
> ...which is why I wrote an entire book and not just a tagline.
Best reply seen on Internet this week, right there.
/CC @npub10a0…epke @npub156n…mvgw @npub1huv…7ssu

2023-09-05 16:53:46 GMT

Call for information: We’re looking for examples of delayed open source publication (DOSP).
That’s when software is published under a proprietary license at first, then later open sourced. For example, at the time version N goes out, version N-1 might be released under an open source license. (But that’s just one way to do DOSP – there have been others.)
We’d like to gather as many examples as we can, both historical and modern, for a whitepaper that will examine the effects of DOSP on open source projects and on open source as a whole. The paper will take no position in the paper on the desirability of DOSP; its purpose is to provide accurate historical description and objective analysis.
If you know of examples of DOSP, please email us:
[email protected]
You can also file a ticket or a merge request directly, if you prefer (account creation is free).
You can see our notes and our current examples list here.
Thanks!
Please feel free to rebroadcast/repost this wherever it’s appropriate.

2023-08-03 19:38:50 GMT

Just mistyped "interface" as "intefarce" and thought for a moment "Hmm, should I just leave that as is? Maybe it's more accurate."