Karl Fogel

Home page: http://red-bean.com/kfogel/
Fediverse:
- <Karl Fogel (npub10r8…gh0u)> (I normally tweetidentitoot from here)
- <Karl Fogel (npub1swk…la06)> (my blog, which is also Fediverse-enabled)

Public Key

npub10r8whdz778q30dxgaj0wl5a6xtcrpchc7fqk9ffqvl8eqveqmm2q45gh0u

NIP-05 Address

[email protected]

Profile Code

nprofile1qqs83nhtk300rsghknywe8h06war9upsutu0ystz55sx0nusxvsda4qpzemhxue69uhhyetvv9ujuurjd9kkzmpwdejhgqg5waehxw309aex2mrp0yhxgctdw4eju6t08dtdp0

Publishing to

relay.primal.net

relay.damus.io

nos.lol

Author Public Key

npub10r8whdz778q30dxgaj0wl5a6xtcrpchc7fqk9ffqvl8eqveqmm2q45gh0u

Show more details

Last Notes

2026-07-24 19:58:32 UTC

I just noticed that you write things down when doing a write-up, and hereby officially give up on making any sense of English preposition usage.

2026-07-08 21:16:27 UTC

The Wikipedia page about [levers](https://en.wikipedia.org/wiki/Lever): much, much better than I expected it to be, and I even expected it to be good!
I ♥️ Wikipedia beyond all normal ♥️.

2026-05-18 05:14:04 UTC

- reply

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.

2026-03-07 16:11:53 UTC

Arrrrrgh. Signal Secure Backups -- a great idea, and I'd love to sign up. Unfortunately, the signup process includes a screen (which Signal prevents the screenshotting of and I don't feel like bothering with the physical moves necessary to capture it via the analog hole) that says:
```
Signal Secure Backups
Signal Private Messenger
CCName-CCDigits
+--- [email protected] ------+
| Enter your password |
+------------------------------+
```
There is *no* indication of whether to enter my password for
* Signal itself, or
* The credit card account indicated by CCName-CCDigits (which happens to use the same email address shown), or
* My email account at example.com.
Since I don't want to expose any of these passwords to any of the other services (because then I'd have to scramble to change a password, and anyway I especially wouldn't want to expose my Signal password even for a moment to some other service), I am stuck. I would happily pay Signal $2 a month for secure backups, but this UX has defeated me. I do not know what they are asking for.
Sigh.

2026-03-03 20:44:07 UTC

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](https://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](https://en.wikipedia.org/wiki/Source-available_software) 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.

2025-12-18 20:14:09 UTC

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 @npub1fdr…lvhs for the link.

2025-11-22 23:35:00 UTC

- reply

(The expression is understandably out of fashion, but I'm almost always willing to make an exception for a punchline!)

2025-11-22 23:33:27 UTC

- reply

A witch doctor?

2025-10-07 16:06:43 UTC

- reply

I had *completely* forgotten about that word until you reminded me! Thank goodness, through cooperation the full pun has been explored.

2025-10-07 16:02:06 UTC

- reply

I think I'd better call @npub1hv5…xq0v's attention to this. Yes. That's obviously what a well-meaning friend would do.

2025-10-07 15:57:09 UTC

Ohmygodohmygodohmygod everybodeeeee! Hey everybody!
Ray Bradbury's "Fahrenheit 451" -- unauthorized sequel:
***"Censor Reheat"***
Thanks. I'll see myself out.
(It helps if you have recent-vintage consumer microwave oven at home.)

2025-09-03 06:42:48 UTC

- reply

Wow, and eeks. Reminds me of one of my favorite quotes:
*"One of the ways the system protects itself is to describe as deviant behavior describing the behavior of the system."*
(from Patrick McKenzie / `patio11`; more context [here](https://kfogel.org/notice/AqoaMiZ2HUBD5AHyim))

2025-07-01 16:00:22 UTC

Dear Lazyweb (Emacs Subcommittee):
I'm a little surprised. While I've never been a power user of `dired`, I use it occasionally, and today I tried to do something that I thought would "obviously" be a feature in `dired`: show only entries matching a certain wildcard pattern. Apparently this is not a thing?
I found a directory into a buffer -- so far so good. There were many things in that directory, but I was only interested in the LaTeX files, that is, files matching the glob pattern `*.ltx`. So I went looking for a filtering command. I just assumed there would be such a function, bound to some key. But I don't see it anywhere, nor can I find this feature described in the documentation.
I am aware of `k` to run `dired-do-kill-lines`, which does the opposite (or the complement) of what I want. I'm also aware of the separate command `find-name-dired`, which does *exactly* what I want but is a separate command from `dired` and is not something one can use to further refine/filter an existing Dired buffer.
But does this functionality exist in Dired proper? Am I actually going to have to write this? It's kind of surprising: Dired Mode is older than several of the people who work at my company. I'm quite surprised it doesn't have this feature already. Am I missing something obvious?

2025-06-18 03:23:17 UTC

Is the fact that "X is a feature not a bug" has now fully entered the non-techie lexicon a featur...
You know what? Never mind. Just, never mind. Let's leave it right there. Have a good night, everyone.

2025-06-12 17:14:01 UTC

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.

2025-02-07 20:39:28 UTC

Dear Fediverse,
What's your favorite FOSS support ticket tracking system? (Note I really do mean *support* tickets, not the kinds of bug-tracker tickets a software development team uses.)

2024-12-31 17:48:15 UTC

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...

2024-12-11 16:44:32 UTC

- reply

I've most often heard the phrase "Benevolent Dictator For Life" used to refer to founding leaders of open source software projects, and in *that* context there is a built-in mitigation of the dangers of a BDFL: freedom to fork.
At this point my best option is to [quote myself](https://producingoss.com/en/social-infrastructure.html#forkability), so... no point being embarrassed about it I guess! Here we go:
> The potential for forks is the reason there are no true dictators in free software projects. This may seem like a surprising claim, considering how common it is to hear someone called the "dictator" (sometimes softened to "benevolent dictator") in a given open source project. But this kind of dictatorship is special, quite different from our conventional understanding of the word. Imagine a ruler whose subjects could copy her entire territory at any time and move to the copy to rule as they see fit. Would not such a ruler govern very differently from one whose subjects were bound to stay under her rule no matter what she did?
Someone who is a BDFL of an organization, or of some other set of physical / financial / organizational resources that can be owned and controlled is in a very different situation from that, of course. I just think that distinction is important: FOSS project BDFLs are not the same as other BDFLs.

2024-11-05 16:06:31 UTC

https://kfogel.org/media/9d97810d-9df0-438c-861b-043e81d71055/clk.png

2024-11-04 04:51:14 UTC

- reply

Do we know what percentage of the British population as a whole is descended from William the Conqueror's army?

2024-11-01 07:56:04 UTC

- reply

It's sometimes important when you publish a proof-of-knowledge hash to come back and publish the plaintext within a reasonable amount of time. Otherwise you might, for example, be posting hashes for mutually contradictory predictions, and then after a given prediction comes true you just post the plaintext for *that* hash while not showing any of the plaintexts for your wrong predictions.
I wouldn't want anyone to think I was doing anything like that! So here's the plaintext for my hash post of October 7th:
```
$ cat foo; echo ""
/me awaits the inevitable local emoji...
$ sha256sum foo
0dac35d5bb2ea7453647e857ff1e115e6e5c2e07f709d1e72f5eaec58d3001fd foo
$
```
The context is too intricate for me to explain, so I won't. I'm just following up here now to retain credibility for any future proofs-of-knowledge I might post.

2024-10-12 18:31:00 UTC

Adulthood is when you realize that nobody knows how to fold a fitted sheet.
Wisdom is when you realize that there *are* people who know how to fold a fitted sheet, but you're just not one of them.

2024-10-08 02:40:47 UTC

0dac35d5bb2ea7453647e857ff1e115e6e5c2e07f709d1e72f5eaec58d3001fd

2024-10-01 16:47:26 UTC

- reply

Excellent post -- the distinction you describe deserves more attention! I've made [this update](https://viewvc.red-bean.com/producingoss/trunk/en/ch07.xml?pathrev=3312&r1=3312&r2=3311#content) to point to the post.
@[email protected] @[email protected] @npub14hs…swac @npub1saa…xm6g @npub19yk…vv0j @npub1yx9…zzja @[email protected] @npub1qfg…mum6

2024-09-11 14:21:46 UTC

- reply

I know. I'm just a waif-like naïf, floating innocently through the sulfurous hellscape of modern web UX, wondering why our browsers needlessly betray us -- but also wondering why we blame web sites (who are naturally often our adversaries) more than we blame browsers (which should be on our side).
C.f.: ["Privacy Promises and Client-Side Betrayal"](https://rants.org/2013/06/privacy-promises-and-client-side-betrayal/)

2024-09-11 05:02:04 UTC

- reply

Why do browsers allow code received from web sites to do that? I mean, sure, the sites shouldn't *try* to do it either, but also the browser should be easily configurable to not let keys be redirected if the user doesn't want them to be (and maybe the browser should ship with a default configuration that protects certain key commands).

2024-09-09 03:22:11 UTC

Is it common for attorneys general to attend State of the Unions?

2024-08-07 17:38:01 UTC

Found an accidental example of the Stroop test in the wild, on the [MacArthur Foundation](https://macfound.org/) home page!
https://kfogel.org/media/f06dd254-e076-4939-a93c-a8bba87ed81f/macfound-home-page-2024-08-07.png