I've got a Framework with 128GB. Sure, I'm probably not going to run Deepseek V4 flash (though, I could run the 2-bit quant). But there are a lot of models (especially MOE models) that run fine. Even Qwen3.5-122B in a 4 or 5 bit quant can run. The Qwen3-coder-next (~80B IIRC) runs fine (IIRC I'm running a 6-bit quant of that one). The much vaunted Qwen3.6-27B is runnable, but kind of slow (~20t/s with MTP), though that's not a memory limitation issue.
Yeah, It's be great to have 256GB RAM, but that's really expensive now anyway and there's no way I could get my wife to sign off on spending 5 grand (or more now) on a box with that much memory.
In general I think from the coding side they're more robust now. However, people generating docs are maybe not as experienced with how to prompt in ways that avoid having the LLM tell you what you want to hear. I think this is still a pitfall that can easily be fallen into. Those of us who are doing LLM-assisted coding for the last couple of years are more aware of this now. Those who are planning/management folks are still kind of susceptible depending on how much experience they've had dealing with LLMs.
I'm getting so many requests to review LLM-generated documents - planning docs, docs intended for end-users, project docs, business plan docs. A team member sent me a zip file with about 30 LLM generated documents in it the other day and asked if I could review them right away. And a lot of it was just repetition and/or stuff that was just out of left field, made-up, hallucinated stuff. They're able to generate this stuff way faster than we can review. It used to be that it would take a significant part of a day for a project manager to come up with a planning doc - now they can generate one in a few minutes and send it out for review. It's just really tiring.
Qwen3-coder-next is very usable. But I don't think it's as good as Qwen3.6-27B (though it does run faster on my hardware). It would be great if we could get a Qwen3.7-coder, but I'm not going to hold my breath.
At this point I wouldn't consider Romney "rightwing", more of a centrist by current standards. Heck, the president probably thinks Romney is a socialist.
To both of your points: the beer hall demagogue wouldn't have gotten to Chancellor if the German elites hadn't decided that he really couldn't do that much damage and we may as well let him be chancellor to quiet down his followers. Even after the putsch, he got a very light sentence because the judge was sympathetic with his right-wing cause. You're both right to some extent. A huge amount of damage was done by one man, but he got to where he did because the German elites thought that he might be useful to their cause.
I read Sinclair Lewis' Babbit last year and it was kind of depressing how little has changed since 1922. The political climate (at least as portrayed in the novel) seemed eerily similar to now. Maybe we continually go through oscillations.
I'm trying to figure out why I should have any interest at all in participating in this? I mean, other than the $1000, but there's really no clear criteria as to what would win that $1000 since AI is judging it.
I've seen this assertion a few times now in the thread here, can you elaborate? The events of 2008 were kind of extraordinary, so I'm not sure we should shape all of our policy based on that - except in terms of trying to avoid another 2008/Great Recession, but I don't see these actions doing that.
> I guess preferential treatment could be the only issue?
I'm not sure it's the only issue, but for now focusing only on this issue: it's a really big issue. The R's often complained about Dems choosing winners and losers (IIRC related to the solar industry). This is now way beyond what the Dems were trying to do by advancing solar. The gov can use it's buying power to sway things towards Intel, for example, over AMD. Huge potential for conflict of interest that will distort the markets.
The other tangential issue is that in some cases these can look like bribes. For example, OpenAI "offering" 5% of their stock to the gov.
I'm old enough to remember that around last Fall people were declaring Google AI dead and then they came out with Gemini 3.1 Pro and there was a bunch of discussion about how Google was back and was going to end up ultimately beating the rest because they have their own TPUs, vast amounts of data, etc.
My guess is that Google will (finally) release Gemini 3.5 Pro and it will be pretty decent compared to OpenAI and Anthropic and we'll see discussions again about how Google is "back". Wash, rinse, repeat.