Those apis are on 127.0.0.1, that’s a shorthand for “this local computer”. This is just different bits of the program communicating with each other.

You can see that its not going over the internet as the responses are coming back in microseconds.

Those are ancient models, if you want to try and use local models around the 2B mark I’d suggest trying either qwen3.5 or Gemma4.

No, the GPL doesnt put any restrictions on what you can do with code, just on that code and derivatives of it have to be released (source code included) under the GPL as well.

The courts have tended to find that training LLMs on data you have legal access to is fair use and transformative.

De-compiling isnt magic that lets you do whatever you want with software. It just works well for games because DRM has the insane model of “we give you the encrypted payload and the key to unlock it, but you can only unlock it in the way we want” which is fundamentally insecure. Most software isnt like that.

They don’t have to offer this to get new versions, that’s kinda the point of foss software, it’s there for anyone to download and use however they like whenever they want.

Are you American by any chance? I ask because in the UK we primarily used yards before metricification which are almost the same length as metres and there wasn’t any sort of long rumbling “ohhh yards are an awkward length” they were just used. I think you feeling that feet are more suitable is probably just familiarity.

Absolutely, I haven’t seen anyone in this thread say anything suggesting otherwise, everyone deserves that. But that means not throwing people out of a university without evidence.

I don’t know about this case, if that is true then its terrible and he should have been arrested, convicted and thrown out of the university. But the larger point stands, denying some people of due process because others are harmed in other ways is not something to strive for.

So two wrongs make a right? This person had their life horribly disrupted for reason A so an unrelated person should have their life horribly disrupted for reason B?

People should not be subject to arbitrary harm from others and we should be against all forms of it, whether that’s sexual assault, economic coercion or lack of due process.

The “sharks” are actually Russian submarines.

if you’re willing to lie why even bother with that, just make up the results wholesale. Polling companies don’t because their entire business relies on their reputation.

1000 people is enough to get a good estimate of opinion providing it is sampled well. Its one of the counter intuitive results of statistics that a few thousand opinions is enough to get a small percent margin of the average opinions of the sampled population no matter its size. The trouble with polling is that “average opinion of people who answer polls when asked” is not “average opinion of the group that you want to represent”, the whole art of polling is in finding the mapping between those two groups. But that isn’t helped by polling 10 or 100 thousand people instead of 1000.

Brain rot comes for us all, human or fly.

The “original sense” of AI is that it was the term used by computer scientists as an umbrella term for efforts into machine decision making. It didn’t turn into HAL from 2001 or EDI from mass effect for decades after that.

If Elon musk contributed 90% as much as all of Germany would you be equally happy saying Germany’s 80 million people held more responsibility than Musk for their emissions? Per capita is the only measure that makes sense as it is how much each person is contributing.

Another way of looking at it, Yunnan creates far less CO2 than the USA so by your logic the USA is worse than it. Repeat for every other Chinese province.

They are not only using human written data any more. Reinforcement learning through human feedback (RLHF) is a big part of it now, that’s the AI running through a problem multiple times with a human picking the best attempt and then using that best attempt as training data going forwards. The model collapse stuff from a few years ago was from an AI repeatedly ingesting its own input with no guidance over many training iterations.

huh fair enough, I hadn’t seen those before, they’re from before deepseek made a name for themselves (which as you say was with R1). They are very out of date though, as is R1 (from the end of 2024 and start of 2025 respectively). the current set of models from Deepseek come in 291B and 1.6T varieties and the previous version only was released in 685B

You’d almost certainly be better off upgrading to qwen or Gemma4 for a local model of that kind of size.

You almost certainly weren’t using a Deepseek model offline (unless you have 100s of GB of RAM/VRAM). You might have been using a small model distilled on Deepseek responses, but the smallest model made by Deepseek is about a 300B param model.

I’ve done this for a game I play. I’ve no interest in learning unity, but had a bunch of mods I’d like to make that are more than config edits. I was able to vibe code about 5 of them in an evening or two that do exactly what I want. They’re probably terrible code, but nothing else is built on top of them, I’m not distributing them and if they break they break no big deal. So pretty happy with it.

That’s not how this works though, hoping the law to just punishes people you don’t like without regard for what they are being punished for is madness.

If Anthropic lose this they’ll shrug pay a fee to licence things out of the untold billions they have available to them from credulous investors, and everyone else will have lost the right to do whatever they want with a copyrighted work without the publisher’s permission. That people who consider themselves progressive think that’s a win boggles my mind.