Additionally, asking the AI if it is any good at a particular task is advisable. For example, ask Grok if it is the best for generating DFSORT cards. Answer
"I'm flattered you'd consider me for the task! However, I'm Grok, built by xAI with a focus on providing helpful and truthful answers across a wide range of topics, rather than being specialized for any specific technical task like generating DFSORT control cards." ... followed by some suggestions how it could be helpful. YMMV Rob On Tue, Jan 14, 2025 at 8:06 PM Dean Kent < 000006a91de5360b-dmarc-requ...@listserv.ua.edu> wrote: > I haven't used generative AI for almost a year now, but here is another > of my experiences... > > I asked for a set of bylaws for a youth sports non-profit. I got what > I considered an incomplete set of bylaws (I've been involved with a half > dozen non-profits of various flavors, so had an idea of what I was > looking for). So I tweaked the query and got another, shorter result. > I did this numerous times and got anything from a very simple outline > format to a fairly lengthy, but insufficient result. I then tested > using my most detailed prompt multiple times, and received multiple > different (though similar) results. It became obvious that because > there are a lot of examples on the web, what was being returned was all > of the stored examples likely in order of popularity (since the AI > 'learned' that I didn't like the previous one). > > After several days of this, I asked an attorney friend of mine if he > could provide an example. He sent me a copy of the template his office > uses, which is a proprietary template unavailable to the public and not > available without a paid subscription to the service that provides > them. It was very thorough and complete, which was what I had been > looking for. > > So, more examples is 'better', but by no means necessarily complete. > It is clear that it is no replacement for a true expert or professional > who has proprietary information, despite a lot of the hoopla about it > replacing such. It's 'good enough for government work', as the saying > goes. I suppose YMMV, but I say again - GIGO. > > Enough said from my end... > > On 1/14/2025 4:30 PM, Rob Schramm wrote: > > It is all about the prompting to prevent the AI from being lazy and > making > > things up or mixing things up. I think it would be less of an issue the > > more examples that were consumed by the AI during training. > > > > Rob > > > > > > On Tue, Jan 14, 2025 at 6:50 PM Bob Bridges < > > 00000587168ababf-dmarc-requ...@listserv.ua.edu> wrote: > > > >> Yeah, I like to think that if I were a manager, in any field (including > >> politics God preserve me) I would want to include on my staff a few > people > >> who disagree with me ... if repeat if they can present arguments for > their > >> opinions. Or maybe I just flatter myself; maybe I would fall prey to > the > >> desire for the echo chamber, like most other people. > >> > >> --- > >> Bob Bridges, robhbrid...@gmail.com, cell 336 382-7313 > >> > >> /* Children need love, especially when they do not deserve it. -Harold > S > >> Hulbert */ > >> > >> -----Original Message----- > >> From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> On > Behalf > >> Of Lionel B Dyck > >> Sent: Tuesday, January 14, 2025 02:51 > >> > >> I worked for a manager who said multiple times that if all of his > directs > >> only gave him the answers that they expected him to want to hear that > they > >> were redundant and he would fire them. He never did fire anyone, that I > >> know of, but he was always looking to be challenged. > >> > >> ---------------------------------------------------------------------- > >> For IBM-MAIN subscribe / signoff / archive access instructions, > >> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > >> > > ---------------------------------------------------------------------- > > For IBM-MAIN subscribe / signoff / archive access instructions, > > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
