AI In the Nov 2023 Enterprise

  • Comments posted to this topic are about the item AI In the Nov 2023 Enterprise

  • The state department I work for has blocked all generative AI. I don't know why. It might even be blocked for all state departments. I don't have access to that information, so all I can do is speculate as to how widespread blocking of generative AI is.

    In my own experiments with ChatGPT I've found it useful. I was trying to configure a GitHub process. Internet searches weren't helpful. So, one evening from home I got into ChatGPT, asked some questions of it in a conversational manner, and got my answers. It was great as I got the answer in 30 minutes, saving me a week of fumbling around.

    Another area where I think Generative AI would help is in trying to predict the spread of diseases. I work in my state's health department. We employ data analysts to run sophisticated algorithms, to predict how various diseases with spread. I'm sure that if we could install a Generative AI in our system, then train it on our data (thus protecting the sensitive nature of the data), that it would help those data analysts in predicting where and how fast some diseases spread. It's sad that we simply cannot do this because someone, somewhere, blocks it for some reason(s) they won't share.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • I was out with a group of 4 friends at the weekend.  All are significantly higher ranking than I am in tech.

    One told a story of using AI to analyse a meeting of senior execs.  It told them that

    • x% of the time was spent talking about the weather
    • y% of time on something else
    • ...etc

    It's summary of the meeting was "There seemed to be no purpose to this meeting".   It's linting for meetings!

     

     

  • That is so funny, David! Thank you for sharing!!

    Kindest Regards, Rod Connect with me on LinkedIn.

  • Rod at work wrote:

    The state department I work for has blocked all generative AI. I don't know why. It might even be blocked for all state departments. I don't have access to that information, so all I can do is speculate as to how widespread blocking of generative AI is.

    We haven't blocked all, but we have some limits on using AI. None of our developers can use AI for code that might be included in a product. Marketing people use it, and we can use Copilot, as you do, which is a glorified search engine. I find the outputs sometimes good, sometimes not.

    I think using AI for predicting and evaluating data is great, but you have to have tight controls on data. If you train the LLM yourself, inside your org, it is great. If you use a third-party service and they might re-use that data, that's bad.

  • I've used AI mostly for things outside of SQL development.  Like git and C# and Powershell.  It seems like it saves time (not always tho) and I go to Google less.  SSC thankfully hasn't instituted any AI policy (afaik) like the other major Q&A technical site.  That policy seems misguided.  It's been a recurring topic in these forums, people use it to help with answers.  I do it too when it's helpful.  Sometimes it's really helpful in describing things going the other way, meaning give AI the code and have it explain what it does.  That's been helpful imo but there are times when it's useless too.  I gave it Jeff's Thanksgiving Day finding code and asked AI to explain it and I didn't find the answer helpful

     

    Aus dem Paradies, das Cantor uns geschaffen, soll uns niemand vertreiben können

  • Steve Collins wrote:

    I've used AI mostly for things outside of SQL development.  Like git and C# and Powershell.  It seems like it saves time (not always tho) and I go to Google less.

    Be interested in examples of what's helped you. I hear people say "time saved", but it's sometimes hard to understand how/where.

    Steve Collins wrote:

    SSC thankfully hasn't instituted any AI policy (afaik) like the other major Q&A technical site.  That policy seems misguided.  It's been a recurring topic in these forums, people use it to help with answers.  I do it too when it's helpful.  Sometimes it's really helpful in describing things going the other way, meaning give AI the code and have it explain what it does.  That's been helpful imo but there are times when it's useless too.  I gave it Jeff's Thanksgiving Day finding code and asked AI to explain it and I didn't find the answer helpful

    I've thought about it, but not sure what to do. I think I want to say no "AI generated articles where an AI has written more than 10%" (20, 30?  not sure). However, I'm not sure what to do in forums. We do see some stuff that looks AI generated as it's not quite on topic, but it's hard to say when it's truly an AI and when it's just someone that didn't real well or doesn't really understand the question.

    A strange world we're in.

Viewing 7 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic. Login to reply