TheRealPomax

Now we just need everyone else who works for X to do the same.

This is an odd comment: the project is right there for you to use, so just try it and see if it holds up to the claims? Then you can comment about the fact that it either doesn't hold up, with numbers to back that up, or on how awesome it is because it works =)

They... didn't? They got a petition response from the EU that was more positive than they had expected, and a response from the commission is not court: the response phrasing literally leaves the door open.

Heck, the parallel movement in the US even had a major legislative victory in California, so... what are we talking about?

A postmortem is for reflecting on something that went so wrong you had to kill it, and you want to objectively describe what led to that, and what lessons can be learnt from the failure.

A retrospective is just the reflecting part, without implying you believe it is going to fail and you're looking for some schadenfreude ;)

Either you mean retrospective, or you're being unnecessarily mean for something that's been getting quite a lot update reports, just on a social medium you probably don't use =P

I can't see any specific "and then AI happened" in this graph? SO made a lot of dumb choices that made it less and less useful for people independent of forcing AI in there, nothing in this graph is a clear indicator that AI was even a blip on the rapid decline of a service intent on making itself as poor an experience as possible for folks wanting to help other folks.

> Input. A sorted list of 32bit unsigned integers vals: Vec<u32>.

Okay, but that's not even remotely like the kind of input that this tree was created for. From the next paragraph, this work is in part

> [...] to make efficient datastructures to index DNA [...]. One such datastructure is the suffix array, that sorts the suffixes of the input string. Classically, one can then find the locations where a string occurs by binary searching the suffix array.

So where is the analysis of how it performs for that use-case? Searching through "already sorted 32 bit numbers" has nothing to do with searching a 3 billion character string (that by definition cannot be internally sorted) for substrings.

If you think those two contexts are equally secure, your machine is incredibly insecure and I hope you run daily incremental full system backups.

You have far too much data in unsecured locations, and you have far too little understanding of what an agent would do, to go "I trust whatever this user account will be doing on my machine".

> On your own machine that leaves two bad options. You approve every command (and babysit a prompt every few seconds), or you run --dangerously-skip-permissions and hope nothing important is one rm -rf or one leaked token away.

Literally everyone has the option to use a VM - it's built into Windows, UTM on MacOS, Docker on Linux. Yes, "a tool that automatically builds a VM" is useful, but we've had a third option (four, if you count "actually I disagree with the idea that it's only useful if it's fully agentic") from day one.

Note that they must be sold with this feature. It does not say you have to keep this feature, you can aftermarket remove it or disconnect it just fine.

Anthropic can't be bothered to actually fix any VS Code extension bugs that people file. They could just let Claude fix them, but instead they let Claude close bugs as stale so they don't have to pay any attention to their users. Who needs users? I guess the retort to that in this case is "who needs Anthropic? Extensions are just web apps, Claude can fix those just fine". So I told Claude to fix the things Anthropic can't be bothered to, and now there's this repo that Claude can use to remove the "whimsy" nonsense that Anthropic refuses to remove, and add in proper syntax highlighting to code blocks, which is literally built into Monaco but Anthropic couldn't be bothered to hook into.

It's nice that you have that luxury, but that makes you an anecdote in a world where folks need a smartphone just to access banking or government services.

> Sometimes I won on points and lost the person. More often I won nothing at all: I’d watch someone grow more certain of the very thing I had just disproven, while the room quietly drifted to their side. I would walk away technically right and completely alone.

This is why debating is taught in school in the Netherlands (and I'm sure other countries, too). Winning an argument is not the same as convincing someone they were wrong: that's something you need to learn how to do and then something you need to actively practice with others.

Just having good arguments makes you a dick. Having good arguments and being able to empathize with the opposition's and conceding their position on any merit, while showing there's a solution that'd they'd prefer even if they don't know that yet, makes you someone helpful and trustworthy.

Not if you had a non-US laptop. Almost every laptop keyboard except for US models have an AltGr on the right.

That's actually the one thing that does make sense: police has always wanted to be able to do this, but they legally can't. But they can reward a private company willing to do it for them, so that they can "ask for the data" without ever breaking the law.

But not so much between "this person walked in my door and stole my stuff" and "I left the door to my house open and then I put my stuff in the doorway."

Flock cameras are roughly that secure.

No, the shortcut was alt+s. That's what people typed. Then on Windows, which used alt-combinations already, it became rightalt+s (as the rightalt wasn't used by Windows itself) but instead of having a dedicated rightalt code, Windows would rewrite that key into a ctrl+alt code combination.

If you're going to tl;dr, at least get the most important detail right. People only ever pressed alt, and Windows went "and now you're pressing ctrl+alt", so that alt+s becomes ctrl+s with an alt that no one's looking for when it comes to intercepting and killing off key events.

And every time you press it, an entire VM gets spun up, fully provisioned, and then set to LLM processing mode even though all you'll be doing is immediately closing the app again.

Thanks Microsoft, stellar!

You forgot the "still" in that sentence.

Funny, that was already true in 2006. It's the thing people keep saying, and yet anime keeps coming out.

Right now they're stuck in the whole "ten shows from the same budget, each run by their own 'committee' and each competing for cash that runs out well before the show's over, and the poor performance very quickly get less per episode". Great for networks, shit for shows. Even worse for animators who need to get paid a living wage.

I just wish someone would look at the input side, too. I want true digital paper that I can draw on in real time already. Not the laggy nonsense that even stupidly overpowered tablets can't seem to get passed.

This is a 2026 update of a 2011 article about writing a tiny zero-width font for doing webfont load detection for drawing text on a canvas element, as the browser landscape, as well as font support, has changed radically in the last fifteen years. IE no longer exists, Opera is just another Chrome, Safari fixed most of the bugs that broke this approach some fifteen years ago, and pretty much everything's using OTS for font sanitization and Harfbuzz for shaping.

In the article we'll run through the byte layout of an OpenType 1.9 font, using TTF vector tables for our glyph outlines (which will contain exactly zero outlines!) and figure out which values we can just rip out, and which ones we can at least null out to make a font that's as small and as compressible as possible, leading to a 520 byte (100% spec compliant!) font file with a Base 64 string that we can pack in 298 bytes of JS for embedding directly on a web page.

"Why would we still need this in 2026??" To be clear: you probably don't. But that's not the fun part. The fun part is to see how far off the beaten path we can get in terms of creating a spec-compliant font that goes about as hard as possible in the opposite direction of what the spec designers had in mind.

IT doesn't even need to be open source, a walled garden that you can afford is perfectly fine. Someone's going to find the cracks in the wall anyway.

But a walled garden that costs $400 for personal use (we're ignoring yearly licensing, because f that noise) is utter nonsense, and the clearest sign you have no idea how to sell and then upsell products to users over the course of several years.

No, it is not. Etymology and current use are not the same thing.

That conveniently ignores the part where it refers to a fracturing where every part is hostile towards each other and unwilling to cooperate on anything. It's not a neutral term, it really is a rather offensive term for everyone from the Balkan today.

I think you'll find XeTeX (with XeLaTeX because no sane person uses plain TeX except for package authors of course ;) predates LuaTeX by almost a decade.

(Something I ran into in 2008 when I was writing a multilingual book and heard about LuaTeX, only to find out it was still extremely incomplete and was roadmapped to not be done until 2012. And I couldn't wait 4 years)

This page really does not like playing nice with reader mode, making it near impossible to read unfortunately.

A dollar is a dollar, and suppliers chose to screw over the world for not-even-profits: they would have sold the exact same number of chips if they changed how they prioritized delivery while making sure everyone got chips instead of just a handful of industries.

So, no: it's nothing like what you replied with and that was a rather dumb thing to say.

"Fueled by greed". It would be trivial to say no to AI companies because dollars are dollars, it doesn't matter who pays them, and prioritizing literally all of humanity instead of "five companies" is a choice that every single supplier could make, but decided not to. This problem was 100% manufactured by suppliers.

Some firms don't understand how to do data management, and if we draw the venn diagram of those and the ones that ban sqlite, it'd be pretty close to a circle.

Yes, databases could have any extension. No sane dev team would accept code that doesn't use an object extension for a sqlite database.

Yes, databases can contain PII but no sane product manager would go "yes, that's a good use of sqlite".

Yes, you can trivially copy database files, but no sane product needs to in the same way that no sane product should require folks to just clone the db just to do some work.

Pretty much every reason a company has for banning sqlite is a red flag for working there.