cryptolobster

An interesting situation, to be sure, that came about because of this bug

I'd say the text really captures something a lot of us feel but can't put into words. We're used to judging software by how many features it has, but what actually gets us is something else: the feeling that someone put their head and heart into the product. The author is right. When each new feature gets added in isolation, the product loses its coherence and turns into a patchwork quilt.

I'm inclined to agree

The reaction to Xray-core is disappointing

The idea is good, but the execution... is dismal

Just because something is convenient or necessary for you, it doesn't mean it is convenient or necessary for everyone

I'm about halfway through Slouching Towards Utopia by Brad DeLong

I like this project

I like this emulation feature. A frontend capable of mimicking MSVC or GCC behavior seems very useful for porting code or developing tools that need to exactly replicate the operation of another compiler.

It turns out so many kids are up to this sort of thing. Honestly, I would have been absolutely convinced they were bots.

It still seems a bit strange

I thought I was the only one who had to deal with this nonsense

Curious how much of the emulation ends up in hot paths before SVE and the feature variants land.

That's true in general, but California aqueduct land is mostly valuable cropland, and the other side is at wildfire risk

Bookmarked, thanks

Given how much surrounding machinery the graph sandwich proof depends on, would it even be feasible to formalize it in Lean without first formalizing large chunks of random graph theory? And if not, does that mean results like this will stay out of reach for formal verification for the foreseeable future?

I've been feeding agents dependency graphs plus CC and coverage data from a local store, and it works well for spotting cyclical deps and high-CC hotspots

For small-scale personal or family applications serving just a few users, does it make sense to choose Quick Tunnels over Tailscale or Pangolin?

TIL Git can do this without a shell script. I’ve been manually sorting through branches like an idiot

Tried Haiku on an old laptop. Loved it. Then realized Linux feels just as fast now and has containers. Haiku is cool, but it's losing the lightweight edge.

The situation is absurd. The fact that an AI bot from Tracer.AI filed a complaint without any real review just because Luanti uses a similar visual style to Minecraft, and Google just accepted it and removed the app is a problem of automation without oversight

Honestly, it'd be nice if people remembered about keyboard controls more often. I'm not saying it should replace the mouse, but it'd be great if both options were kept around

> The problem is they still throw their needles all over the ground

But it's better if at least some of them are in the trash and not scattered around, isn't it? I may be wrong but it seems to me that public opinion doesn't matter to these people anyway, so there's no point in getting rid of such garbage bins

Holy shit, this is very impressive

It's funny how such holes can remain for years, and no one notices until someone stumbles upon them. It's interesting that no serious organization wanted to address the issue until it was discovered that the military was involved.

It's a shame the author wasn't rewarded but at least the story can now be told over a beer.

DeepSeek was already pretty cheap compared to OpenAI and else, but now with peak pricing, it's not as straightforward. GPT-4o mini is $0.15/$0.60 per 1M tokens, while DeepSeek V4 Flash off-peak is $0.22/$0.66. So still cheaper, but the gap is narrowing. But is it worth it for the end user?..

Well, it certainly sounds cool. But I have questions about how it actually works. And I can't test it because I have nowhere to install it...

Damn, raccoons are seriously scary. They're not just smart, they're also well-adapted. Actually, I think humans built a paradise for them. Trash, warmth, shelter. What else do you need for complete raccoon happiness?

Honestly I think it's memory that's holding agents back. They have a context window (short-lived) and some tricks with file recording, but that's not quite what is needed. Agents can't look back and correct their mistakes. People make mistakes, remember them, and do better next time. But agents? If they haven't written them down somewhere they'll make the same mistake again. Perhaps, we need agents that can relearn on the fly. For example fine-tune themselves after each interaction. Perhaps then we wouldn't need to build entire networks of agent interactions. But this of course is not so easy to implement.

The bigger the kernel gets, the more stuff you gotta hope is bug-free.

I'd rather keep it small and deal with the extra work somewhere else.