In the meantime, highly encourage folks to checkout grim dawn. Still gets updates even though it released years ago. Game worthy of buying to support these types of devs.

Awesome thanks! I will check this out :)

Man hopefully they just mercilessly rake companies that don’t support Linux because of kernel level anticheat over the fires as well. We need more advocacy in this space I think (and honestly I’d like it if kernel level anticheat was banned from steam on account of security).

TLDR:

Qwen model is popular and DeepSeek R1 seems good, and they’re based in China. There is investment because they think that’s where the future commerce will be (which to me is a little like the dot com bubble but, whatever). Lack of chips can be problematic.

Is there any issue for current owners? I was going to get one as a gift for a family member but if they’ll somehow lose vital support for it I suppose I’ll look for an alternative.

I’ll check this out, thanks! I really just need to figure out how to build in the driver level stuff for my chipset. Even this I think just pulls from lm-sensors which needs the low level drivers to populate the appropriate files to read from.

HWMonitor / cpuID / cpuz. One of the frustrating things is not having good driver level support for certain mbs with system monitoring utilities, so you can’t see fans and some cpu stats (like per ccd temps etc on Ryzen processors). Specifically things like it87 boards

This is cool, thanks for sharing

Thanks! Looks like they don’t specify any fine amounts just saying that it’s probably coming and could be leveled before leadership change in the fining body in EU.

Looks paywalled or something, anyone can provide a tldr?

Ah interesting — again happy to help out if there’s anything I can contribute to. I can make a feature request on github if there’s interest.

Anybody know of a program that would let me remap the keys on the razr naga? I’ve been using this to deal with the rgb, but would really like to edit the mouse button mappings (not averse to running a virtual machine to do it, if it’s one off, or maybe in bottles if that works).

Is there any interest in getting local models to run using this? I’d rather not use Gemini, and then all the data can reside locally (and not require a login).

I’d be happy to work on this, though I’m a python developer not a typescript one.

I personally love PWAs — why the hate for them? Personally I think more apps should be PWAs instead.

Already voted — hope everyone gets out there. Would be pretty awesome to have Allred win our senate seat; super long shot, but a Texan can dream.

You just equated doing things that you have at least semi active control over to someone’s genetics predisposing them to certain medical conditions, which they have 0 control over. In their markets, risk is supposed to balance out and make people make less risky choices. You can’t derisk your alleles.

Health insurance is a fundamentally flawed idea, and not because of preexisting conditions, but because of profiteering. We should just optimize the health of our citizens directly by taxing wealthy individuals and companies and paying for the most effective healthcare for everyone. It’s more cost effective for society at large and also serves the greatest cross section of our community, but there just won’t be a profit motive (well there is a motive that by doing better healthcare for everyone in more cost effective ways you lower the overall cost of healthcare for the society, which isn’t so much maximizing profit as it is minimizing cost-benefit ratio).

And by the way, it may be a paper next year finds an allele you have increases your risk of some horrendous disease. The people in this thread are arguing with you that you should still be able to afford healthcare. You’re arguing you shouldn’t.

N-95 masks are protective, and to a certain (most likely lesser degree), KN-95 masks are also protective.

I don’t believe this is quite right. They’re capable of following instructions that aren’t in their data but appear like things which were (that is, it can probabilistically interpolate between what it has seen in training and what you prompted it with — this is why prompting can be so important). Chain of thought is essentially automated prompt engineering; if it’s seen a similar process (eg from an online help forum or study materials) it can emulate that process with different keywords and phrases. The models themselves however are not able to perform a is to b therefore b is to a, arguably the cornerstone of symbolic reasoning. This is in part because it has no state model or true grounding, only probabilities you could observe a token given some context. So even with chain of thought, it is not reasoning, it’s just doing very fancy interpolation of the words and phrases used in the initial prompt to generate a prompt that is probably going to give a better answer, not because of reasoning, but because of a stochastic process.

Bookmarked and will come back to this. One thing that may be if interest to add is for AMD cards with 20gb of ram. I’d suppose that it would be Qwen 2.5 34B with maybe less strict quant or something.

Also, it may be interesting to look at the AllenAI molmo related models. I’m kind of planning to do this myself but haven’t had time as yet.

So glad I have Tesla shorts lol