During inference ?
Tetsuo
To my surprise almost if not all LLM are set not to be deterministic and have one unique input result in always the same output.
They all are set to have a "temperature" setting so that they sound more natural.
Personally I think quality of the output tokens of LLM is surprisingly not as much their priority as the quality and truthfulness of the result.
I would 100% prefer a LLM that is purely deterministic and repeats the same answer exactly to the same question. Instead LLM are constantly choosing the next likely token more TL appear human rather than being accurate.
These LLM are designed as sycophants and set up and trained as such.
So a fingerprinting in the output seems quite realistic. An LLM is not giving you it's best most likely answer. It's taking one of the most likely answer and adds a sprinkle of uncertainty and randomness on top of it so it looks natural...
I applaud the open source Chinese makers. They might be "distilling" from the closed weights models of competitors but since they themselves pillage the internet with no oversight in my book that's fine. In some way the Deepseek, minimax and kimi makers are forcing the weights of bigger proprietary open.
Right now on OpenRouter a large part of the processing goes to Deepseek. Deepseek is still weaker than the Claude "frontier" models but they are way cheaper. Since they are open weights anybody can compete as providers and the cost stays "low". These AI Chinese makers are a constant threat to the american proprietary models and that's a great thing for the consumers.
IMO OpenAI should stfu considering they completely dropped their good open source form for pure greed.
If they want to compete in the open source arena they should push their own models there ? IH wait that would mean releasing open source and not making as much money ?
Fuck spez.
There isn't enough fingers in this cookie jar.
Personally I downvoted you because you completely forgive/forget the older devices from Apple.
They had sold approximately 2.2 billions of impossible to repair devices before they kind of changed course with the iPhone 14 line.
And to say they had "Don't be evil" as their official motto in the early days...
Be careful, because we will all pay the price when the bubble pops. It won't be just the CEOs paying the bill, it will be governments and people.
I disagree I think the fact the boss of wikimedia is very well paid and that he is benefiting at the end of day from free labor.
They could reward/pay the most important contributors and they don't. Yes it's voluntary but it's also a comparatively very "" small "" group of people that write a big chunk of wikipedia.
I absolutely can't stand that the boss of such utility service as wikipedia thinks first about his salary rather than rewarding their contributors.
Also you say this is not a problem but I'm sure the Union would have pointed out the same thing I just did.
That's not the problem, 'tho. Wikipedia is volunteer-written and needs money to pay for the SW development and the servers.
Also you are missing an important point if you just say "hey they are volunteers, they choose to work for free". It's not like that because as I said it's a few contributors that do the bast majority of article edits.
I'm wasting my time if you think that wikimedia has one problem and that's it. Multiple problems that would need to be adressed by a Union.
While the "volunteers" making it happen for real get nothing...
Why buy these kind of monitors in the first place ?
I think there are plenty of monitors that don't have weird online requirements or spyware in it so why buy them at all ?
Everytime someone buys a monitor like that they normalize the practice.
A monitor has one job -> displaying images.