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?
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.
Thanks, I'll give that a try. I basically have the same use case, only on Strix Halo.
Interesting, I had assumed it'd be too large to fit. What quant and context size are you running?
It's also possible if contributors agree to waive rights to their contributions, thus having multiple contributors and a single rights holder.
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.
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.
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…
No, silly. If you pray for something and it comes true then praise be to god; if it doesn't then you did it wrong.
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…
Unfortunately, I don't think there's getting away from just understanding value semantics to get the correct and/or performant behavior.
So they go for a wildly shitty LLM translation?
>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…
Those people were never going to pay a human translator, though.
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…
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…
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…
>the Moon's orbit is always concave towards the Sun Huh?
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…
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.
What exponential?
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.
>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…
Should get a second one, then, if that's the meta.
Fair enough. I've only ever used XMPP when Google chat used it and at work in the days before Slack.
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?
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.
Thanks, I'll give that a try. I basically have the same use case, only on Strix Halo.
Interesting, I had assumed it'd be too large to fit. What quant and context size are you running?
It's also possible if contributors agree to waive rights to their contributions, thus having multiple contributors and a single rights holder.
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.
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.
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…
No, silly. If you pray for something and it comes true then praise be to god; if it doesn't then you did it wrong.
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…
Unfortunately, I don't think there's getting away from just understanding value semantics to get the correct and/or performant behavior.
So they go for a wildly shitty LLM translation?
>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…
Those people were never going to pay a human translator, though.
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…
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…
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…
>the Moon's orbit is always concave towards the Sun Huh?
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…
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.
What exponential?
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.
>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…
Should get a second one, then, if that's the meta.
Fair enough. I've only ever used XMPP when Google chat used it and at work in the days before Slack.