Menu Sign In Contact FAQ
Banner
Welcome to our forums

ChatGPT discussion, and ChatGPT-generated post examples

I bet you wish you bought Nvidia shares ChatGPT is a huge factor in their trillion $ valuation.

Administrator
Shoreham EGKA, United Kingdom

I bet you wish you bought Nvidia shares

I did! Though not as many as I wish I had.

LFMD, France

johnh wrote:

I did! Though not as many as I wish I had.

Same here….

alioth wrote:

Which kind of puts the AI panic into context. It’s going to be a while before the Cylons come and murder us all, I suspect

I don’t know. It must be ages ago since I read a book or small essay about the “singularity”. Computers taking over at some point in time is inevitable IMO. The wrong assumption we do is to think that they have to be self aware like us, creative like us, have wisdom like us (or some of us ), have feelings like us and so on. We think nothing will happen before they become like us. What has self awareness, creativity, wisdom and love got to do with any of that? Seriously, the only criteria is efficiency IMO. Whenever they can do something more efficient than us, they will take over that subject. Bit by bit all the different subjects will emerge together until there is nothing left for us to do than to play video games, have nice holidays, and fly SEPs, then hang around at EuroGA

Funny thing, as this thread have shown is that we don’t even need to hang around at EuroGA. We can have “bots” do it for us It defies all logics to do that, but it proves the point, bots are able to do it more efficient than we can, so they will. Entire forums may eventually be habited by bots exclusively.

The elephant is the circulation
ENVA ENOP ENMO, Norway

If EuroGA wasn’t moderated (by a human, well, sort of, some may disagree) you could definitely post entirely with ChatGPT, and to an outsider whose IQ is approximately at the Ibiza 18-30 Club level, it would all hang together. ChatGPT produces very correct grammar, and a sickly-sweet prose devoid of any opinions so nobody will get annoyed by it.

And this is already happening. It’s pretty obvious, too.

Specialised tasks, well suited to neural nets, e.g. MRI analysis, will be taken over by “AI” but that isn’t “AI” in the classic sense. It is a dedicated bit of code which has no idea of what a “tumour” is, just like ChatGPT has no idea what a frog is.

Administrator
Shoreham EGKA, United Kingdom

If you haven’t read it yet, i think this paper is a very interesting study of how to measure ChatGPT “intelligence” I think it shows it is quite better than one may think initially. Or at least what I thought: it’s just regurgitate some “kind” of linear interpolation of existing texts and content. it’s not.

I am no expert, but it seems at a high level the concept of embedding that turn semantic relationship into linear algebra is fascinating, and the fact just these “embedding” trained purely on scientific paper text can make actual scientific prediction is quite interesting and quite a new step if you ask me.

Good overview, but of course some of you may know more from the original word2vec papers…

Last Edited by roznet at 31 May 20:54
EGTF, United Kingdom

Peter wrote:

ChatGPT has no idea what a frog is.

ChatGPT begs to differ:

Yes, I am familiar with what a frog is. Frogs are amphibians that belong to the order Anura, which means “without a tail.” They are known for their distinct features, such as a short body, long hind legs adapted for jumping, and a unique life cycle that includes both aquatic and terrestrial stages.

Frogs typically have moist, smooth skin, which allows them to breathe through their skin as well as their lungs. They also have specialized skin glands that secrete toxins, serving as a defense mechanism against predators. These toxins vary in potency among different frog species, with some being highly poisonous.

Frogs are found in various habitats worldwide, ranging from tropical rainforests to deserts. They are known for their ability to jump long distances and catch insects with their sticky tongues. Frogs are also well-known for their distinct vocalizations, which they use for communication and mating purposes. The sounds they produce are often referred to as “croaking.”

However, it then proceeds to demonstrate that it has absolutely no clue:

It’s worth noting that my knowledge is based on information available up until September 2021. If there have been any significant developments or discoveries about frogs since then, I may not be aware of them.

Last Edited by Cobalt at 31 May 22:51
Biggin Hill

If you haven’t read it yet, i think this paper is a very interesting study of how to measure ChatGPT “intelligence” I think it shows it is quite better than one may think initially. Or at least what I thought: it’s just regurgitate some “kind” of linear interpolation of existing texts and content. it’s not.

Thanks roznet, interesting. I also thought that these LLM’s are far away from intelligence as we know it and, to be cynical, are a great tool for many politicians (“know nothing about the subject but make sure to sound smooth”) but I need to change my view then.

We can all see the tremendous advantages of ‘narrow’ AI (autopilots, medical diagnosis and so many other areas) but AI beyond that poses some great challenges. As usual, tech is way ahead of any kind of legislation of AI, if legislation can even be applied. And how to implement legislation on a worldwide basis? I can see big differences between different countries’ ideas on this coming up. Not to mention misuse of AI by certain regimes of course.

Private field, Mallorca, Spain

As usual, tech is way ahead of any kind of legislation of AI, if legislation can even be applied.

By far the largest problem with legislation of any kind is that those who bring up laws are usually absolutely clueless about the topic. At least that’s how it is in Croatia – the majority of our parliament and politicians in general are incredibly low educated, ill informed and at the same time absolutely confident how they are the smartest part of our population – typically suffering from Dunning–Kruger effect.

The process of adopting the law (the technical part) goes like this:
- the government defines initial proposition
- it’s printed and delivered in paper form to representatives in parliament
- they read it and make a list of amendments, print, sign it and send back
- then some back office people take this amendments, photocopy them, cut each with scissors and glue them in original proposal printed with additional space/blanks for amendments
- then back office scan this Frankenstein monster, makes a PDF and send it as a final proposal to representatives

Can you imagine brightness of mind who invented this procedure? Looking that, any level of AI can replace such genius.

LDZA LDVA, Croatia

I don’t think the world will be worth saving, because the IQ will decline to zero all by itself, as a result of massive false eyelashes blocking out sunlight

Administrator
Shoreham EGKA, United Kingdom
Sign in to add your message

Back to Top