I am using Chatgpt and Grok and they have full internet access. Even deepseek has internet access now.
which hasn’t be your statement and which doesn’t address potential biases or even the fact that - as a LLM developed in China - is faced with the very same limitations the Chinese citizens are confronted with. And that is a strict censorship, without access to Western media.
I did not intend to make a political discussion out of it. But both news sources on the censorship of deepseek address the very same fact that deepseek simply doesn’t answers questions related to Taiwan and most likely. And also other topics that won’t align with the agenda of the communists party. Therefore it’s not really relevant to make the distinction between “de facto” or “de jure”.
Not every LLM is subject to censorship, indeed. LLMs aren’t proactively updating their parameters by gathering the latest web results. In that context - specifically related to the so called knowledge cut-off date of an trained LLM - there is, as a matter of fact - no re-evaluation happening on the fly. In short, almost every LLM can provide an outline about almost every significant pieces of literature. But legally the LLMs are prohibited to cite the literature directly.
If by “intelligence” one means “problem solving skills via pattern recognition”, then yes, i think we can talk about an intelligence in AI. But if your definition of human inteligence is beyond that, maybe then the use of the word is a misnomer.
Well, I see that you are still unwilling to admit that you use a “non true” claim to discredit a LLM which very well may be censored and biased.
For doing that, you don’t have to recourse to a statement that cannot be backed up by fact and sources.
As I said, I don’t dispute your argument about “non reliability” of LLMs wherever they may originate from. I dispute your statement which is intended to backup that argument.
Anyways, enough said from my part on this matter. Thank you and so long!
That is your subjection and your assessment that Taiwans legal status is “undetermined” at best.
It might be a non true claim. It totally doesn’t make a difference if it’s true, false or “complicated” within this context.
Simply ignoring the fact that LLMs are prone to the knowledge-cut off of their training stage just by stating that those are capable to perform web searches upon request, is simply not solving the issue. And that has been the statement which I was replying to. The major culprit is that those models are outdated pretty fast, maybe faster than the GPU clusters they were trained on.
From my point of view: The chase of General A.I. or such things - is definitely not an sustainable approach in the long term. It’s chasing a myth, in anticipation of the next breakthrough and a potential paradigm shift.
Prove me wrong with objectivity. Fact, resources…?
And you know what, in my real “real life”, I take it as a sign of maturity if people admit that they’ve got something wrong. I appreciate myself to be corrected if someone points out to me that in one or another matter I am mistaken. I will stand corrected and learn something from it.
I don’t see any such sign from your side. You make an obvious false claim to prove a point and you don’t back off from it.
You can still write GBs of text about this and that LLM and their flaws. That’s your time and you know how to spend it best.
I, personally, wouldn’t waist a second more engaging with a person who chooses consciously to ignore facts and attack by accusation of subjectivity on my part.
Doesn’t AI, given the proper circumstances, always choose to control/enslave and/or eliminate/kill humans to resolve whatever issue is at hand?
Or, have they resolved that AI peculiarity?
The rush to AI seems a bit foolish to me but I do not know enough about it to make an educated decision one way or the other.
I certainly have used ChatGPT.
I have also used Video2x 4K Video upscale software to.
I do not have an AI based search engine though, for example.
I do not want AI integrated into my OS as some kind of personal assistant.
I do not trust it enough to allow it into my personal space yet.
Ah well.
It is coming, for better or worse and this kind of caution by the general masses for a new invention is always the case whenever something this life changing is introduced to our society.
For example, the same kind of caution came when the they invented gun powder, the gun/rifle/cannon or any other explosive based projectile weapons, the steam ship, the chainsaw, the automobile, the airplane, the space ship - ALL of it was met with suspicion and caution when they were invented.
sidenote: AI ? what is that thing you allways call AI? LLMS i bet ? thats not really the same .. AI.. real AI is not something you already tried in most cases.
[AI-content]
AI is the broad field of creating intelligent machines, while LLMs are a specific type of AI focused on understanding and generating human language. Think of AI as the entire forest, and an LLM as a specific, very important type of tree within that forest. LLMs are a subset of Generative AI a branch of AI focused on creating new content.
Thanks for this. I am a Trek fan too. Spock and Data were my favorite characters. I always thought Data was more human than anybody on that ship. And that actually confirms it. We always fall prey to our own storytelling biases and when we try to create a character that is not human, it ends up being some kind of human, and in Data’s case, the perfect human.
