ericd

I mean, it's a good deal for the boss, but I also wouldn't want to waste two years of my life doing something I could do in a couple months or weeks, would you? That's not a good use of my life, regardless of whether someone's willing to pay me.

Yes, but I think with bafflingly huge numbers, it's a useful reference point. ~the value of all the stuff and services made in Australia for the year!

California rate payers are being forced to essentially pay for fire insurance for CA residents who live in tinderbox areas, by way of PG&E rates, and burying lines that serve widely dispersed populations. That's the big difference. The vast majority of US residents have much lower utility rates.

Well of course. Wars tend to destroy wealth quite efficiently.

Yeah, fair. We’ll see!

I don’t think they’re getting enough press on this to justify this level of investment. But I could definitely be wrong.

That's usually in hindsight? And I said it wasn't obviously insane to them, not that it would pencil out in the end.

Can also just be an asymmetric bet - relatively small outlay with low chance of working, but massive upside if it does, such that it's still positive expected value.

But mostly, I think all these armchair quarterbacks should not have very strong opinions about something that they almost certainly don't know much about compared to the people who build these. The person I responded to cited a random blog that ran some numbers, and includes such gems as "In space, a broken fan or a fried motherboard is a crisis." Just... lol

The threat of prison if you don't turn over SSH keys probably still works.

https://xkcd.com/538/

I guess we'll see if their idea warrants your ludicrous comparison soon. Google is putting out a lot of real resources to test theirs out, so presumably it wasn't obviously insane to them when they ran the numbers.

The real point is power limitations, putting stuff in space in the right orbit lets you turn solar power into reliable base load.

And here I was thinking that prisons should be about keeping people who can't operate in society without damaging society away from society. I mean, even better would be to find a way to help them operate in society, but apparently we're not even all at the baseline.

Excited to try this, I've wanted for so long to have a properly performant gaming VM without having to do all the VFIO nonsense, thanks very much.

Well, the guy is a professor in AI. So I'd not look for 100% human-written prose from him.

But I'd also not dismiss what Tim has to say. For much longer than this stuff has been super hyped, he's reliably been one of the best sources of info on some of this stuff, especially about GPUs and how to run things locally.

> No ones shown there's any real ROI in a more virulent intermediary that has even less reason to view ads.

Good riddance?

Very happy to have my agent totally ignore the ads trying to hijack my attention.

It seems that a pretty clear first step is that thinking traces are required, along with sourcing all evidence, so that things can be easily double checked. Anthropic/OpenAI have business reasons for not sharing those, but it also probably means they can't be trusted with any vital decision making.

That’s awesome! I especially like RECAP as a method for freeing things from PACER.

Seems like you all are already doing a lot of what I’ve been aiming for with mine. Are there useful ways to contribute, or have you all gotten it to a pretty good place technically, and it’s mostly a matter of spreading it at this point?

OK, I just bundled it into a torrent, available here: https://law.agentlookups.ai/downloads/

If you decide to give it a download, let me know if you have any issues with it/suggestions for improvement.

Yeah, going to the courts and municipal code seems like it's going to be a heavy lift. Many of them seem to hang off of municode, though, so maybe it's not a huge number of unique crawlers.

Cool, thanks for the pointer!

Right now it's just a bunch of crawlers for the individual states. If there's interest, I could periodically stand up snapshot torrents or something. That something you'd be interested in?

Alternatively, if someone else knows an all-in-one option that exists, I wouldn't mind retiring those crawlers...

Yeah, it should be freely available, you have to be able to know the rules you're supposed to obey in order to obey them well. I've been making a free API for US law search, you can point whatever model you want at it: https://law.agentlookups.ai/

Very much a work in progress, only federal and state so far, no municipal codes yet, and no case law yet. Big hole, I know. Also working on making the search ranking work better.

If it powers a heat pump water heater, it's pretty efficient, and much less plumbing involved.

Our house is all-electric. We have a 60 panel, 24 kw array that would probably cover all of our needs if it weren't shaded in the afternoon, and we have a real winter here, so the heat pump runs hard, and we use a substantial amount of electricity. About 1250 square feet of panels in total. Equipment costs about $20k-$30k in the US, not including batteries or electrician install. The actual array was pretty easy to set up.

On the global market, you can get the equipment for much less, the same inverter is a fraction of the cost that the UL listed American company version costs, the panels are a good bit less. But in the US, this stuff gets marked up pretty hard.

If your main load is for summer A/C, with a mild winter, the generation and usage line up much better.

Nice, have a github repo? This is something I've thought about trying to build, seems like the future.

This is my read, too.

Ah good catch on the total tokens, was going off vague memory there, and I thought people had gotten qwen 3.8 27b up to similar decode speeds as ds v4 flash.

>One data point that seems relevant to me is that the previous gen qwen Qwen3.6-27b was not so different in performance from its sibling model Qwen3.6-35b-a3b. We never got a qwen3.8-35b-a3b, but if we had, would the gap have stayed the same or gotten bigger? I.e. would the quality gains by improving training coming up against a hard limitation with 35b, or not.

Yeah good question, kind of shocking that a 3b active model would perform as well as a 27b dense.

Thanks! That's very helpful as a way to discuss.

First off, I'd include Qwen flash-next and GLM 5.3 to show some of the other strong open weight models, and they predictably dominate it, but they're much larger. But, it shows up right next to DSv4 Flash 0731 on the overall index, and that's much larger. It's a great model! But then scroll down and hit Time Per Task, and you'll see that DSv4 Flash takes 3.6 seconds per task to Qwen's 21.1. That's what I meant when I said this:

>speed due to excessive thinking maybe to make up for the smaller amount of world knowledge baked in (qwen 27b's main issue iirc), etc - they're tuned for different things.

It can make up for its shortcomings by iterating a lot longer, and using way more thinking tokens. And that's a great trade if you don't have the vram to run the bigger models, but speed is pretty important for getting things done... And that's why DSv4Flash is great, too, despite being much larger, and scoring similarly on the intelligence index.

Ah sorry, I should've continued, the bigger recent models are commensurately smarter. If you really want to make the point, then you'd need to show 27b being smarter than similar vintage bigger models. And in that case, there's confounding issues like efficiency, speed due to excessive thinking maybe to make up for the smaller amount of world knowledge baked in (qwen 27b's main issue iirc), etc - they're tuned for different things.

Gpt-oss—120b is like 1000 years old in AI years, whereas Qwen 3.8 27b is pretty young. What you’re seeing is that parameters aren’t apples to apples, and at a given parameter level, the new models are much, much better than the ones from a year or two ago. Like, to a comical degree.

Yeah, it's really tough, there's so much potential good, and so much potential bad. My intuition says that decentralized will probably be better, the other way seems to concentrate power while also not really keeping it out of the hands of anyone. It doesn't seem likely that the US firms will be able to convince the Chinese to agree to an international regulatory framework for this stuff. Opinion polls there have the populace being quite optimistic about the benefits of this stuff, and I think the cat's too far out of the bag to really put it back in, anyway.