fluoridation

6 ms·

1.
by fluoridation 5d ago
I'm pretty sure the company that owns the compute is the one paying for the power. Who would agree to have an ugly noisy cube in their property that they have to pay maintenance for? What would they get out of that deal?
2.
by fluoridation 5d ago
It didn't work. DeepSeek failed to load on LM Studio due to a tensor format error. I don't have the time to do deeper troubleshooting. Qwen 3.8 Next did work, but the performance was abysmal.
3.
by fluoridation 7d ago
Thanks, I'll give that a try. I basically have the same use case, only on Strix Halo.
4.
by fluoridation 7d ago
Interesting, I had assumed it'd be too large to fit. What quant and context size are you running?
5.
by fluoridation 10d ago
It's also possible if contributors agree to waive rights to their contributions, thus having multiple contributors and a single rights holder.
6.
by fluoridation 14d ago
Correct. We can either accept the claim or disregard it. The comment I replied to opted to accept it and then committed a fallacy, hence my response.
7.
by fluoridation 14d ago
Okay, but the anecdote states that every model repeated the pseudo-factoid about Foobar square, not just the 4 GB open source model equivalent of a tabloid.
8.
by fluoridation 14d ago
Yes, of that I'm sure. This optimization is only possible if the compiler has control of both sides of a call. If the function may be callable from other translation units or modules I imagine it generates a thin wrapper that's ex
9.
by fluoridation 14d ago
N/RVO works by (at the machine language level, of course) rewriting the function signature to return void and take an extra pointer parameter, which is written to before returning. If you're returning a newly-constructed object, t
10.
by fluoridation 14d ago
Unfortunately, I don't think there's getting away from just understanding value semantics to get the correct and/or performant behavior.
11.
by fluoridation 15d ago
So they go for a wildly shitty LLM translation?
12.
by fluoridation 15d ago
>It's reference material of all sorts. Technical books, manuals, software localizations, etc. Yes, I covered that under "formulaic content". As I've already said, a lot of this was already done decades ago. East Asian
13.
by fluoridation 16d ago
Those people were never going to pay a human translator, though.
14.
by fluoridation 16d ago
Manuals have been machine-translated for ages, well before algorithms were anywhere close to ready for the task. No change there. >a limited amount of diplomatic communiques With zero data to support it, I'd bet good money the vast
15.
by fluoridation 16d ago
AI will definitely not take over translation for a long time. Machine translation has improved to the point it's mostly no longer incoherent, but for non-formulaic content it's still very hit-and-miss, not that much better compare
16.
by fluoridation 16d ago
The thing is, there is no real evidence to back up the classification. The assignment between words and objects is arbitrary. A consistent definition for "planet" can be given and immediately after define the Solar System as being
17.
by fluoridation 16d ago
>the Moon's orbit is always concave towards the Sun Huh?
18.
by fluoridation 18d ago
It's human (as an adjective), yes, but it's not a human, because a human is a person, and there is no personhood without a brain. Is this line of questioning going somewhere? If the next thing you ask is whether a simulated huma
19.
by fluoridation 19d ago
So it's an exponential with a negative exponent? The thing about those kinds of curves is that since they have asymptotes to zero, it's hard to tell when you've plateaued.
20.
by fluoridation 20d ago
What exponential?
21.
by fluoridation 20d ago
I understand that. That review process is a bridge that imposes a maximum separation between design and implementation (that is to say, they can be separated so far, and no further). The two can't be fully decoupled.
22.
by fluoridation 21d ago
>The current LLMs are making it quite easy to seperate software design from implementation. I don't really agree. This kind of division of labor has always been possible, with architects doing the design and engineers/programme
23.
by fluoridation 23d ago
Should get a second one, then, if that's the meta.
24.
by fluoridation 23d ago
Fair enough. I've only ever used XMPP when Google chat used it and at work in the days before Slack.
25.
by fluoridation 23d ago
What do you mean? You have to accept a friend request before someone can message you. It's more spam-resistant than email, not less. That said, I don't know if there's anything that prevents spamming friend requests.
26.
by fluoridation 23d ago
What are you doing taking on that kind of risk and starting ventures if your earnings are low? I'm sorry, but if you're a delivery boy living from paycheck to paycheck, you should be trying to get an education, not start a busines
27.
by fluoridation 23d ago
I don't think any bank will give you such a huge personal loan that you could end up in debt for your whole life, since they typically want the loan repaid. The only way I can imagine that happening is if you're a sole proprietor
28.
by fluoridation 23d ago
Do they not have LLCs?
29.
by fluoridation 24d ago
You could get more bang for your buck. I'm currently freelancing at ~$50/hr and for every sixtyish hours I can live for two months, so I could afford to work half the year at a leisurely pace and not work the rest of the year, or
30.
by fluoridation 24d ago
I think there are actions a user may sometimes wish to take that software should simply refuse, ideally. Some actions are, and can only be, mistakes. I don't think it's a bad idea for software and hardware to defend itself against

More ›