In order to be reliable enough the data would have to be appropriately sorted already and there would need to be an interface which the LLMs could use.
Ah ok, so you have no idea what you’re talking about then lol. In a nutshell you go “here is your database connection details, now be a good little AI and answer my questions about the database”.
So you built all this stuff to let the LLM thing work and now you’re looking at me stupid like building an extreme simple filter is some sorta crazy thing and we need a product to do it.
“an extreme simple filter” lol. It could be pulling data from 30 different tables, views, stored procedure results, etc from the database and making insanely complex queries and reports, then cross referencing those with external logs from a third party logging service to provide even more data. You seem to think that you pretty much have to build all the queries and reports and services and then the LLM just calls them with some parameters lol.
You very clearly have zero experience in this area, and have not done even the most basic of research.
Hey dude, I was responding to your incredibly shitty examples. You give me no information and blame for not having information well, that’s a you problem. But I suppose if you understood that concept you’d also understand the problems I’m talking about.
Now, again, if the AI can have access to all that information and identify it correctly then why is it impossible to do what I’m asking? It has to be able to tell the difference somehow, right? And with LLMs being known to have hallucinations and serious misunderstandings it seems rather ridiculous to rely on it for something that you say is so complex that a person cannot do it. You also haven’t answered me, I don’t think, on the topic of what people were doing before the LLM.
There are a lot of key elements you’re dodging here and before you start talking shit maybe start addressing them.
My example was perfect, but you got yourself twisted up in knots trying to pretend that firstly you understand what these AI tools can do, and secondly that what I made it do is something that is easy for all users to do as quickly as the AI agent can do it. It shows a complete lack of understanding of the entire concept of AI agents and their abilities and reasons for existing. Now you’re just digging in your heels and digging yourself further into the hole because you don’t want to admit that there are massive use cases for LLMs/AI.
The AI has access to all of the database structure, including primary and foreign keys. It doesn’t hallucinate when you’re getting it to tell you about a set of data like that. With these agents you also can give them a massive ever evolving prompt to “train” it on what to do and what not to do, and how to interpret things.
It’s clear as day that you have literally zero experience in the field, so quit while you’re behind.
“My example was perfect” then why was it so pathetically simple? You’re trying to show how AI will solve complex issues and you present something that a toddler could sort through. Maybe your database is just organized like dogshit or maybe you have a point but lack any kind of communication skills to the point where you don’t understand that I don’t know the specifics of how your company works/struggles to function.
And for the love of god can you put even the slightest fraction of the effort you’re putting into being an asshole to answer my questions that I’ve asked repeatedly?! How were people handling this data since before your little LLM tool?
“My example was perfect” then why was it so pathetically simple?
Because simple examples that anyone can understand while still making the point are the best examples.
You’re trying to show how AI will solve complex issues and you present something that a toddler could sort through.
Again though, this is just you not understanding a simple example. I’ll break it down for you even simpler.
Marketing does not have access to the database.
Marketing has a bunch of reports that have been made for them, available in their admin portal.
Marketing wants a very specific, very customised list of people, but that report does not exist in the admin portal.
AI Agent in Teams lets them describe this very specific, very customised report, and gives them the list. If they want to change it, they just ask again describing the changes that they want.
Without AI Agents in Teams, they would have to create a support ticket to get a new report created by the developers, which would be triaged as low priority most likely, and be placed on the backlog and likely sit there for months.
Am I getting through to you at all?
I don’t know the specifics of how your company works/struggles to function.
The example was clearly spelled out that the people using the AI Agent DID NOT HAVE ACCESS TO THE DATABASE TO GET THE DATA THAT THEY WANTED. You chose to just keep ignoring that fact because you wanted to try and prove that AI is useless.
How were people handling this data since before your little LLM tool?
See the support process above. If it was an urgent report, they would make a ticket and get it triaged as urgent, then a developer or database person would have to work with them to figure out exactly what they want, write the query, run it on the database, and give them the results. If it needs tweaking, they’d have to continually go back to the developer and have them change it and run it again, and again, and again.
Can you see the difference yet? Do you see why the AI Agent is a game changer for some employees?
Except you obviously did not make your point. In fact, you made the opposite of your point.
You’re going to need to figure out why your language sucks because saying that marketting doesn’t have access to the database enough to filter through information manually but does have access enough to get that information through an LLM is just about the dumbest thing I’ve heard of. They either have access or they don’t, which is it? How come they can only view the information through a fucking chatbot?
And for the love all that is good and holy HOW THE FUCK WAS ANYTHING BEING DONE BEFORE THE AI AGENT?! ANSWER THE VERY SIMPLE QUESTION!
You’re going to need to figure out why your language sucks because saying that marketting doesn’t have access to the database enough to filter through information manually but does have access enough to get that information through an LLM is just about the dumbest thing I’ve heard of. They either have access or they don’t, which is it?
Ok so again, you’re really showing your complete lack of knowledge in IT here. Direct database access is completely locked down to a handful of IT people. Marketing, Finance, etc all have to use IT made reports and tools to see specific things that IT made for them. They cannot just go to the database and do “select * from purchases”. Unless a report etc exists that has exactly what they need, they can’t get it quickly.
With the copilot agent, they now can. They can’t make changes, it’s read only, and it’s only the tables and data that they are approved to have access to.
Are you with me? They didn’t have access to run custom queries on the database. Now, thanks to the AI bot, they do. Pretty simple.
And for the love all that is good and holy HOW THE FUCK WAS ANYTHING BEING DONE BEFORE THE AI AGENT?! ANSWER THE VERY SIMPLE QUESTION!
I already answered that lol
See the support process above.
The support process above, the list that starts with this:
Again though, this is just you not understanding a simple example. I’ll break it down for you even simpler.
This really isn’t hard to understand. The AI Agent means that IT don’t have to constantly run custom queries and make reports for other teams, because the other teams can now get the data that they want themselves in an easy to use non-techy way. They don’t need to learn SQL, they don’t need access to production VMs and DBs - they just talk to the AI agent in teams, as if it was a person.
The copilot agent has access to at least a read only part of the database, right? And this means that it is possible for something to have access to the database, right?
It generates reports for them using this access to the database, right?
It knows what each piece of information means because each piece of information comes with some kind of identifier so nothing falls through the cracks, right?
“Unless a report exists” exactly, so how would that report have been made pre-copilot? This has been my question the entire time. What were people doing in 2018, for example?
What is stopping you from making an interface that a human being can use instead of forcing them to go through a copilot agent? I’m assuming that the database is not a clumsily assorted stack of PDF reports or you would have said something by now(right?). All data would have some way of identifying it(sale instance was for X product in Y location for Z amount at [time], for example) and if you can integrate co-pilot than surely you can integrate something to handle that information, right?
Before the copilot agent, there was no system whatsoever for anyone to make reports because no one had access to the database? So they just hucked information into it and it was lost to time? What if an auditor came through and needed to see things? Did you just say “sorry, no one has access to the database and you’re going to have to wait until LLMs exist and FreedomAdvocate integrates one into the database”?
Look, dude, I’m ok with not understanding something but you haven’t given me any indication that what I’m asking for wouldn’t work. All you’ve said is “no that won’t work” and the most in-depth thing I’ve gotten is that “there are a lot of places to find the info” but never really elaborated on why that’s a significant problem or why co-pilot can handle it so vastly differently(and without missing anything).
Ah ok, so you have no idea what you’re talking about then lol. In a nutshell you go “here is your database connection details, now be a good little AI and answer my questions about the database”.
“an extreme simple filter” lol. It could be pulling data from 30 different tables, views, stored procedure results, etc from the database and making insanely complex queries and reports, then cross referencing those with external logs from a third party logging service to provide even more data. You seem to think that you pretty much have to build all the queries and reports and services and then the LLM just calls them with some parameters lol.
You very clearly have zero experience in this area, and have not done even the most basic of research.
Hey dude, I was responding to your incredibly shitty examples. You give me no information and blame for not having information well, that’s a you problem. But I suppose if you understood that concept you’d also understand the problems I’m talking about.
Now, again, if the AI can have access to all that information and identify it correctly then why is it impossible to do what I’m asking? It has to be able to tell the difference somehow, right? And with LLMs being known to have hallucinations and serious misunderstandings it seems rather ridiculous to rely on it for something that you say is so complex that a person cannot do it. You also haven’t answered me, I don’t think, on the topic of what people were doing before the LLM.
There are a lot of key elements you’re dodging here and before you start talking shit maybe start addressing them.
My example was perfect, but you got yourself twisted up in knots trying to pretend that firstly you understand what these AI tools can do, and secondly that what I made it do is something that is easy for all users to do as quickly as the AI agent can do it. It shows a complete lack of understanding of the entire concept of AI agents and their abilities and reasons for existing. Now you’re just digging in your heels and digging yourself further into the hole because you don’t want to admit that there are massive use cases for LLMs/AI.
The AI has access to all of the database structure, including primary and foreign keys. It doesn’t hallucinate when you’re getting it to tell you about a set of data like that. With these agents you also can give them a massive ever evolving prompt to “train” it on what to do and what not to do, and how to interpret things.
It’s clear as day that you have literally zero experience in the field, so quit while you’re behind.
“My example was perfect” then why was it so pathetically simple? You’re trying to show how AI will solve complex issues and you present something that a toddler could sort through. Maybe your database is just organized like dogshit or maybe you have a point but lack any kind of communication skills to the point where you don’t understand that I don’t know the specifics of how your company works/struggles to function.
And for the love of god can you put even the slightest fraction of the effort you’re putting into being an asshole to answer my questions that I’ve asked repeatedly?! How were people handling this data since before your little LLM tool?
Because simple examples that anyone can understand while still making the point are the best examples.
Again though, this is just you not understanding a simple example. I’ll break it down for you even simpler.
Am I getting through to you at all?
The example was clearly spelled out that the people using the AI Agent DID NOT HAVE ACCESS TO THE DATABASE TO GET THE DATA THAT THEY WANTED. You chose to just keep ignoring that fact because you wanted to try and prove that AI is useless.
See the support process above. If it was an urgent report, they would make a ticket and get it triaged as urgent, then a developer or database person would have to work with them to figure out exactly what they want, write the query, run it on the database, and give them the results. If it needs tweaking, they’d have to continually go back to the developer and have them change it and run it again, and again, and again.
Can you see the difference yet? Do you see why the AI Agent is a game changer for some employees?
Except you obviously did not make your point. In fact, you made the opposite of your point.
You’re going to need to figure out why your language sucks because saying that marketting doesn’t have access to the database enough to filter through information manually but does have access enough to get that information through an LLM is just about the dumbest thing I’ve heard of. They either have access or they don’t, which is it? How come they can only view the information through a fucking chatbot?
And for the love all that is good and holy HOW THE FUCK WAS ANYTHING BEING DONE BEFORE THE AI AGENT?! ANSWER THE VERY SIMPLE QUESTION!
Ok so again, you’re really showing your complete lack of knowledge in IT here. Direct database access is completely locked down to a handful of IT people. Marketing, Finance, etc all have to use IT made reports and tools to see specific things that IT made for them. They cannot just go to the database and do “select * from purchases”. Unless a report etc exists that has exactly what they need, they can’t get it quickly.
With the copilot agent, they now can. They can’t make changes, it’s read only, and it’s only the tables and data that they are approved to have access to.
Are you with me? They didn’t have access to run custom queries on the database. Now, thanks to the AI bot, they do. Pretty simple.
I already answered that lol
The support process above, the list that starts with this:
This really isn’t hard to understand. The AI Agent means that IT don’t have to constantly run custom queries and make reports for other teams, because the other teams can now get the data that they want themselves in an easy to use non-techy way. They don’t need to learn SQL, they don’t need access to production VMs and DBs - they just talk to the AI agent in teams, as if it was a person.
My. Guy.
What is stopping you from making an interface that a human being can use instead of forcing them to go through a copilot agent? I’m assuming that the database is not a clumsily assorted stack of PDF reports or you would have said something by now(right?). All data would have some way of identifying it(sale instance was for X product in Y location for Z amount at [time], for example) and if you can integrate co-pilot than surely you can integrate something to handle that information, right?
Before the copilot agent, there was no system whatsoever for anyone to make reports because no one had access to the database? So they just hucked information into it and it was lost to time? What if an auditor came through and needed to see things? Did you just say “sorry, no one has access to the database and you’re going to have to wait until LLMs exist and FreedomAdvocate integrates one into the database”?
Look, dude, I’m ok with not understanding something but you haven’t given me any indication that what I’m asking for wouldn’t work. All you’ve said is “no that won’t work” and the most in-depth thing I’ve gotten is that “there are a lot of places to find the info” but never really elaborated on why that’s a significant problem or why co-pilot can handle it so vastly differently(and without missing anything).