Matumio

I know some European "automatic scribe" projects of the government sector. Their IT buys a physical GPU server hosted locally. Pretty sure it wouldn't be accepted otherwise (or maybe I'm just naive, but it sure is a topic they care about). The software stack is mostly open source, I think. It sure as hell doesn't talk to a big American cloud provider. (Well, the transcription service doesn't. Who knows what they do with the automated transcript. Probably the same thing they did with the manual transcript.)

Well said. There is a social context, there is a process and a struggle that can be more important than the result. It is sad to reduce art to the final product, or to approach it with an industrial mindset: maximizing commercial value while minimizing effort.

I can't write well. Let someone say it who can: (Ursula Le Guin, 5min) https://www.youtube.com/watch?v=s2v7RDyo7os&t=337

My doctor probably thinks we software developers do a very narrow job. And she is kind of right, we always turn up with those back problems from sitting too much, or RSI or whatever. While doctors have all those medical specializations and different roles and employers.

If it can go online, I'd prefer to use an android work (or user) profile with only auth apps in it, and nothing else.

As a separate device, it should be offline always IMO, and perhaps the size of a passkey. Or one of those banking devices with a display that show an authenticated text saying what you are confirming.

I don't know if it will work, but Pl@ntNet Identify (which I use often) seems to have an API: https://docs.plantnet.org/en/reference/api-plantnet/

Saying how many lines of code you can write this way is also a bit like bragging that you are building world's heaviest airplane.

Say you have an ad-blocker and you don't allow it to touch your forms. Five years later, the ads have moved all into form fields.

Never mind the technical challenge to allow doing anything with the DOM but disallow reading the forms. Like, prevent the forms leaking its text when you do funny things like testing character width via line breaking or font changes.

AGENT SMITH: And tell me, Mr. Anderson, what good is DRAM if you are unable to boot?

The question unnerves Neo and suddenly he feels his phone vibrate as it unexpectedly reboots. The standing Agents snicker.

In Switzerland, on some trains there are trilingual announcements: German, Italian, French.

Disagree, Linux is too big to fail. Too many people depend on it. It may get chaotic, but worst-case distributions will start collecting patches, as they already do for many unmaintained projects. Eventually one or two of them will emerge as the new upstream.

Remember the venomous, desperate BEEP! when the keystroke buffer was full. (Or was it when pressing too many keys at once?) Like a tortured waveform generator constantly interrupted by some higher-priority IRQ. Good times.

I've had password login enabled for decades on my home server, not even fail2ban. But I do have an "AllowUsers" list with three non-cryptic user names. (None of them are my domain name, but nice try.)

Last month I had 250k failed password attempts. If I had a "weak" password of 6 random letters (I don't), and all 250k had guessed a valid username (only 23 managed that), that would give... uh, one expected success every 70 years?

That sounds risky actually. So don't expose a "root" user with a 6-letter password. Add two more letters and it is 40k years. Or use a strong password and forget about those random attempts.

I once had weird results with searching specifically in the Switzerland region, it didn't find an obviously Swiss site. IIRC it was solved it by switching back to international search. I'm using Kagi exclusively, and I don't remember having such trouble recently. Maybe they fixed it.

I just did a quick test: local search for a specific law term. Kagi, Google and DDG all found the roughly same relevant sites in the top five. Each has a different top result. Google's and DDG's are a private law company. Kagi's first is an official government site. (With a suspicious non-government domain, so I had to check, but yes it's prominently linked from the main government site.)

You may also like this how this mostly hand-designed CA rule can produce plausible mutations when disturbed: https://www.youtube.com/watch?v=qwJNeq-WABU ("From One Cell to a Multicellular Organism", Part 2, 25min, Simulife Hub)

Those displaced workers need an income first, job second. What they were producing is still getting done. This means we have gained freedom to choose what else is worth doing. The immediate problem is the lack of income. There is no lack of useful work to do, it's just that most of it doesn't pay well.

Uh, a company not paying money for something they can legally use for free? There are so many MIT-licensed software libraries that everyone is using in a critical place, for profit, with zero money flowing back into the ecosystem that created them. It should surprise nobody, it has been like this for over a decade now.

I don't think pointing out something that goes wrong under the current flavour of capitalism is the same as being against capitalism. Similarly, reporting a bad police officer doesn't really mean I'm against having a police, it can also mean that I want them to do a better job.

That, and the cognitive load. You need to buy the right amount, remember where you stored the $5 replacements, or else spend $100 worth of your time to figure out where you ordered from five years ago. And if they are no longer available you need time to figure out which of the replacements isn't total crap.

Only one way to find out... float some ice cubes in a glass of water and observe.

(Edit: I'm back to report the results. There was either no change in the water level, or a change below my measurement tolerance ;)

(Edit2: Here is a more serious take of that experiment: https://skepticalscience.com/Sea-level-rise-due-to-floating-...)

Given those time frames, maybe don't send primates. Send a computer babysitting a diverse zoo of bacteria and algae, with a variety of landing devices and instructions in what order to deploy them under which circumstances.

Same problem: the best-case outcome is that we never hear anything interesting from that rocket ever again. But it should be a lot cheaper.

If you prefer short video lectures, Complexity Explorer also has a Random Walks tutorial: https://www.complexityexplorer.org/courses/46-random-walks

It's a bit more visual, but still gets quite math-y enough for my taste.

NTP will typically have about 10ms error which would be visible on this display. So for more authentic giggles PTP is recommended. With compatible switches to compensate for queuing delays and cable length. And if your provider doesn't give you a good time, put your own GPS-backed network clock on the roof. For science.

The post looks interesting, but I'm still building my triangle castle so I haven't read it yet. This probably means I should read it later when I'm less distracted. Anyway, already love the layout and the playful visualisations.

Now if we only could disambiguate words based on context. But you'd need a good language model for that, and we don't... wait.

Other people do, though (not mine): https://github.com/facundoolano/feedi

I'm not using that one yet, but I very much like their idea to highlight infrequent sources. If I hand-added the feeds, I feel they should by default have at least an equal chance of me noticing a post in them, without clicking through all the feeds.

I also like the idea of subscribing to everything and then privately filtering the for special-interest keywords. I'm not doing that currently, but I subscribe via RSS to some searches or tags. E.g. here via hnapp.com, or on StackOverflow you can subscribe to tags, or the RSS feed of low-traffic subreddits.

This kind of thinking, that everything can be broken down into pieces and studied in isolation... it has brought us very far, but it has some hard limitations. Especially in biology, where any leak you find may serve some function.

In medicine, the primary goal is to help, rather than understanding why something works exactly. Sure, understanding is an important goal too, it's just much harder to achieve than being able to help. And less important than knowing that your treatment will work, without any major side-effects, for the kind of patient in front of you.

Not so sure. Imagine you have a base64 encoded payload and it just happens to encode the forbidden word. Good luck debugging that, if the payload only gets silently modified.

I suddenly understand why it makes sense to integrity-check a payload that is already protected by all three of TLS, TCP checksum and CRC.

They don't exist. If you want out-of-the-box, use Jetbrains and pay for it. If you want free-as-in-beer, VSCode proper. For free-as-in-speech you'll spend hours to learn/configure neovim/helix/emacs, or even hunting down free VSCodium plugins, I guess.

I think working on out-of-the-box experience is not very attractive for volunteer contributors, so I guess the situation won't change unless we find a way to sponsor that work.

Agree, Undertale is absolutely brilliant in that aspect. Especially the beginning (the part that is in the demo version). The mood changes after that, for the worse I thought. Things got a bit more silly/naive than I like. The ending is absolutely brilliant again, tough, in the same way (it is a reflection on game mechanics). It is not 4h because you'll want to retry some parts.

Yet we have laws against fraud, rape, and so on. Where do you draw the line for those? There are some crystal clear cases, and there are unclear cases where you could argue forever.

So it is for advertising. You don't need to draw a clear line for every case before you can make a law.

I like how it turned out with email advertising, actually: spam is defined to be whatever people put into their spam folder.