Any quantization beyond Q4_0 goes mad (losing context, making category errors, malforming markup tags, mangling and inventing words) long before the context reaches a size I would consider "useful". I've tried running 31B at IQ3_XS and it looks alright from a distance, for a few prompts at a time. Unfortunately the errors compound and semantic drift goes into runaway. I have no interest in playing with quantizations any smaller than that. The only thing "massively wrong" is the suitability of a 2-bit quantization for my actual use. I've investigated it, and it can be an amusing toy, but it is far from usable for any meaningful work.

Even Q4_K_M was too error-prone for meaningful use, and I was using Q5_K_M as my baseline of "usable for a while". Fortunately, most of the heavily quantized models are based on QAT releases now, and a 4-bit QAT model is pretty much on par with a 5-bit non-QAT model.

DeepMind has released Gemma 4, which doesn't include their largest models, but 31B is a creditable model. I wish I could say the same about the 26B-A4B model, which is just smart enough to be entertaining for a few days or weeks, then everything it does starts to sound the same. Unfortunately, 31B on my hardware (i5-8500, 48 GB DDR4, 12 GB RTX 3060) runs at 0.3 to 1.2 t/s. So while I'm not a big fan of Alphabet's business practices at large, they aren't regressing. Switching to the MIT license basically means they're abdicating all control over derivatives.

The only thing Grok ever really did for open weight AI was show up for the party a couple times. At first, this helped establish a baseline that could never be retracted, but it hasn't proven to be particularly important. Everything since has been far off the bleeding edge, but they collect their participation trophies. I think their subsequent actions have gone a long way toward demonstrating their purposes, which are wholly selfish. They'll do as little as they can to contribute while retaining the benefits of being perceived as open and competitive.

Large electricity users drive up the costs for everyone else. They should be made to pay for those extra costs; similar to how a lot of water districts charge heavy users more per gallon, electric utilities should charge heavy electric users more per kWh.

Stupid tends to drag the rest of us along for the ride, even when it doesn't intend to.

The average (median, whatever) IQ would increase above 100... until the test was re-scaled to account for current demographics.

This exists in the Mac world as well, but the primary reason I'm tied to Windows for my DAW is because I can't register my instrument because Native Instruments never bothered making a registration app for Linux.

Not that MS Project Server is necessarily the greatest tool in existence, but what is the Linux equivalent? And I do mean the Server/Cloud editions, i.e. supporting multiple projects across shared resource teams, with ticketing and websites per project, etc.?

I don't know what expectations these are defying unless they're from those created by anti-EV FUD. I thought it was pretty clear that EV batteries usually last longer than the cars themselves. If 250K is exceeding expectations, then the expectations are wrong and haven't been supported by the data for a long time.

Depreciation is high, meaning you're mostly paying for a name—which people probably aren't going to do anymore. I was just looking at the market for used Polestar 2s last week: a 2024 AWD model with 48k miles (so someone leased it for three years and put every mile allowed onto it) can be had for $29k. This was almost identical to the pricing of a used Hyundai Ioniq 5 with AWD, and the Ioniq started about $5k cheaper.

OTOH I don't hate everyone who's committed a crime, and I don't think the purpose of our system should be "punish the wicked" or exact retribution or anything, either. If we think a person can be returned to life on the outside without undue risk to the public, I'm all for that. but if OTOH we can identify that someone is likely to recidivize(?), then maybe it's best for everyone if they stay locked up. Meanwhile, I don't think we can look at the penal system in isolation to determine that - see e.g. comments elsewhere on this page about the quality and capability of the society at large to provide an environment that lessens the likelihood of recidivism.

You seem to be assuming that the difference between Nordic countries and the US is only about the penal system itself, as opposed to e.g. more social democracy i.e. a stronger social safety net for people once they are back out in society, which might them less likely to resort back to crime?

"In other words, they haven't fixed the problem of recidivism, they've just kept people in jail longer."

If they're keeping the habitual criminals in prison, and it means fewer crimes committed against decent law-abiding folk, then what's the problem? Isn't the purpose of our laws and penal system to protect the innocent?

Case in point: Long Island NY residents and activists got the Shoreham Nuclear Power Plant shut down before it ever even went fully on line. Now there are wind turbines, solar panels, and battery storage located there, as well as a connection point for imported power from New England.

That parts gonna happen anyway.

Yeah, that's not really a thing in LEO where debris clears itself fairly quickly due to atmospheric drag.