goodusername

Seems promising, will have to try it out. It really is grating to read output from opus 5 currently. Never heard of the term "mannered prose" before. I'm constantly debating with myself whether to add more to my claude.md, or whether to simplify it before adding new stuff.

Under Questions: "which languages" only lists 7, with no possibility to view more, that I could find.

Yeah, would be good to have confirmation that this happened to others as well.

But it really seems like an own goal if true.

Although I'm slightly relieved there is a way out of Googles verification system, it's still pretty wild if you compare this to installing software on a Windows pc. I'm sure Microsoft is heading in the same direction with Windows, but today its still "only" a few confirmations to install anything.

This will sadly still put a major damper on adoption of open source apps, while giving a false sense of security that apps from the Play store are safe.

Years down the road, the low usage of apps installed from outside the Play store will be used as an argument for removing the functionality completely.

Very uplifting read in these times. See myself in some of these mental traps of sorts.

This looks pretty interesting from a laymans perspective. Can anyone with more in depth knowledge relate this to the current state-of-the-art auto regressive models? are diffusion models closing the gap in a meaningful way, or are they still in the experimental stage?

Thank you for this well written comment. I was not aware that passkeys are developing in the exact direction i feared, when i first read about them.

The court subtracted a fee for the artwork and for the display of the work, which he does not have to give back - so the artist did in fact get paid for his empty frame artwork. Now he just has to give back the money he lent for the originally agreed upon artwork that he did not create.

Seems like a fair deal, since he also got a ton of publicity out of it.

I wonder if this could be used to finally get universal audio streaming from the device, kind of like airplay. Perhaps using the received audio stream to create an internet radio station, which could then be played by most multiroom speakers, without needing specific support for each player app on the device.

Wow, this sounds like a really important decision. I wonder what Metas next line of defense will be, if any. The article says Meta can appeal, so I guess they will try that.

It will be interesting to see how many users will decline to opt-in to Personalized ads. Meta will probably still try a range of dark patterns to keep users from making that choice.

To avoid syncing my calendar and contacts to the cloud, I'm using wireguard as a always on VPN on my phone, to connect to a locally running CalDav/CardDav server. Wireguard on Android can be setup to only affect specific apps, so my DavX app is the only one using the VPN.

You might be interested in https://radicale.org/v3.html. Runs on my odroid board with 26mb memory. The documentation is particularly good. I've used it as a replacement for the Synology CalDav and CardDav services.

It's very easy to install and does not require a DB. As a bonus, it stores everything as files which can be read and edited manually. It does require python.

Thunderbird mobile will be open source, so they, or anybody else can fork it, and continue development independently of Mozilla, using a different brand. It will not be "Mozilla property".

I'm happy to support Thunderbird; it's great to see a healthy project supported by donations.

It would be pretty great if Firefox development could be supported by direct donations too. Although maintaining a browser is probably orders of magnutide more expensive than Thunderbird.

It was pretty amusing to try and guess how to smash the pieces into order, thanks!

This will be very dependent on the specific apps you need, but I have done almost the exact same process with some success. I switched to CalyxOS which has microG to "replace" Google Play Services. For banking, goverment ID, payment apps etc., this mostly seems to work. I use Aurora store to install and update the apps that are only in playstore, which sadly is still a lot of them. I Use f-droid as much as possible. The odd app that doesn't work, can sometimes be replaced by using the website of the service instead, even though the experience is likely worse.

Some banks or government services allows you to buy a physical token to generate TOTP, and that comes in handy if the app doesn't work.

Dowsing has in fact not been proven to work. There is instead overwhelming evidence that it does not work better than chance.

Wikipedia sums it up nicely, with citations: -- The motion of such dowsing devices is generally attributed to the ideomotor phenomenon,[7][8][9] a psychological response where a subject makes motions unconsciously. Put simply, dowsing rods respond to the user's accidental or involuntary movements.

The scientific evidence is that dowsing is no more effective than random chance.[10][11] It is therefore regarded as a pseudoscience. -- https://en.m.wikipedia.org/wiki/Dowsing

For an entertaining time, check out the late James Randi and his debunking of dowsing

I highly recommend The Skeptics Guide To The Universe: How to Know What’s Really Real in a World Increasingly Full of Fake https://www.theskepticsguide.org/our-book

Its somewhat of a reference book and very accessible

A really hard problem that often arises when doing E2E tests, is creating and managing test data.

If you have one or more integrations to external systems, where you cannot control your test data, it becomes much harder to write stable E2E tests.

Some don't have test environments, some have too few. Most don't allow you to setup data easily either way.

You can, of course, mock the external systems, but if they play a large enough part, your tests start looking more like integration tests again, but with the added overhead of something like browser automation.

It's a hard balance to strike.

To me, this seem like a fine initiative, all things considered.

There's a bit of a false dichotomy in some comments, where Mozilla Rally is preventing or hampering Mozilla from improving Firefox. They can and should do both at the same time, especially if both are positive things.

It doesn't mean all their initiatives are equally good, and that we shouldn't still criticize the bad ones. But this seems benign at worst, and could possibly do some good. it's completely opt in, transparent, and with very strong user control over the data.

Instead of chastising them for all their other shortcomings yet again, I prefer to be happy that this might benefit ethical research.

Wow,i never realized you could do that, such a game changer! Thanks for the tip.

Interesting. As a counter argument, linguistics and grammar also involves a great deal of abstraction, which seems closely related to OOP concerns.

But I will say that familiarity with equations and mathematical symbolics tends to be very useful for reading and understanding code.

Perhaps this also depends on the type of software being built. If you are solving complex problems with original solutions then math and problem solving might be very important, but if you are building more "mundane" software for lack of a better world, then readability and maintainability might be more important. And in my own biased opinion, those aspects might rely more on good language skills instead.

Lately I've been focusing more and more on writing code which is easy to read, and it seems to require good language skills more than anything else. Things like descriptive naming, keeping the same level of abstraction in a function, writing good commit messages or documentation all seem to require language skills rather than technical skills.

But it might just be my personal bias for what is considered good code that gives that impression.

"German de-Nazification would be called an ideological tyranny if it were ever objectively described, and it brainwashed a generation of Germans. The German soul of romantic discomfort with the modern world was murdered, with disastrous consequences for European culture. Liberal democratic cosmopolitanism didn’t win any arguments; it won a war, and got to impose itself by force."

It reads like the writer is trying to say something, but isn't really concrete about what that is.

Without any concrete criticism of "Liberal democratic cosmopolitanism" it's hard to know what this is trying to say. As others have done in the comments, this vague language leads one to guess what they mean. The principle of charity should be used, but it's hard when the same complaints are often used as a dog whistle by the extreme right.

What exactly is the problem with de-nazification? Explicitly violent and genocidal isms like Nazism should not be tolerated in good faith discussions. They should not be "debated" as Equal viewpoints.

To call it brainwashing is a stretch. Nazism has lead directly to millions of deaths in the name of Nazism and not something else. If that's not something we can agree to fight, then what is?

Yeah, signal is my preferred messenger app, so would be hard to go without. Might have to try this and see if they got it working. My countries Personal-ID-login app refused to open on rooted devices, and several apps for banking and public transportation relies on Google Play Services. There's a bit of way yet.

I'm continously impressed by the effort put into custom ROMs for Android. If this was supported on my phone in the drawer, I'd try it out in a heartbeat. Does anybody have experience using an OS with MicroG on a daily basis? A whole bunch of banking, work and government provided apps are still only available on the play store

Have only just turned it off so don't know yet, but my bet is on no. I'm right there with you on reflecting in peace though. Sometimes you want to just take a breath and think about what you saw. That same "feature" was one of the things that drove me away from flow TV, as they increasingly pushed the next program more and more aggressively, until it would actually start the trailer/speaker BEFORE the current program was finished. Drives me insane. Who watches a 2.5 hours movie, and as soon as the credits start rolling, goes "alright next movie, play play play!"

I have had thoughts around that same idea as well. Could be a start to just define a subset of existing good semantic HTML elements to use for a blog, a news site etc. My dream is to be able to browse a subset of the web, that only uses this kind of document format. Something like docuweb.example.com or whatever :)

I had Agner as a teacher when studying IT-engineering. At the time I didn't know anything about him really, but always found him pleasant and knowledgable. Looking back now, I wish I had spent some more time picking his brain about his expertise.

The webpage is awesome in my opinion. It's functional, simple and with a good amount of humor :) It a good example of a content-first approach.