Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)MG
Posts
0
Comments
195
Joined
1 mo. ago

  • You're missing a crucial detail- the discussions of LLMs as just being probability machines ignores the fact that when we're talking about "what is the most likely next word?" The answer to that question isn't merely just "What, in the training corpus, was the most likely word to follow in this instance?" But rather there is a "finger on the scale" so to speak in favor of certain types of responses, and this is frequently updated. You cannot have a useful LLM without this because it will talk as if it is a human with a sense of self, display blatant prejudice just because it's common, and say creepy things (see the Microsoft "Sydney" fiasco) because the humans who wrote the sources in the training corpus do have a sense of self, do hold prejudices, and express thoughts and feelings that are inappropriate coming from a chatbot. When done intentionally and carefully, this creates a much more useful product, but when done poorly it potentially makes things worse. It seems that at least part of that weighting is based on user interaction, which is what I was talking about with models getting dumber the more they interact with the general public.

    Furthermore, the newest versions of ChatGPT attempt to include "reasoning" as an actual feature of the response. I've played around with them and they are definitely a lot better at logic and math problems than older models but not necessarily less prone to "hallucinations" when it comes to factual information. I haven't read a whole lot about how the "reasoning" works because I have been a lot more interested in non-LLM methods lately but it is intended to combat the issue you described. Personally I am not convinced this will fix much of anything in its current strategy but it's certainly interesting to see.

  • Yeah the "nobody understands why" is an absurd statement when it's pretty obvious, they're training it on itself and keep trying to "improve" it with shit that's making it worse. Plus, if it's using reinforcement learning based on interactions with the general public (learning based on user responses to and ratings of bot responses) the more "like" the general public the LLM will become, aka stupid. Furthermore, I have a personal theory that LLM power users who aren't programmers are dumber than average or at the very least less creative, and this will also skew results if user responses are given a heavy weight.

  • This is largely untrue. I have worked in manufacturing my entire adult life and in my experience people are very eager to do manual labor type jobs if those jobs pay well and provide stability. The problem is that most of them don't. It would make everything more expensive to pay everyone doing these jobs better but it would be worth it longer term by making a society that doesn't just rely on there being a constant supply of an artificial second class of people that can be underpaid and exploited with impunity. When people say "nobody wants to work manual jobs..." the implied rest of that sentence is "...to make the same amount of money as someone working retail."

    I hate Trump and his idiotic tariffs, but this argument that we need immigrants to do all the jobs Americans don't want to do is based on the racist idea that Americans are too good to do these jobs- the reality is, they are simply not desperate enough to take them for the amount of pay that is being offered. It's a blatantly false narrative and it only serves to harm anyone left of Mussolini.

  • I feel like I'm taking crazy pills with all of these comments talking about this as if it's a real photo and not AI slop looks like complete shit. Are the comments AI generated spam too? Are they just stupid? Am I AI generated spam?

  • I stopped using mainstream social media in 2019 but my accounts are still active so I can snoop on random people I went to college with and holy shit every time I get on Facebook it's so much worse on ways I don't even understand. Most recently I got on to look at something and my feed was completely unrecognizable because it was all AI generated slop from pages I have never heard of and not any updates from people I know. It's crazy what people will accept if it's done slowly enough I guess. I legitimately don't understand why anyone would use Facebook as it exists today. At least when I quit I could at least understand why people used it.

  • This is one of those bizarre Lemmy echo chamber things. I've never seen this sentiment that advertising is evil and should be stopped at all costs anywhere else but on Lemmy it's super common. Idk where it comes from. I get that advertising kind of sucks but it just seems like a weird thing to get so passionate about especially considering how many other things are wrong with the world. Sorry you're getting downvoted to hell, you're not crazy, Lemmy is.

  • FYI according to CPI inflation $2,500 in 1972 is worth about $19,000 today. Based on a comparison of median individual income that would be similar to about $25,000. Either way, the truck cost about the same as a Honda Civic does today.

  • I will say a lot of what you've discussed here is actually illegal but very rarely enforced. Pretty much every small business owner I know is pulling shit like this but it's basically never enforced even though it's illegal fraud.

  • Pretty much any tax avoidance loopholes. The more money I have the more I see how ridiculously skewed in favor of the rich everything is. My income is taxed at a lower rate than my capital gains, meaning that not only did I make several thousand dollars last year on stock sales I did literally nothing to earn, but I paid very little on taxes for it. There is also a scheme a friend of mine uses to reduce his tax burden even more by recording losses that only exist on paper by swapping between essentially equivalent assets. The system is designed to punish poor people for being poor and reward rich people for being rich.

  • That is a conspiracy theory that can easily be disproven if you look literally anywhere outside of Western Europe. Muhammad started Islam and composed the Qu'ran at the beginning of that time period, and the Islamic calendar begins in 622 AD. According to the Phantom Time theory, basically none of early Islamic history actually existed and it's a fabrication by... The Pope?! It's an absurd premise. The Abbasid Caliphate was founded in 750 and still existed in 911, with all sorts of written and archeological evidence in favor, and let's not forget about the other Roman Empire that existed in that time frame and have their own historical and archeological evidence, including about events that overlapped with Western histories such as Irene of Athens being pressured to marry Charlemagne. Again, it stretches credulity that the Byzantine Emperors and Eastern Church or any Muslims whatsoever would be swayed to fabricate so much of their own history by the Pope. This theory is absurd at best and racist at worst because in erasing so much Western European history it just kind of assumes literally nothing was happening anywhere else in the world either.

  • As someone who has used it for viewing legal copyright free content and nothing pirated, I will say that it definitely provides a service that regular torrents don't- ease and speed. It makes viewing legal and copyright free content just as easy as using Netflix to the point where you don't have to go onto various websites looking for what you want to watch, you can just get it all in one place. Further, if you have roommates or relatives who would like to view this legal and copyright free content as well, it's easy enough for them to use as well without having to learn how to find and download said content. I have done both, and I must say that the fee is well worth it to me compared to the amount of time it saves.

  • It's a French company that you pay a small amount of money to use their service to access perfectly legal and open source content, and not anything resembling pirated movies and TV shows. If your son is using it to view pirated TV shows, that would be very bad and against their instructions for how to use their service. However, it is possible that this is what he is doing.