suby

I don't know where else to write this, but I want to throw the idea out there. I have long wanted to take old broodwar televised matches, many of which are terrible quality 240p, and use machine learning to convert them to into perfect Broodwar Remastered frames. This seems tractable to me because you should be able to map the terrain sets to their remastered equivalents, and the game is just a series of sprites rendered at specific frames. Even if the source quality is terrible, I imagine this is able to be extracted at high quality since you can, eg, set up an automated pipeline which generates training data. Maps from original graphics to remastered, and then again for 240p -> tilemap positions for frame camera center + sprite positions / animation index.

I found the post offputting because he is insinuating a few things about the author which are clearly untrue (eg, unfamiliar with the formal verification landscape) and he's using this untrue speculation as evidence of the dangers of vibe coding. Even the edit where he says swap it out with a hypothetical person who fits the description - it's still leaving in the untrue claims about the author.

It doesn't seem unreasonable to think it mean if I were to make up speculative negative assumptions about you which undermined something you've been working hard on for a long period of time.

I know nothing about the author, I'm just watching this whole spectacle unfold.

Linux Mint mentioned on their blog that they would be using the Rust versions.

My problem with flatpak is that the underlying runtime has a limited lifespan. I used the Sublime Text 3 flatpak for many years, and it was deprecated a few months ago because the runtime it was using was no longer supported. This caused people to update the flatpak to Sublime 4, which is good, but I only have a license for 3. In the end I just downloaded the binary for 3 from their website, and it still just worked. Flatpak still doesn't save you from bitrot.

I have had my account since like 2013. I browse the site and hit like on things, but I only have a handful of posts going back over the years. Occasionally, when I do try to post, it stops me because they think I am a bot. Meanwhile, I see what are clearly and unambigously bots follow me all the time, I see identical messages repeated across obvious botnets. Their bot detection is just very bad.

Also, while we're on the subject, I used to have two factor enabled. After Elon bought the site, they forced me to disable two factor for insane reasons, iirc you needed to pay for a blue checkmark if you wanted two factor. Today for the first time they've started blocking my access to even view the site until I authenticate with a qr code from an android or ios device. The future of the internet is bleak.

You're comparing the challenge of building game engines in the 90's with the more open-ended AGI research. It's different enough that I don't think it's valid to conclude he's less productive now based on impact over time.

People point to the equivalent API costs to show that they are getting a great deal on the subscription, 10,000 dollars worth of tokens for 200 dollars. I do wonder if it's the other way around though - are the API users simply getting ripped off? I have seen Dario say in multiple interviews that they are profitable on inference, which maybe he was only meaning to refer to API usage, but that's not the impression I got.

It's not a 98% margin loss if your users are unwilling to pay 50 times the cost that they were previously paying, and if they have other options like open source providers. The calculus isn't so simple because some portion of users would switch to API, and so it's about how many would continue using the service rather than leaving for a competitor.

I'm aware they need to recoup the enormous cost of training and data centers, but on a purely inference cost level I'm not convinced that the 200 dollar plans are unprofitable.

I'm so jaded at this point. The AI translation from Bun to Rust doesn't bother me, I think it's interesting, but that this blog was so clearly written by LLM's is offputting for some reason. I think after having to interact with LLM's for much of the day, it's exhausting to read LLM speak in so many things I see online. It feels almost disrespectful to the reader. It's written from a first person perspective, but Jarred did not write these words.

I was looking forward to this blog post too, but in retrospect I don't know why. I could have had an LLM generate a hypothetical of what this blog post might have looked like and it would have probably been able to get close.

I feel like we've replaced unique voices on the internet with the same style / author, which might be more tolerable if the breathless LLM writing style wasn't so jarring. Contrary to the amount of times "But honestly" or "genuinely" is mentioned, nothing about having your LLM speak for you feels honest or genuine.

I know it's not cool to leave responses like this, but I'm really tired of all of this at this point. The ironic thing too is that it might actually be better to have LLM written text be so distinct so that you can still pick out when a human has actually authored something. Again, this is a blog post from Anthropic about having an AI translate 500k+ lines of code in 11 days, so I guess my disappointment is my fault for expecting otherwise.

Because it is general purpose. We did not have the ability to create a single robot form which could do all of these minor, finicky, and opened ended tasks. Now that seems within reach. The nice property of humanoid robots is that the world is already made for human form, and so if you're trying to replace people naturally this is what you'd want.

Hi Lunixbochs, thank you for the comment. One of my take aways from the article was that you were considering removing the ability for Talon to work on x11 in the future. I appreciate the correction. I understand that you're dealing with a lot of unpleasant support requests (especially because people who have to use Talon are probably frustrated at having lost the ability to use their machine normally). I am sorry if I added to that, I do genuinely think you've made the world a better place.

I also think it's unfortunate that Talon is closed source. You see this even in the support practices adopted, where all support is routed through a slack chatroom which doesn't let you view history older than I want to say a few months but might be wrong on the length of time. The author seems to want to force all support requests to go directly through him, presumably because it increases his income if he can create a direct connection with his users.

He's created an incredible piece of software, and that's entirely within his prerogative to do this, especially because him being able to work on it full time leads to more work going into the system. He's made the world a better place so I'm not trying to criticize too harshly. But it's also super unfortunate right, because now if I run into an issue with Talon I am unlikely to find a search result of someone else who has solved it, but rather I have to interact with the creator of the software in a silo'd manner that will not be useful to anyone else other than me.

Tthreatening to remove x11 support entirely (as the article alleges) is also unhinged, yes. We're in a situation in which the best accessibility software is being threatened to be removed from a working platform because the author is (justifiably) frustrated with support requests that he cannot fix because of the transition to Wayland.

I expect that sooner or later we're going to get a better solution to accessibility than Talon, I'm not sure exactly how but probably using local LLM's in a heavy way.

I read this yesterday wanted to raise awareness for it - https://nocoffei.com/?p=451

It describes the regression in accessibility software for Linux from x11 to Wayland. Unfortunately, judging by the pace of protocols being accepted, I think we're years out from having a solution.

The most notable thing not working is Talon, which is a voice input system that lets you insert speech to text, manipulate windows, call scripts, etc, all via voice. It's software that works on Windows, MacOS, and x11, but not Wayland.

I think unfortunately right now the best bet is to, if you need the software, stick with X11 for as long as possible. An environment like i3 will probably be maintained for decades to come. Alternatively it might make sense to build some type of bespoke solution on top of a specific wayland stack, like re implementing what you get of talon in a kde plugin or via sway IPC. This seems viable to me but an incredible amount of work.

For people that need this, having to be a developer and build your own tooling in order to use your computer... it's not a future of Linux I'm particularly excited about. I don't want to leave people who need accessibility software behind, and I don't think any security justifications are actually real roadblocks which would prevent being able to serve these people. We have a coordination problem. It's less of a technical issue and more of an issue of getting people to agree on protocols which would let software like Talon work against the entire ecosystem.

I am happy the ecosystem is moving to Wayland, I think we're going end up in a better place. Wayland does solve some real problems for me (x11 screen tearing / frame pacing issues on Nvidia). I'm happy that KDE exists, it's great software.

To be fair, they're allegedly experiencing under attack by the Iranian government. I doubt Guix or OpenBSD would be able to sustain service under a similar attack, though granted there's a reason Cannonical is the target and not OpenBSD.

https://arstechnica.com/security/2026/05/ubuntu-infrastructu...

This strikes me as a good idea I've never seen articulated before. Something like a sticky scroll which accrues all off screen cursors, limited to some max to prevent things getting out of hand.

This is a valid concern imo, but I'm not too afraid about nvim itself being compromised. I do think it is risky to be depending on many plugins, which is why I'm hoping nvim can integrate more of the popular plugins into nvim proper.

I am somewhat dismayed that contracts were accepted. It feels like piling on ever more complexity to a language which has already surpassed its complexity budget, and given that the feature comes with its own set of footguns I'm not sure that it is justified.

Here's a quote from Bjarne,

> So go back about one year, and we could vote about it before it got into the standard, and some of us voted no. Now we have a much harder problem. This is part of the standard proposal. Do we vote against the standard because there is a feature we think is bad? Because I think this one is bad. And that is a much harder problem. People vote yes because they think: "Oh we are getting a lot of good things out of this.", and they are right. We are also getting a lot of complexity and a lot of bad things. And this proposal, in my opinion is bloated committee design and also incomplete.

There is a roadmap and github issue tracking what is needed for 1.0.

https://github.com/neovim/neovim/issues/20451

https://neovim.io/roadmap/

Quality is not the issue. We should be more specific - Microsoft has been consciously employing dark patterns that they know will be harmful to users but they do not care because their incentives are misaligned. Employees of Microsoft have surely had meetings scheming of ways to degrade user experience for some internal metric they are trying to hit. A decade+ of conscious, willful decisions which negatively impact users.

I personally will never forgive them for uploading the entirety of my users dir to OneDrive without asking for permission. They're --still-- doing this. Whatever decision making process they have in place that not only cooked this scheme up, but allowed it to continue for years must be broken beyond repair. It's contemptuous, backwards, and hostile to users. It cannot be condemned enough.

This blog post talks about taskbar positioning and vaguely gesturing at quality, which is whatever. I'm not mad about removing features or even a higher incidence of bugs. I'm mad about hostile dark patterns that they have consciously chosen to employ at an ever increasing rate. I don't think you can fix this without drastic company wide changes.

For as long as I live, if I have a choice, I will avoid Microsoft products. They cannot be trusted.

> And how would they be able to "push stuff down people's throats" if people could walk away towards alternatives?

It's a forcing of their narrow opinion on what should be allowed onto the ecosystem at large, because all of these things are connected. You can leave to a different DE/distro, but if every DE is doing its own thing for global hotkeys or whatever, then software in the ecosystem is going to be hacky/bespoke or have an unreasonable maintenance burden.

Even if you in particular can move elsewhere the ecosystem is still held back. We only recently got consensus on apps being able to request a window position on screen, which is something x11, macos, and windows all allow you to do. CSD and tray icons are other examples of things found everywhere else that they did not want to support. Some applications are just broken without tray icon support.

This bleeds over into work for folks releasing software for Linux in general. By not supporting SSD they were pushing the burden of drawing window decorations onto every single app author, and while most frameworks will handle this, it's not like everyone is using qt or gtk. App authors will get bug reports and the burden of releasing software on Linux needlessly climbs again.

Hard to convey how unreasonable I feel their stance was on tray icons / SSD. It should be the domain of the DE from a conceptual but also practical point of view, even from just the amount of work involved. It reminds me of LSP's enabling text editors to have great support for every language. And again, Gnome was the odd man out in this, they want extra attention and work when Linux is the lowest desktop marketshare by far, and they themselves are not the overwhelming majority but they are large enough that you really do need to make sure your software runs well on Gnome even if you want to support Linux.

People think Gnome push stuff down your throat because they have the power and influence to impact the ecosystem, and they use that power and influence to die on absolutely absurd hills.

It's possible I have no idea what I'm talking about, but my understanding is that nixos relies on fetching things from third party URLs which may simply die. I feel a bit misled by the promises of nixos, because I cannot actually take the configuration files in 10 years and setup the system again due to link rot.

I was also under the impression that I could install DE's side by side on nixos and not have things like one DE conflicting with files from another DE, but this apparently isn't true either - I installed KDE, and then installed Sway and Sway overwrote the notification theming for KDE.

NixOS is very impressive but the marketing around it feels misleading. The reproducible claim needs a giant asterisk due to link rot.

I don't know why Blade was decided on, but it was started by Kvark who worked on WGPU for Mozilla for some time. I assumed it would be a good library on that basis.

I harbor similar sentiments, but I understand why OpenAI, Anthropic, Zed, etc begin with a macOS version. They're able to target a platform which is a known quantity and a good jumping off point to Linux.

I'm writing software for Linux myself and I know that you run into weird edge case windowing / graphical bugs based on environment. People are reasonably running either x11 or wayland (ecosystem is still in flux in transition) against environments like Gnome, KDE, Sway, Niri, xfce, Cinnamon, labwc, hyprland, mate, budgie, lxqt, cosmic... not to mention the different packaging ecosystem.

I don't blame companies, it seems more sane to begin with a limited scope of macOS.

I would have a very hard time accepting it to be true. They don't even strongly make the claim that it is true, they are correlating areas of the brain associated with memory as growing larger, and they are associating a larger brain area with better cognition, but later in the article indicate that there are some areas of the brain which are associated with memory which actually shrink.

I think we should question research which is overwhelmingly against our common experience of life. My memory is absolutely shot when I am consuming weed regularly. It's not particularly subtle, it is noticeably worse. I suppose there is room for a situation in which it is worse while I am heavily using, but if I were to cease maybe it will rebound and settle at a point which is better than it would have been had I not consumed any cannabis... but I don't see any reason to believe this.

The reason Wayland progress is slow is not technical. We have a coordination problem, people have differing priorities and views on what should be allowed.

There are people opposed to things like a allowing windows to specify their own bounds, and unless all the stakeholders agree to implement such protocols in their respective projects, the ecosystem will remain fragmented. Multiply this against every feature that people want.

This is from 2019, prior to the finalization of modules in the standard. I'd be interested in how many of these issues were unaddressed in the final version shipped.

I think if you were to poll people, a significant portion would be repulsed by this catgirl aesthetic, or (though this isn't the case for Anubis) the cliche inappropriately dressed inappropriately young anime characters dawned as mascots in an ever increasing number of projects. People can do whatever they want with their projects, but I feel like the people who like this crap perhaps don't understand how repulsive it is to a large number of people. Personally it creeps me out.

I don't think Helix can currently match the C++ experience you get with VS Code.

* Debugging is rough. There's experimental DAP support, but it isn't ready to be used. I was able to set breakpoints and step through the code, but the UI for exploring variables / state while paused felt missing or was unintuitive enough that I couldn't figure it out. I use CLion for debugging.

* Goto definition works with clangd as long as your CMake setup outputs compile_commands.json, which you already do.

* Renaming symbols (variables / functions) via clangd works fine.

* Intellisense is decent, but I had to tweak clangd settings. By default, it would stop returning results after scanning a certain number of symbols, so some valid functions just didn't show up. I was using Helix for a few days before realizing this problem, it isn't obvious that you are getting an artificially constrained filtered view of your symbols via default clangd. Maybe this is a distro packaging issue though?

* The order of intellisense completions is not great. CLion is smart about surfacing relevant suggestions first. In Helix + clangd, I often get obscure symbols that obviously have nothing to do with my project or context. It's not the worst thing, but it is mildly annoying and noticeable.

* "go to error" doesn't surface errors in files that aren't open. In Helix, space D brings up workspace diagnostics, but it only shows errors for compilation units already open. This appears to be a clangd issue, as space D in other languages will show all project errors. CLion does not suffer from this problem.

* I think you can get LLM suggestions via an LSP, but I've not tried personally. Assuming it's true that you can get LLM suggestions, it's not clear to me that you can run two LSP's on the same file, so it might be a choice between clangd and an LLM LSP? Not sure.

* No integrated build support. You'll probably end up building from a terminal. I use Wezterm with a custom lua script that is invoked on a hotkey. I put a lot of thought into the build UX, and what I've done is both extremely hacky and still not good enough.

Helix is not flexible, it's uncompromising. I like it, but I think it's hard to beat CLion or vscode for C++ development.

I'd say right now, if you have a good setup already, stick with what you have.

That's how you know LLM's aren't AGI.

Well, I meant how much is lost financially. Ah, unless you mean that people would watch less videos if they were subjected to ads, which is a great point I didn't consider. You're right, you can't just linearly extrapolate as I suggested due to that.

Well, if the recent drop in views was due to adblockers, we now have some data about what percent of viewers block ads. There would have to be an effort to collect this data, and the view discrepncy is probably going to differ by genre of video (eg, tech youtubers probably experienced a greater dip), but this should roughly tell us how much is lost to adblockers.

Creators have stated that while their viewcount is down their ad revenue is not - but a lower viewcount still presumably hurts youtubers for in video sponsorships, and if some genres of video have a higher portion of users with blockers, that probably hurts that entire genre in the algorithm. It sounds like viewcounts are returning back to normal though.