miggol

And wondered if tailscale was doing a bit more magic than tinc is here?

Yes, tailscale, rayfish, zerotier and all use an existing network of relays to do nat traversal. Tinc doesn't provide that, but allows you to be completely independent if you get (or already have) a $5 VPS.

Considering the guideline that QUERY should be idempotent and cacheable, file querying requests come to mind. Reverse image search, for example.

It's the combination of geographical data (maps) linked to its visual representation in the world (footage of structures, roads, landscape features) that is useful.

The geographical data already exists in digital maps. And I would expect competent militaries already have maps of enemy territory. It's the second part that was so far missing.

This combined set allows the training of AI models that can say, "When my surroundings look like x, that looks like y on a map".

So when your drone's GPS gets jammed, it can look at its surroundings, reference its (internal and offline) maps, figure out where it is, and navigate.

From the FAQ: "What is SDF"

While we did initially start out on a single computer in 1987, the

SDF is now a network of 8 64bit enterprise class servers running

NetBSD realising a combined processing power of over 21.1 GFLOPS!

Which piqued my interest about how that compares to today's computers. nVidia's venerable 1080Ti from 2017 measures about 11300 GFLOPS, or 11.3 teraFLOPS. About a fifty times increase.

Yeah I read further down in the comments that Aeon and Kalpa have actually diverged quite a bit, the installer might be one of those divergences. How are you liking Kalpa?

The main benefit of sleeping to RAM is of course the resume speed, which makes it more suitable for when you just left your computer inactive for 15 minutes. That goes double if you use encryption without TPM unlocking.

For leaving your computer overnight, hibernate wins on all fronts. I'm enthusiastic about sleep-then-hibernate schemes, but haven't gotten them to work on my devices yet.

Been running Aeon, the Gnome version of Kalpa, on my personal laptop for about six months now. I came from Tumbleweed so the learning curve wasn't steep. Overall the experience has been good!

The one major issue I had from the start was non-free Bluetooth codecs like AptX. That required me to taint the base image and add a non-official repo. It was messy but that was mostly down to it being a learning process, if I had to do it again I could probably do it with a single run of `transactional-update shell`.

The installer is super minimal and surprisingly user-friendly. One thing I remember is that there was zero partitioning choice: just use the full disk for encrypted btrfs and you get no swap (but zram swap is on by default). If you use OpenSUSE with secure boot enabled (as intended) then hibernate is prevented by `kernel_lockdown` anyway.

Snapper by default is nice, but you also get that with Tumbleweed. I ran into no applications that I couldn't get from Flatpaks or export from a distrobox, the latter being mostly for obscure stuff I need to compile myself. And my main toolbox hosts my Emacs environment that I spend most of my time in besides Firefox.

It's hard to recommend a MicroOS desktop over Tumbleweed, the latter being a great all-purpose distro as it is. But I'm hoping the benefits of forcing this "rootless" paradigm on myself will appear when it's time to move to a new machine. Just copy over my home directory and distroboxes and I'm golden, I could even switch to ARM without hesitation.

The distroboxes help with migrating because if I want to compile a newer version of that obscure program from earlier, I don't have to hunt down all the arcane requirements again. They're all still there waiting for me, in a Fedora/Ubuntu/Arch/whatever distrobox, depending on what works best for that program. At least that's the theory.

Happy to answer questions.

Wow, it feels like this argument rewired my brain.

When I first read about the chardet situation, I was conflicted but largely sided on the legal permissibility side of things. Uncomfortably I couldn't really fault the vibers; I guess I'm just liberal at heart.

The argument from the commons has really invoked my belief in the inherent morality of a public good. Something being "impermissible" sounds bad until you realize that otherwise the arrow of public knowledge suddenly points backwards.

Seeing this example play out in real life has had retroactive effects on my previously BSD-aligned brain. Even though the argument itself may have been presented before, I now understand the morals that a GPL license text underpins better.

Yes, the entire point is that requiring Google's permission to sideload anything is very bad.

The linked post by F-droid additionally points out that even that very bad case is not certain. We shouldn't trust that Google will even allow sideloading at all based on their words on their own blog.

Media has a responsibility to report that there is no evidence that Google will even allow anyone to sideload at this point.

Just wanting to say thanks to the whole team for creating such an inspiring and useful creative tool!

I'm most excited to try the perceptual analizer, which was something I found always had disappointing performance in plugins.

Which of the new features would you say posed the most interesting engineering challenge?

No need to be rude. The person above is adding a new insight to the conversation.

Vertical integration makes it possible but motivation makes it happen. Where is Samsung's ultra LTS Exynos device?

Good point, considering the nature of floppies I suppose it technically mustn't have been immutable. But I feel like it would have been wise to mount your OS root read-only to prevent yourself from accidentally ruining your (possibly only) copy of your OS. At least before you had a reasonably sized hard drive.

I'm surprised that, having something immutable, you'd choose to go the other direction.

I can somewhat imagine that having been limited by space and having to swap out disks all the time one would jump on the train of mutability without fully appreciating the benefits of immutability.

I've only used modern immutable Linux (Alpine, MicroOS) and wondered why of all places `/var/` was chosen as the location for rw stuff. It's fun to be reminded that there was of course a time when an immutable OS was the default, and you ran it off of floppies. So there's a lot of history to using `/var/` for that. Guess we've come full circle!

I think the issue with dependencies is a bit overstated. There's uv for that. And shell scripts have dependencies as well! You're not guaranteed to have bash, jq, and curl on a minimal install.

That being said, for real portability of programs that have slightly outgrown the script moniker I really like Janet.

https://janet.guide/scripting/

It can compile your script to a static binary for distribution.

But what actually being banned is accounts, not access.

Is this implied, or is this detailed in the law? I can see why this would make sense. So many businesses just have a link to their facebook page as their business website, which you should still be able to view. And presuming platforms like YouTube fit the definition, banning kids from watching anything on there would be pretty rough.

One might argue that this removes the algorithmic feeds. But I would wager that social media companies will just use browser fingerprinting to continue to serve algorithmic content to logged out users, if they aren't doing this already.

On the subject of YouTube, I wonder if they would allow for the creation of teen accounts, which restrict all social media functions but allow subscriptions. But would that then also remove algorithmic recommendations? What about data harvesting off those accounts? If so, I might have to look up how to get a teenage fake ID.

And what country is that? If you're comfortable sharing.

Of course the coin thing somewhat relies on coins of mentionable value being in common circulation, which rules out the US from the start.

Here few people still carry coins since contactless, but most still have a fake euro coin somewhere on their keyring or in their car for shopping carts.

They hand out these fake coins, which are usually branded with the supermarket's logo, for free inside the shop. But the system still works because having to go inside for a new coin negates the saved effort of not bringing back the cart. And there's probably some feeling of discomfort associated with abandoning the cart with your coin still in it.

Do people not return carts in your country? I enjoy Cart Narcs but they would have a hard time making content in The Netherlands.

We do have small parking lots here, and many carts have a coin deposit mechanism. It's not just that we're a morally superior people.

I wouldn't be surprised if astral's next product would be something like this. It's so obvious and there would be much interest from the ML crowd.

My current hobby language is janet. Creating a statically linked binary from a script in janet is trivial. You can even bring your own C libraries.

This invites a game where models have variants with slightly differing system prompts. Don't know if they could actually sample from their own output if instructed, but it would allow for iterations on the system prompt to find the best instructions.

I wonder if these will get better over time. Fun idea and I kind of want to join a table.

For now at least, some can't even determine which hand they have:

LLAMA bets $170 on Flop "We have top pair with Tc4d on a flop of 2s Ts Jh. The board is relatively dry, and we have a decent chance of having the best hand. We're betting $170.00 to build the pot and protect our hand."

(That's not top pair)

The author dedicates an entire paragraph to how much they trust the Obsidian team. It isn't open source purism, they are warning users that good intentions don't prevent a developer from writing software containing vulnerabilities.

Usage of user-created plugins and access to cloud accounts aggravates the risk posed by a vulnerability.

Open source reduces vulnerabilities over time, so those who want to heed the author's warning may indeed want to switch to an open-source Markdown editor. Just not because the Obsidian team is Evil.

Love the idea but the thing is, if it's just one file then it's probably easy enough to work that contribution out of the repository.

It is quite fun to try and think of ways that this could work though. Perhaps a bot that code-paraphrases (paracodes?) every accepted PR. Or maybe there's some crypto magic you could do to make the only option a clean room rewrite.

Qubes really is the trailblazer in this regard. You can get pretty close with distroboxes on Linux as well.

When a project requires a certain Python version a virtualenv suffices. But when you need a specific Python and NPM version then I might as well make a distrobox. Set a custom home and the project is isolated, speaking only to my IDE over LSP, and also to my web browser I suppose.

This only protects the developer themselves of course, but it's a start.

Thanks for sharing that link. The categories suggest that this system is mainly used to give priority to certain kinds of vehicles. Public transport and emergency vehicles already had a system to turn lights green, so this seems like an addition to extend the system to "milder" priority vehicles like construction without giving them access to the previous system.

The location of the lights doesn't exactly point to usage by cyclists.

As far as I know, almost all traffic lights in the Netherlands work with magnetic sensors under the road for presence detection. Then there are a few optical cameras for congestion detection at specific intersections.

The magnetic sensors don't pick up ultralight or carbon road bikes, but bikes usually have an additional push button anyway.

I have been using an untranslate add-on like this one and have been absolutely loving it. Since YouTube has been AI translating video titles and dubbing contents it has been suggesting foreign videos in users' home feed.

The way these untranslate add-ons work (layman's explanation) is that they fetch the original title and audio and reinsert it, but the recommendation for the video stays in your feed. This has resulted in loads of super interesting foreign language content in my feed which is just awesome.

Cars are one of my YouTube interests and seeing loads of cool old car content from different parts of the world has been fascinating. Not only were different models popular in different places but the things people value in a car are also wildly different across the globe. And I get to listen to a cool foreign language while discovering this!

One downside is that to the YouTube algorithm, it probably seems like I absolutely LOVE this autodubbing feature, going crazy for all these translated videos. That could not be further from the truth: my youtube feed has become completely unusable without an untranslate add-on since this update.

I don't want to defend Django here, surely this should be categorized as a bug. But on the other hand, for this situation to come up you have to be the following:

- The kind of person to dive into the schema and worry about an unnecessary index

- Smart enough to heed Django's warnings and use `Meta.UniqueConstraint`

- Dumb enough to ignore Django's warnings and not use `Meta.Indexes`

I think it's funny that the kind of dev that 100% relies on the ORM and would benefit from this warning would probably never find themselves in this gritty optimization situation in the first place.

That being said, I enjoyed the article and learned something so maybe I'm the target audience and not them.

That does make sense: an application server like Gunicorn talks ASGI with the application, but Gunicorn itself speaks plain old HTTP with the outside world (or reverse proxy). No more need for FastCGI inbetween.

That is in line with what the article is saying. Thanks for clarifying.

Similar protocols like ASGI for Python are still ubiquitous. I assume because they offer separation of concerns and as a result good performance.

You would think the separation argument would still hold true for compiled languages, even if performance is no longer as relevant.

On top of this inaccuracy in actual measurements, many fitness tracker brands invent their own metrics.

"How did you sleep?" I asked a friend who stayed over, - "I don't know" they replied as they raise their smartwatch to their face. "Pretty well, apparently! I got a sleep score of 84."

I've also heard of an acquaintance that they cancelled plans because their (Garmin™) Body Battery was too low.

Instinctively I flinch whenever people place such importance in what is essentially an arbitrary techno-horoscope. But I also find it hard to make an argument against it as long as it inspires healthy behaviour.

Group stories are great fun and a feature I seriously miss on Whatsapp. They work well from meme chats to family groups.

There being a killer feature that Whatsapp users are missing out on won't convince everyone but it sure makes me feel less like a nerd when encouraging the switch to Signal.

I find it quite funny that such an obvious feature likely hasn't been added to Whatsapp yet because Meta thinks Instagram is for stories. That's pure speculation on my part though

All this marketability tuning may some day result in models which are extremely finely attuned to our current societal norms and taboos.

At which point the models will stop reinforcing (racial/gender) biases and start reinforcing said taboos instead. I don't think anyone wants that either