How are you reacting to the AI explosion of tools?

🙋 Questions
1

I just added another prompt to my prompt list and that made me wonder what others are doing about this explosion of technology.

  • Are you interested in it?
  • Worried about it?
  • Playing with it?

I’d love to hear some thoughts

2 Likes
2

If it’s ChatGTP…

I’m definitely interested in it.
I’m definitely not worried about it.
I’m definitely playing with it.

I find it speeds up some tasks, and slows down others.

I’d say use with caution. Validate and sense check what it’s saying, much like you would with a spell checker or Grammarly. Understand what its doing if you’re taking bits of code out of it. Use it to compliment your own research, don’t just take what it says as a given. Appreciate it’s limitations.

5 Likes
3

Totally agree with the above reply…

Its exciting, although I haven’t been playing with it
But I am enjoying reading/hearing peoples stories of how they’re using

I’m being very mindful to hear from and read more by the critics of AI - In particular James Bach has been quite vocal and I believe coined the term AI - “Automated Irresponsibility” he makes a really great point about being people not validating the responses and having blind faith in its abilities

3 Likes
4

AI can be useful - when I’m moderating on the SQA stack exchange I spend a fair bit of time using an AI tool to detect AI-generated answers (although honestly anything that’s sufficiently abstract, waffly, and talks around the question seems to generate a bunch of flags for being AI-generated, whether the author normally writes that way or not).

It’s not truly artificial intelligence, no matter how much it looks like it, so treating AI tools as if they’re actually intelligent in a human way is… risky. Along the lines of the aborted experiment teaching monkeys about money risky (they discovered prostitution rather quickly).

I prefer to see it as a quasi-AI that can do things in its core competency area very well, but is shaky at best outside that range. From what I’ve seen, ChatGPT is very good at synthesizing publicly available information into a more or less coherent overview, but not so much at determining whether its source is factually accurate (now that would be a fun test, although the results could be… interesting).

Some of the other tools are getting closer to the ability of a human to detect patterns of various types, although the amount of processing they need to do it is still both daunting and awe-inspiring (we have that much processing power crammed between our ears?)

So in summary, interesting, exciting, and use with caution.

1 Like
5

I’ve been looking at a few AI chats with a view to seeing what help they could be and I believe it is very dependent on how you ask and what you ask for.
Prompt Engineering is all about how you get information from the AI. I just asked ChatGPT what is the best method to get answers from it and this was the reply.

As an AI language model, I am designed to provide answers to questions based on the text provided to me. You can ask me anything related to the topics I am knowledgeable about, and I will do my best to provide a helpful and accurate response.

To get the best answers from me, here are some tips:

  1. Be specific: The more specific and clear your question is, the easier it will be for me to provide a helpful response.
  2. Provide context: If your question relates to a specific situation or problem, providing relevant context can help me better understand the question and provide a more relevant response.
  3. Use proper grammar and spelling: Using proper grammar and spelling can help me better understand the question and provide a more accurate response.
  4. Ask follow-up questions: If my response doesn’t fully answer your question, feel free to ask follow-up questions or clarify any confusion you may have. This can help me provide a more complete and helpful response.

So for me success is based on the quality of the questions asked.