AI megathread

Are you active in this area?

David

Well I work for a company that does use some creative aspects in its software.

And I am looking at the prospect of using Large Language Models, since I think everyone in software has to get up to speed or be left behind...for better or worse.

OTOH I would do my best to avoid pure AI art but I can see some circumstances where an artist is using some AI based tools for the same reason I need to learn about LLMs - to stay in business and put food on the table.

Unclear how much life will be left in these machine "learning" companies at the end though ->

Congress Wants Tech Companies to Pay Up for AI Training Data

David, if you haven't invested a fair amount of time in using the LLM AI tools, you're really incapable of opining on its uses and limitations.

Well I am a retired software developer - but since I worked for a small firm, I still have some involvement in the work I did.

LLM's are pretty new and I have explored them in a few directions by now. They are very interesting, and I have done explorations into getting ChatGPT to write small quantities in Python code in relation to a particular library. What I discovered was that its first attempt was syntactically invalid, but when I pointed this out, it apologised and came up with a viable solution.

I have also explored ways to persuade ChatGPT to discuss some of the parapsychology literature by using a form of language that gives it license to explore non-materialist possibilities. I have, of course, written about the results here.

You must decide if my advice is going to be of use to you!

I've only recently begun to try and "use" it in my professional career.  Its choppy in terms of how much efficiency I've gained (likely negative as there is a learning/creative-use curve).  However, just yesterday I was struggling with how to phrase an email to a prospective client.  I ask ChatGPT with a simple prompt to suggest some language.

Well I don't usually have 'delicate' emails to write, so I leave that job to myself - perhaps that is why you and I tend to jar a bit - LOL!

I was amazingly good, but certainly not a final draft.  That said and for me at least, when I am writing for my job the first draft is far and away THE most time consuming part.  This example allowed me to move from initial drafter to draft editor.  Took me 30 seconds to edit the AI draft and bam..... off went my email.  Probably saved me 10 minutes.

While the bigger questions around AI continue to be a point of potential concern, the day to day application potential seems exponential.

Well do keep us posted! Maybe you should explore its use in writing your posts on this forum?

David

Well I work for a company that does use some creative aspects in its software.

And I am looking at the prospect of using Large Language Models, since I think everyone in software has to get up to speed or be left behind...for better or worse.

OTOH I would do my best to avoid pure AI art but I can see some circumstances where an artist is using some AI based tools for the same reason I need to learn about LLMs - to stay in business and put food on the table.

Have you explore this site?
https://www.img2go.com

Unclear how much life will be left in these machine "learning" companies at the end though ->

I don't either, but remember some of the previous computer fads - think of Mandelbrot (and other) fractals, the acres written about RISK architectures, fifth generation computers (if you can remember that far back), etc.

It probably takes a while to see these things in perspective.

David

Well I work for a company that does use some creative aspects in its software.

And I am looking at the prospect of using Large Language Models, since I think everyone in software has to get up to speed or be left behind...for better or worse.

OTOH I would do my best to avoid pure AI art but I can see some circumstances where an artist is using some AI based tools for the same reason I need to learn about LLMs - to stay in business and put food on the table.

Have you explored this site?
https://www.img2go.com

Unclear how much life will be left in these machine "learning" companies at the end though ->

I don't either, but remember some of the previous computer fads - think of Mandelbrot (and other) fractals, the acres written about RISK architectures, fifth-generation computers (if you can remember that far back), etc.

It probably takes a while to see these things in perspective.

David

Apparently if you ask ChatGPT if it uses copyrighted material for its answers it will reply "Yes".

DPD error caused chatbot to swear at customer

DPD has disabled part of its online support chatbot after it swore at a customer.

The parcel delivery firm uses artificial intelligence (AI) in its online chat to answer queries, in addition to human operators.

But a new update caused it to behave unexpectedly, including swearing and criticising the company.

https://twitter.com/ashbeauchamp/status/...9104450874

DPD error caused chatbot to swear at customer



https://twitter.com/ashbeauchamp/status/...9104450874

Yay! ...there again, it could be clever viral marketing for dpd... you never know these days.

DPD has disabled part of its online support chatbot after it swore at a customer.

OMG ! It's becoming ....

Huge Proportion of Internet Is AI-Generated Slime, Researchers Find

Maggie Harrison

The internet's steady fall into the AI-garbled dumpster continues. As Vice reports, a recent study conducted by researchers at the Amazon Web Services (AWS) AI Lab found that a "shocking amount of the web" is already made up of poor-quality AI-generated and translated content.

The paper is yet to be peer-reviewed, but "shocking" feels like the right word. According to the study, over half — specifically, 57.1 percent — of all of the sentences on the internet have been translated into two or more other languages. The poor quality and staggering scale of these translations suggest that large language model (LLM) -powered AI models were used to both create and translate the material. The phenomenon is especially prominent in "lower-resource languages," or languages with less readily available data with which to more effectively train AI models.

In other words, in what the researchers believe to be a ploy to garner clickbait-driven ad revenue, AI is being used to first generate poor-quality English-language content at a remarkable scale, and then AI-powered machine translation (MT) tools transcribe said content into several other languages. The translated material gets worse each time — and as a result, entire regions of the web are filling to the brim with degrading AI-scrambled copies of copies.

Huge Proportion of Internet Is AI-Generated Slime, Researchers Find

Maggie Harrison

Amazon Is Selling Products With AI-Generated Names Like "I Cannot Fulfill This Request It Goes Against OpenAI Use Policy"

Victoria Tangermann

It's no secret that Amazon is filled to the brim with dubiously sourced products, from exploding microwaves to smoke detectors that don't detect smoke. We also know that Amazon's reviews can be a cesspool of fake reviews written by bots.

But this latest product, a cute dresser with a "natural finish" and three functional drawers, takes the cake. Just look at the official name of the product listing:

"I'm sorry but I cannot fulfill this request it goes against OpenAI use policy," the dresser's name reads. "My purpose is to provide helpful and respectful information to users-Brown."

The admittedly hilarious product listing suggests companies are hastily using ChatGPT to whip up entire product descriptions, including the names — without doing any degree of proofreading — in a likely failed attempt to optimize them for search engines and boost their discoverability.

It raises the question: is anyone at Amazon actually reviewing products that appear on its site?