montroser

Sorry, but they can think and reason.

They can have empathy too. Meaning, you can describe a circumstance with many parties involved, with different motivations and conflicting goals -- and it will tell you how these humans would likely feel about each other and the situation at hand.

Well, it will be interesting to see where Qwen3.8-Flash-Next ends up landing, also released today. These are exciting times!

Hoping this is real. It's too bad to see the signals from Qwen that they will not be releasing a 35B-A3B for the 3.8 lineup. The MoE architecture makes a huge difference for being able to run these local models on reasonable consumer hardware.

You can do so many different things with a computer. If you spend hours watching AI generated YouTube shorts instead of playing outside with your kids, then yeah, that's too bad.

But if you spend all those hours building a retro/modern sim city game pair programming with your 12yo, designing the artwork, crafting the story arc, writing the music, finding and fixing bugs, building new levels -- you don't need to regret that you weren't idling around at the park all that time.

I am pro AI, and also deeply share this sentiment.

I think sometimes people have a big long thread with Claude where they feel enthusiastic about the back and forth being very productive -- and then they genuinely want to share the summary of the conversation with their colleagues so that they can feel it too. It's not just typing a short prompt and copy/pasting the response.

But, Claude, especially Opus is so unnecessarily flowery, and smug -- and it's so identifiably characteristic.

It somehow rubs salt in the wound: "I didn't take the time to write this out in my own words, so now you get to listen to this petulant asshole mansplain it to you".

If it would get to the point, and with some humility, it would be a much different proposition.

Wow, here we are! I do not want this in any way. Soon enough, we will not have a choice.

Or at least, we will have to pay dearly for devices that afford us the luxury of not participating in mass corpo-state surveillance...

This would be extremely difficult to detect and enforce.

I think their heading `box-decoration-mode` is actually meant to be `box-decoration-break`. That's the real property that exists, and is documented on MDN.

https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/P...

> This is intended to handle the increase of calls that are related to one incident, so callers get automatically routed to an AI agent who asks if they are calling regarding the incident. If the answer is yes, then callers can receive information or updates, and if it's no, then the callers are transferred to a human.

This is where it starts. Soon enough we'll have robot EMTs, and then robot police after that. Only for the poors, of course...

It's this here: https://www.nordstjernen.org/northstar-browser/

> You must not store or reuse URI structures in the client. An API could change them without notice, and your client would obviously stop working. Instead, parse the links and follow the relative identifiers (rel).

This part is a complete and total fairy tale. No API consumer wants to operate in a world where the API could "change without notice." In real life, no provider would expect that of their clients, and no client would seriously accept that before every API request, they must first hierarchically rediscover the provider's dynamic url scheme of the moment, in order to traverse down to resource they ultimately want to request.

There has never been any serious commercial or high-traffic API that implemented this. And it's not because "lame corporate tech teams aren't hip to the true meaning of REST" -- it's because "it doesn't actually make sense to do it that way so nobody wants it to work like that."

I want a beautiful eink clock that gets its time from gps/gnss. It seems too much to have to collect a clock to wifi. Just turn it on and it should work.

Well, the whole thing is theater anyway. It does not matter what you choose.

They will fingerprint you with or without cookies. They may or not try to honor your preferences, but their "partners" will not try, and by the time you see that banner, it's all out there.

"Accept" is the close button.

Well, many EU countries like Italy and Germany officially freaked out about DeepSeek, ordering it to be banned from app stores etc.

This is why God gave us 1.58-bit ternary quants?

This was always where this was heading, but we got here much faster than expected.

Once western governments declare it to be a "national security" risk for citizens to have access to open-weight frontier models, and once they classify using these models as acts of terrorism, what will that world be like?

Will using Kimi K3 come to be like how napster was in the olden days? Everybody knew it was technically illegal, but come on -- any track at your fingertips? But surveillance is quite more evolved now.

Or it will be like cannabis, where a guy in the neighborhood will low key rent you metered access to the 8x5090 rig in his basement he cobbled together from parts on ebay? Or everyone will flock to VPNs?

Or will the oppressors actually succeed? The same way that napster is long gone, and everyone accepts that they must pay spotify for a homogenized collection, where artists must take only a minuscule cut (more than napster though)... We'll be stuck with nerfed Cohere or Mistral models for open-weight options, as if they need more lobotomizing. Or else we can pay through the nose for Anthropic/OpenAI for "American Frontier" models which will fall increasingly far behind China.

Or else, like how Kindle Fire was subsidized by ads, we'll have "Kindle AI" where influence is sold to the highest bidder, where the LLM will tell us that smoking is actually healthy if big tobacco can engineer its renaissance by turning its lobbying dollars to pay-to-play, pumping its propaganda into the training pipeline for Amazon's extra commercialized line of ultra budget LLMs.

One time I hired two different random guys on craigslist to assemble three IKEA bureaus. They didn't know each other beforehand, and I hired them separately to work together.

They started by jockeying to see which one was going to be in charge, arguing about how to proceed, throwing shade, making claims of superiority. One guy wanted to follow the instructions, and the other guy just wanted to go on intuition. They couldn't agree and asked me which one should be in charge, and I said, I don't know just work together, follow the instructions, and get this stuff built.

The instructions guy assumed the lead, and then they settled into a rhythm. Three hours later, they couldn't stop gushing about each other, and a most beautiful bromance was born. They exchanged telephone numbers and made plans to hang out on the weekend.

This is very cool! Slightly off-topic though, I miss technical people writing in their own voice about the awesome things they've built.

Welcome to the future. I think open weight models are our only hope for LLMs being net positive for society.

Well, it's not that hard: just give the LLM a user-scoped access token, same as if the user themselves were asking their own LLM to act on their behalf.

Basically, just like we don't show users information they shouldn't be able to see, and don't let them take actions they shouldn't be able to take -- we can use exactly those same explicit mechanisms (scopes, roles, permissions) to limit what the LLM can see and do.

The LLM could try to do more than what's allowed, but they get shot down with an access denied message just like anyone else.

The anti pattern is to think that you can reimplement access control with prompt engineering and give the LLM root access. That is doomed to fail every time.

I'm ready for my home helper robot that makes dinner and does the dishes and takes out the trash.

But I'm scared for when those home helpers get drafted to fight in wars, either for or against me...

This article only promises to get into "the coming AI margin collapse" in a yet to be published "part two". This part only makes the point that GLM 5.2 is pretty good (no shit).

I'm glad that your trivial migration was trivial.

All this was a big source of churn for plugins and frameworks: https://github.com/vitejs/vite/discussions/16358

Vite had five major version in the four years 2022-2026. Version 3 => 4 => 5 => 6 => 7 => 8. Each one of those had breaking changes and required devs to go through a migration. It's too much. And for what? It's not as if it is dramatically better now than it was in version 3.

I can't say I would really look forward to bringing this level of needless churn and constant disruption to the rest of my development toolchain. Anyway, Vite+ is really just wrapping existing tools into an abstracted command-line interface? And so I have more layers of indirection to wade through in order to get the thing to do what I want? So far I am not optimistic about this prospect...

That "thank you" at the end is particularly classy. Thank you for getting fucked and giving us your money.

Yeah, but they're talking about fine-tunes.

Sounds cool. How do agents know what else is going on in the doc? They have an embedded browser and they do like mutation observer type stuff? Or does the integration do polling?

Nice. Can we get `nub --compile` up in there like Bun has?

I've been driving glm-5.2 for a day or two now. It feels like a mature, seasoned colleague.

It could be luck, but I don't know -- it keeps one-shotting relatively hard stuff. And taking initiative to think about what potential regressions it should look out for, and choosing to do strategic refactoring when it should do. It is not confidently incorrect hardly at all, doesn't tell me that it's fresh risky pile of changes is ready for production without having exercised all the code paths and writing a bunch of tests, etc.

We might be reaching the next level here...

Reading the paper, by "Fertility" the author really means "rate of pregnancy" which seems kinda like a different thing.

Basically, access to the Internet is correlated with less teenage pregnancy. Idiocracy in motion...