August 26, 2024 at 12:00 am
Comments posted to this topic are about the item Big Northwind with AI
August 26, 2024 at 6:10 am
Thanks for showcasing the reality of the accuracy and usefulness of AI. I wish more such posts come into mainstream and reduce the hype caused by the linkedin posts, youtube videos, and other online content.
August 26, 2024 at 1:58 pm
I have similar experience, ChatGPT generated T-SQL code to calculate local min and max nodes from a function that was defined by set of values in a given table. As expected code had some syntax errors and was not runnable in SSMS.
When I pointed out in dialog that code is not working and requires debugging the next version was ok. I got a slight sensation that instead of ChatGPT I was talking to a “mechanical Turk” person who has access to SSMS and managed to debug code. It like old Russian proverb - do not trust and always double and triple check!
August 26, 2024 at 6:33 pm
I've found ChatGPT to be most useful when asked to summarize the tools and steps involved in implementing a process.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
August 26, 2024 at 6:49 pm
To date, I'm the least impressed with BING AutoPilot and Claude 3.5 Sonnet. I'm especially disappointed in Claude 3.5 Sonnet because it doesn't seem to provide references like the others do.
I've found ChatGPT 4o to be better but only a little.
I like the "language" that Perplexity.AI uses and how well it explains things. It also seems to me that it comes up with better answers and it tries to avoid RBAR answers, in a lot of cases (and explains that it has avoided the RBAR answer).
They all still have a "suck" factor, though. Although Perplexity.AI seems to dig out the "better" answers, NONE of them have correctly answered the following question correctly nor similar questions. The reason why is because of the overwhelming number of people, including large numbers of supposed "experts", that actually don't get the answer right either.
Here's the question and a copy of one of the answers that they all boil down to. The answer is patently incorrect but, doing a normal Google search, you'll find out why it came up with that answer.
--Jeff Moden
Change is inevitable... Change for the better is not.
August 26, 2024 at 7:35 pm
To date, I'm the least impressed with BING AutoPilot and Claude 3.5 Sonnet. I'm especially disappointed in Claude 3.5 Sonnet because it doesn't seem to provide references like the others do.
I've found ChatGPT 4o to be better but only a little.
I like the "language" that Perplexity.AI uses and how well it explains things. It also seems to me that it comes up with better answers and it tries to avoid RBAR answers, in a lot of cases (and explains that it has avoided the RBAR answer).
They all still have a "suck" factor, though. Although Perplexity.AI seems to dig out the "better" answers, NONE of them have correctly answered the following question correctly nor similar questions. The reason why is because of the overwhelming number of people, including large numbers of supposed "experts", that actually don't get the answer right either.
Here's the question and a copy of one of the answers that they all boil down to. The answer is patently incorrect but, doing a normal Google search, you'll find out why it came up with that answer.
What is the correct answer?
If you tell me I'll see if I can drag it out of one of the AI sites with an improved prompt.
August 26, 2024 at 8:16 pm
What is the correct answer?
If you tell me I'll see if I can drag it out of one of the AI sites with an improved prompt.
I'd like to hold off a bit on that because I'm writing an article about it.
--Jeff Moden
Change is inevitable... Change for the better is not.
November 21, 2024 at 6:41 pm
AI is the future
November 22, 2024 at 1:02 pm
This was removed by the editor as SPAM
Viewing 9 posts - 1 through 8 (of 8 total)
You must be logged in to reply to this topic. Login to reply