Hi Chris, that’s true, but even if the answer is wrong, the suggestions often provide some very useful clues as to where to look for causes (and the AI model has not changed anything, so has done no harm … and can ask several).
#DigiBlogChat

Thank YOU, Carol!! #DigiBlogChat

Oh heck, Larry! All the best for getting over this very soon! #DigiBlogChat

AI models can find (or at least suggest) why some software is doing something incorrect or not doing something correct in seconds, which would have taken me, not minutes but, hours or possibly days to figure out.
#DigiBlogChat

A10 #DigiBlogChat
Who knows?! I expect that people will find that the most useful applications are different from those expected today.
For example, in the software world, everyone seems to focus on generating code, but my experience is that by far the most useful application is in fault diagnosis.

BTW in a joke interview, when asked that: after a mathematician replied "exactly two plus exactly two equals exactly four" and a physicist replied "approximately two and approximately two is approximately four", the system analyst leaned forward and asked: "what do you want it to be?"
#DigiBlogChat

That could be a lucky prediction to fool us all! #DigiBlogChat

A7 #DigiBlogChat
The LLMs just produce (mostly) coherent words and sentences. The risk is that, because it refers to relevant topics and appears to be coherent, that some people believe it.
But that's not different from people spewing nonsense.

True, but it's all just trying to control the processing based on text ("natural" [but to whom?!] language), even in fields where that language does not apply. What happens when I ask an LLM: "What's two and two?' ... or similar more complex questions? #DigiBlogChat

Q6 #DigiBlogChat
People forget that current AI systems are **language** models!! They manipulate text!!
AI systems of the 1980s would not be making these kinds of nonsensical mistakes/hallucinations/confabulations ... they were "expert systems" operating on rules specified by subject domain experts.

It's a fast moving area of innovation, so people are at a wide range of stages of knowledge and adoption, and in a wide range of aspects of the subject ... so it's multidimensional.
#DigiBlogChat

A4 #DigiBlogChat
Showing how far behind the (b)leeding edge I am, I haven't heard of any of those (except for Zapier itself).
I'm applying AI to software development, but even there, unlike others, I can't not look at the code!

A friend didn't see a winter for several years, alternating between UK and NZ, and I've heard of others. #DigiBlogChat

A3 #DigiBlogChat
Pretty much anything involving text (after all they're Large *Language* Models): so reading and writing of textual messages, including programming language source code.
When applied to anything else they are just blunderbusses

People have done it, and still do. E.g. between UK and Aus or NZ. #DigiBlogChat

A2 #DigiBlogChat
Only a little, for software development. But I don't trust it enough (yet). People I know have them running all the time (some on local machines for which they pay no fees to the AI companies).

Of course, it's geometry ... we can't have the longer summer days without the shorter winter ones. Actually, we can, but it involves migrating between hemispheres. #DigiBlogChat

A1 #DigiBlogChat
I expect that we're all at different stages of understanding this fast moving field. But, as I understand it, they're both generative and the difference relates to whether it waits for a prompt before acting.

Hi Larry! Are more pensioners joining #DigiBlogChat this week?