ChatGPT: The Future of AI is Here! (1 Viewer)

[Isaac]

Mostly I don't see that it comes up with new things. It offers nothing in the way of real creativity. Everything it produces consists of stuff that is already available in some form elsewhere.

What it seems to have down is combining that old stuff into novel strings (or pixels or beeps). That's the essence of language and, by extension, coding - a performative utterance of sorts. So it's good at coding and putting words together. Yet its record on producing new information is spotty as hell: specifically, it likes to just make up whatever sounds good whether it's true or not.

And that would make me hesitant to trust it in providing ROI-positive decorating advice. Can it suggest a color for the wall? Maybe. But does it have anything backing up that color choice or did 'rainforest tangerine' just make a nice-sounding string of words?

And we are back to the question of what is the basis of the value of its output?

How can we know whether or not a big chunk of that basis for its value currently rests on whether people recognize, and therefore a firm and trust the output? It is definitely very good at outputting responses that make you think gee that sounds perfect and well-rounded.

Now I'm starting to put a little effort into recalling what if any is the average difference between what I usually find when I research a topic versus what it outputs when I ask it a question. And precisely what the meaning of that difference may or may not be.

These are all important questions to keep asking as artificial intelligence or anything that has that label continues to expand in usage.

Like other subjects, if asking the question gets a person in a lot of trouble, that tells you something. And the something that it tells you does not bode well for the outcome, with the exception of temporary revenue.

 

[Jon]

Mostly I don't see that it comes up with new things. It offers nothing in the way of real creativity. Everything it produces consists of stuff that is already available in some form elsewhere.

What it seems to have down is combining that old stuff into novel strings (or pixels or beeps). That's the essence of language and, by extension, coding - a performative utterance of sorts. So it's good at coding and putting words together. Yet its record on producing new information is spotty as hell: specifically, it likes to just make up whatever sounds good whether it's true or not.

And that would make me hesitant to trust it in providing ROI-positive decorating advice. Can it suggest a color for the wall? Maybe. But does it have anything backing up that color choice or did 'rainforest tangerine' just make a nice-sounding string of words?

What would it have to demonstrate for you to believe it is creative? When it writes a poem, for example, are you saying that the poem is not creative? Or if it comes up with 20 different product names for something, is that not creative? Are these generated product names, or new poem not an example of new information?

The colour choice for a wall would be an example of emergent properties that these Large Language Models provide, the property being intelligence. The corpus of human knowledge gets distilled into the recommendation.

Edit: Incidentally, the fact it gets into the top 10% of people taking the bar exam suggests that what backs up its answers is in most cases accurate. Or you could say, it is more accurate that most humans. It won't be long before it is more accurate than all humans.

 

[Jon]

Great video on more big AI announcements that have happened in just the last 2 days, including Large Language models for drug discovery, Google's competing Baird system, and some amazing image creation, such as text to video and image manipulation.

 

[Jon]

Some great prompt engineering ideas...

 

[Jon]

Thought I would try out one of the Q&A prompt engineering ideas. It's amazing!

 

[Isaac]

That is pretty cool ...

Though I enjoy playing devil's advocate to raise questions about things that I sense the public is 'assuming', that does not mean I don't think it is pretty darn awesome.

One thing your posts have served to teach/emphasize for me is how much the ability to give it precise instructions can be leveraged more than first meets the eye.

This thing is pretty cool ...
I'm sure as time goes on, people will figure out more and more ways to subject it to Tests of all kinds.

I am trying to imagine a scenario where a person in a field who is knowledgeable, versus a more entry-level person in a field, each tries to solve a problem. One uses ChatGPT and the other uses more manual research methods. They are timed and the business scrutinizes the results, process, methodology, etc. Figures out how much they can use ChatGPT and how much they can't, or more accurately generates cost savings/etc.

Can't imagine all the nuts and bolts of how that would work, but that sort of thing is bound to become needed to help businesses determine the proper approach to Reliance on it, risks, rewards, savings, etc.

 

[Jon]

Some analysis has already been done with Github Copilot. They state how much time programmers have been saving. However, GPT 4 is another magnitude more advanced. Check out how it compares on exams to GPT 3.5.

 

[AccessBlaster]

I would like to try Copilot in VS Code but I can't justify the subscription price for just fooling around

 

[Jon]

There are other AI tools that are free, like Tabnine.

 

[AccessBlaster]

Thanks, I might give it a try. Some of the extensions I'm using have some form of intellisense or autocomplete already.

 

[Jon]

ChatGPT is now going to open up an app store for plugins. This is a game changer. For example, one of them will allow for ChatGPT to go to a website, retrieve the data and then act on it. e.g. provide a summary, or answers and so on. I see ChatGPT becoming a central hub where people go for everything. If they want to visit the searched site for extra information, they can. But in many cases, why would they? ChatGPT will give them the answer to anything.

It kind of raises concerns for website owners, because their hard work becomes like a hard drive of answers for ChatGPT. They will miss out on ad revenue, which then brings into question the ability of the sites to survive. No idea how this is all going to play out, but I see ChatGPT as a real headache to the likes of Google, who provided a demo of their own system recently which received mixed reception.

 

[jpl458]

I gotta give you guys the heads up on this new piece of amazing tech. OpenAI have released the most incredible chatbot ever! It is so smart I guarantee you will be amazed. It is currently free during this phase, but will probably end up as a paid thing in the future. I recommend you sign up now and have a play.

Just go to https://openai.com/blog/chatgpt/ and click TRY CHATGPT.

To give a taster for how good it is, I am pasting my question to it and its answer. It wrote this thread title!!

View attachment 105112

You can ask it just about anything, ranging from coding questions, tutorials, recipes, and so on. Give it a go and let me know what you think.

This has been the most profound new piece of tech that I have come across since the birth of the search engine. No wonder they have over one million signups in the first week!

I used it to create an EXCEL spreadsheet that that has over 200 lines of VBA in it and I didn't write one line. I did it in a few hours and it runs like a champ. I just asked it how to do something and it wrote the code for me. And not once did it ask me why I was doing it that way. Had to make a few little corrections, but it works great for VBA. But you have to be able to read the code because occasionally the way you ask a question is really important, and the resultant code can be a little off. I have 3 monitors on my rig, and CHATGTP replaced Google on the left monitor when I am writing code. Plus, spelling errors virtually don't exist.

 

[Lightwave]

Its particularly good with code

 

[Jon]

Another nice prompt I picked up.

 

[Isaac]

AI expert warns Elon Musk-signed letter doesn't go far enough, says 'literally everyone on Earth will die'

Eliezer Yudkowsky, a decision theorist and artificial intelligence expert, is calling for a complete "shut down" of all AI development on systems more powerful than GPT-4, arguing it is obvious that such advanced intelligence will kill everyone on Earth.

www.foxnews.com

 

[Jon]

Read a good idea today. If you want a summary of a YouTube video, get the transcript and paste it into ChatGPT, and ask it to give the key points, or summary. Clever!

 

[Jon]

This is kinda cool. Ask it for a table format.

 

[Jon]

Some more stuff just for kicks.

 

[The_Doc_Man]

I disagree, Google will show you where the answer may be though links, it's up to you to decipher it.

Ai actually answers your question in the form of working code.

In the past (through ChatGPT 4.0) that has not always been the case.

 

[CJ_London]

Over the last few weeks, I've been getting phone calls from AI trying to sell me life insurance. They sound very human, but 'don't listen'. If you answer one of the questions with a bit of nonsense it assumes yes or no (whichever is in their favour) and you can't ask questions.

AI: Hi - how are you today?
Me: Well, not so good, had an accident with my lawn mower and chopped my feet off, I'm currently under general anaesthetic having them sown back on.
AI: That's great! How are you covered for life insurance?
Me: How are you today?
AI: That's good to hear, I'm sure we can provide something more cost effective

and if you stay silent, it just burbles on

Still got some way to go