OP asked about ChatGPT literally no one said anything about any other AI model.
You made the argument that others were “tarnishing ChatGPT’s reputation for no good reason”
You then go on to clarify “AFAIK” meaning you’re hedging your bets and you don’t really know. Then you go on to say you assume ChatGPT is “good at aggregating info”.
Within your first three sentences you “stated an argument” then stated two assumptions to back them up. Great, we’re on a level playing field or whatever bullshit doubletalk you feel like throwing around here. But then you’re immediately told that you’re wrong and then told how it works. Instead of realizing your base assumption is wrong you keep doubling down with more assumptions in the face of people who know better than you and aren’t assuming.
So it’s all well and good that you want to pretend you’re trying to have a conversation in good faith while pretending your basic assumptions mean anything, but since your assumptions are wrong your arguments are simply invalid and you come off like you’re just trying to be right instead of having an actual conversation.
I completely understood what you posted, and do the other commenters here. We just all know that you’re wrong and we don’t have to make assumptions. ChatGPT isn’t aggregating a fucking thing, it’s using a predictive model. You fundamentally don’t understand what you’re talking about in the face of someone breaking it down for you Barney style.
Again, as far as I know, it’s the same model. Different end products doesn’t mean different AI models behind the scenes, so their computations could be the same (but the result can be different — e.g. DuckAssist always giving sources whereas ChatGPT doesn’t (unless asked?)).
Again, I interpreted OP’s question about AIs in general. You don’t have to become angry that I did so. It’s okay to have discussions that are a bit more encompassing than what was specifically mentioned so long as it’s relevant or helpful.
Don’t really know but have some knowledge and stating my limited knowledge and being transparent about the limitation of my knowledge is, IMO, contributory to the discussion rather than hindering it as you’re accusing me of.
That’s how arguments are made. Hell, that’s how very solid arguments are made in general. Do you know about axioms in math? Mathematicians advance the scientific world and make great things, and even they make assumptions. At least they state their assumptions, though. I’m doing the same.
Chill down.
See @ScionOfUrza411 for how to have a thoughtful and polite disagreement with me. Don’t speak on behalf of other people like you’re some kind of authority.
Could you have stated this thoughtfully and respectfully? Next time, do that please.
I don’t get this reference.
You’re complaining about ChatGPT while I’m saying DuckAssist is good. There’s nothing wrong, contrary to the “self-conflict” you’re trying to point out here, in suggesting people to use the supposedly better product (DuckAssist) rather than the product which you yourself think is bad (ChatGPT). That’s not hypocrisy, it’s expected.
Because of your disrespectful tone, I’ll disengage from the conversation effective immediately. I’ll mute the thread so pinging me won’t work anymore.
Don’t bring me up, I completely agree with @bobross419 here, the main difference is that I was going at specific points and not your entire argument, and that I didn’t curse, which is just because I don’t curse in general.
Just to chime in for no particularly good reason. But I just don’t get the impulse to use ChatGPT or any other generative garbage creator for literally anything.
They will lie confidently to you. I can’t stand that in people, I avoid people that do that because I can’t deal with bullshitters. And it defeats the purpose it is trying to fill. What is the point of a tool that I have to double check all the time?
And, finally, why not ask people with decades of experience that are more than happy to talk about the game? People on this forum alone have put in hours to give advice to people that comes from deep understanding of the game. Then there is reddit, discord, etc.
LLMs have no concept of domain knowledge. They do nothing to create loci of related concepts that inform the correctness based on triggers with adjacent concepts. They literally just predict a statistically “most likely” word (after inserting some randomness), with no judgment of correctness involved whatsoever in determining “most likely”. So it really doesn’t even make sense to say that LLMs “lie”, because “lying” requires having some way to know the correctness of the lie. They also don’t really reflect reddit mistakes, except in a far-removed way that is pretty negligible. Consider the “tomb strategy with mummy of fedhas using cleansing flame” advice, a strat made of separate concepts so wildly unrelated to each other that it’s inconceivable to think someone would string them together even to make bad advice.
Any use of an LLM will be subject to this, and LLMs are what the current “AI” explosion all are, so this statement applies to every AI on the market. It’s really important for everyone to understand this before they start giving these things high-stakes responsibilities.
Unfortunately, we’ve already let them do all sorts of critical things and it is going to be terrible for everyone. On mastodon a doctor was talking about how their institution is using a AI for charting purposes. People are going to die.
Our world is run by know nothing bullshitters and they’ve build a machine that reflects them. They are too stupid to realize they are being conned.
And don’t get my started on the climate change implications.
chatGPT is pretty good at writing SQL queries for example.
With code in general chatGPT will look up the function and method names you need and how to call them, and implement it all, and it will call ~80% of functions and methods correctly. That’s a massive time-saving.
chatGPT is also nice for paraphrasing something you’ve written or writing a first draft of a letter. In both cases I wouldn’t want to keep the chatGPT response entirely, but it provides a useful series of sentences.
I think Ed Zitron is rubbing off on me. Also, I’m very bitter about our sclerotic political class and their inability to deal with fundamental and human life on this planet ending problems.