atonse

Omarchy seems to be doing a really good job with just that. But like you said, Linux.

I don’t fully agree with this sentiment. Just because we were told something and it didn’t come true before, it doesn’t mean it can’t come true now, and that the capabilities are there now.

At this point we have enough of real evidence from project glasswing like the massive Firefox security patches from Mythos findings. This isn’t crying wolf.

I’m very glad that they’re actually being grownups and not yolo’ing something this important, and are working with groups until we can secure critical infrastructure before making this more available.

I don't know, I feel like for a few weeks before the SpaceX datacenter, I was just constantly checking my weekly limits. And now after that miraculously, I rarely even come close to hitting my weekly limits. and I still have 5-7 claudes open a day (defaulting to Opus 4.8 xhigh, sometimes ultracode).

So I feel that the additional datacenter caused them to just ease up a bit. But demand is also insane, so who knows...

Yep!!! It really has changed and has become exhausting. The most flippant, negative, cynical, and antagonistic replies nowadays, rather than genuine curiosity about the news or optimism.

Too bad because HN has been my “home” on the internet for 15 years.

Sources for your numbers? I'm curious to learn more about how widespread this issue is. I personally have heard of 2-3 cases in my small social circle, so to me it feels more common. But I am having trouble finding actual numbers.

Don't agree that any of these cases are "most obvious" given that it's gone all the way through various appeals courts to the supreme court - and that's the mission of the Supreme Court, to interpret all the various situations for these cases and how they apply constitutionally.

Yeah I considered whether I should use the term "illegal" in my original post, but in this case, I believe these models are actually banned for use in China, right? Like there are probably export controls (at least with the NVidia chips)

I honestly don't know ... yeah if it's just technically a terms of use violation (which isn't illegal, just a violation of one company's rules, for which Anthropic has every right to stop), or do we now have export controls applied from the various government actions, etc making them truly illegal now.

I can't imagine any world where we put this AI stuff back in the box. It is simply too useful and too powerful. And as we start seeing all his upheaval where models are getting banned, etc, I can even see the appeal of on-device AI increasing for a lot of use cases.

So I think Apple has the right instinct. In fact, I've had the thought multiple times that I really want a lot of workflows just running on my device. Workflows like fast vector search (already fast on the m4, but I want it more common place), or realtime transcription and summarization to be even faster, on device, etc.

Oh not at all. To be totally clear what we do is build websites for research centers, help write articles about their research and run their newsletters. We aren’t claiming (or trying) to do any research or take any research funds.

But occasionally those research centers have needed help on the cloud side setting up servers to do their research. So we’ve supported them there, but again mainly in an IT/programming capacity.

We aren’t competing for grants. We leave that to the scientists. :-)

Our company is KZN Consulting (kznconsulting.com) and our product is SimplyScholar.com

We have the opposite problem. We've tried this slack integration (first with NanoClaw, then Hermes, but now just building our own), and we actually want the OPPOSITE.

In private context, I want to have a per-person conversation with durable context for that person's private chats. I also want that person's permissions to extend. Like contractors in our slack should only be able to ask and get back information about clients that they've been attached to, not our entire knowledgeable.

And we've implemented ALL that, but just with a lot of custom code. We've put in interceptors that put in per-user keys into the MCP connection so only certain tools are even exposed, etc.

We have worked with many (NSF funded) research centers over the past decade. It (accidentally) became our speciality. If you know research centers that are looking for software dev firms that understand how to work with NSF funded grants and research centers, I’d love to connect.

That made me wonder, since bun natively supports typescript, is it already using some kind of rust implementation of typescript?

yeah jq was a good analogy. I tried so many times to like jq, but found the cli and syntax too complicated to use. But I only finally saw and appreciated how powerful it was when my agent started using it.

Same with ffmpeg. I recently built an entire video out of 3 other source videos, just talking to claude, and it did everything (including audio adjustments, EVERYTHING) with ffmpeg. It was unbelievably cool.

It's not just about better diagnostic outcomes. Currently MRIs are a horrible, claustrophobic experience, and MRI machines are so insanely expensive that it is a bigger deal just to prescribe one.

So even if it is only as good as an MRI, or even 80% as good as an MRI, if it is much cheaper and much more pleasant to go through, you will get MORE people doing it, and get it prescribed in more situations.

That's at least how I read the benefits, democratization of imaging techniques rather than just improvement.

To be fair, almost zero people praise the dx for ffmpeg. but the utility and value is so massively high, that it overcomes the famously complex dx of ffmpeg. I'm not even insulting ffmpeg, if it does a million things, then there are going to be a million knobs.

I think of git as the same. The git cli is not intuitive at all (unless that lightbulb goes off) but the utility is so good, that people just kind of suck it up and use it.

I've built an entire app (and Swift UI companion app) to run my consulting business. It's my chief of staff that I use daily. It checks my emails, keeps an eye on client relationships, finds opportunities, helps me qualify them, etc etc.

While I am using Hermes now (and was using nano claw before), they really aren't doing much and so I'm considering folding the little functionality they do into the app, and I think it'll work better.

Part of the app is also a personal agent that checks my schedule and inserts in things into my calendar like "20 min neighborhood walk between meetings" or "break for lunch" between meetings.

I also built an entire document creation system that I use to build analytics reports or proposals, that generates very polished proposals with data about the company with me having to do very little work. All the numbers add up, the cover pages are beautifully designed and full bleed, and the documents always have perfect formatting.

Mythos useful for infiltrating networks of China and Iran?

Anyone who’s read anything about Mythos would’ve said the same thing on day one. Didn’t need some anonymous vague source to tell you that.

Or Astro? or Nuxt.js?

That's why I use NextJS... it seems to be the clear winner for React (at the moment at least).

Tanstack Start seems to be gaining steam.

At least many are in the react family though, so switching is less of a pain than Ember to React, for example. And especially so with agents now.

I think Apple Music (the web version) runs on ember.

I thought the main LinkedIn website runs on ember? I can't tell anymore cuz so much of it is obfuscated.

You’re so right. I had the same experience early on finding elixir developers so I just thought, I don’t care anymore, I’m just going to try to hire smart people.

And so the 2-3 devs we had working on elixir learned on the job just fine.

But it’s sometimes also the flip, where some candidates don’t really want to learn a tech that isn’t attractive on a resume. But those are probably the wrong candidates anyway (only saying this with experience)

I loved ember for years. Even attended EmberConf in portland around 2015. In fact, I'm just weeks away from retiring my last ember codebase that was in production and worked great for a decade (but I haven't updated in 5 years, since I was unable to keep up with all the changes). But after years of the community dying the most slow and boring of deaths, and having an absolute nightmare needing to hire Ember consultants, I really soured on it.

It is the main reason I completely stick to the boring mainstream (like react) now, so I'm never again stuck between "nobody knows Ember" and "this one consultant is charging $28k a month cuz I'm competing with LinkedIn, Netflix, and Apple" and then am stuck with them implementing engines for fun and then I don't have the time to undo it months later - all left me wanting to flee.

Basically, left it for non-technical reasons, just practical "literally nobody except billion dollar companies use this, I've painted myself into a corner" reasons.

But I do have fond memories of building things with it, personally.

No I actually do think this is AI generated. I came here to say the same.

Brokovich might not know it. But her web people certainly used AI to build this site. From the Emojis, cards, to the single colored left border.

But aren't they able to do incremental builds and separated x64/arm64?

They know which OS version is requesting an update, at least the version number part.

Yeah I’m honestly not sure why macOS updates seem to be so huge. Often gigabytes. Do they actually have thousands of changes, so they basically ship out new versions of almost all system libraries? Or is it that they don’t have good diffing in place? Or is it a BSD thing where you basically ship everyone at once since it’s all sort of “one version” of the base system?

My argument wasn't that the test suite was perfect, my argument was that this is far from "YOLO" - this is a textbook example of being able to do bigger refactors, etc when you have an extensive test suite.

You've never changed your mind once you got new information?

You can go read all the details on Jarred's X account - including the progress, how it was thought out, strategy, that they're aware that it looks like zig still, etc etc etc.

Speaking of environment though, everyone neglects to mention that the Bun core team now has access to Claude Mythos. You think they haven't already run Mythos against this? So they have private access to the best cybersecurity scanner known to man.

Suffice to say, I'm yet to see anything that really worries me in any major way with this.

Mostly unfounded in my opinion. They took the _exact_ same strategy that go did when they moved from C Go to a Go based compiler.

First they essentially wrote a translator that preserved the C idioms (so it wasn't idiomatic go) until they had byte for byte output.

Then they started changing code one by one to be more Go-based.

Bun is doing the same. Right now it is mostly a one for one translation of the zig code. Over time they'll make it more idiomatic Rust.

Yep, it's one thing if there was some project that saw severe regressions in Bun.rs and actually showed data about regressions.

But it's been available for a week. And so far, seems like crickets on actual data on any regressions. It's more "I just don't like this!" style grumbling.

YOLO? Bun has an extensive test suite and this implementation passed the test suite.

Can we at least try to be a bit more accurate and less hyperbolic?

I will continue to use Bun because the same people that made bun have made this decision. I trusted them one week ago. I have used bun for the past 2 years, and so have many others.

I'm not about to just assume they've become immature idiots yolo'ing stuff overnight. They're still the same people they were a week ago. Or two weeks ago.