I don't think it's going away. I do however doubt how useful it will be moving forward.

nai06 said:

---

I don't think it's going away. I do however doubt how useful it will be moving forward.

---

Based on what expertise?

Lathspell said:

---

Lol... if you think AI is going away, you live in a fairytale world.

All enterprises are slowly adopting it in all aspects of business. Every CIO I talk to these days has their own AI protocol and is looking to further implement effectively. Most companies, especially in Houston, will slowly adopt it, but if you are a company who refuses to do so, you will eventually fall behind.

---

I didn't say it was going away. I just think all the hype and investment is going to burst.

Everyone said "not a chance" in 1999 and we all remember what happened.

Lathspell said:

---

nai06 said:

---

I don't think it's going away. I do however doubt how useful it will be moving forward.

---

Based on what expertise?

---

Just experience with same as everyone else and follow the news and evaluations of people that work with it in ways different than I do.


I'm not convinced it's the magic bullet people make it out to be. AI has some major problems that no one seems able or willing to address

AI is going to be a very big deal. Monumental shift, both for those who understand it and use it to their advantage and the laggards who discount/ignore it. Just in different directions.

@NFLPlayerProps said:

---

AI is going to be a very big deal. Monumental shift, both for those who understand it and use it to their advantage and the laggards who discount/ignore it. Just in different directions.

---

Ok in what ways? Be specific, what is this going to change?

So far I am using it to increase productivity by automating repetitive tasks and increasing creativity, communication, and employee engagement/development. Also using it to decrease my rate of making mistakes from 1-2 a year to non-existent (so far).

I've also used it to enhance my options trading, specifically selling 30-40 DTE covered calls and balancing the Greeks to roll at the best possible time. About to turn it loose on the family budget, but I'm not sure I want my wife to know how much I'm spending on my M3 and 4Runner

The biggest thing I'm tired of is the sentence structure it uses. I can pick out an AI scripted post immediately at this point. AI posts all sound the same regardless of the topic.

And it's very tiresome because posts are becoming commoditized into using the same sentence structure on everything.

@NFLPlayerProps said:

---

So far I am using it to increase productivity by automating repetitive tasks and increasing creativity, communication, and employee engagement/development. Also using it to decrease my rate of making mistakes from 1-2 a year to non-existent (so far).

I've also used it to enhance my options trading, specifically selling 30-40 DTE covered calls and balancing the Greeks to roll at the best possible time. About to turn it loose on the family budget, but I'm not sure I want my wife to know how much I'm spending on my M3 and 4Runner

---

Do you think it does a good job of this?

I ask because I see AI generated communication and creative writing on a daily basis and it's always dog *****

Maybe don't assess it based upon its ability to write interesting stories and rather on its ability to create a paragraph on technical explanations/completing discrete tasks. Don't think does it do a good job today, but rather what it does relative to what it did a year ago or two years ago. How does it handle data? Coding?

People look at it now and see how well it writes poetry, haikus, or something similar. They don't look at it from a replication of human emotional capacity, but rather a human technical/logical capacity.

I just got laid off from my job. I got a severance package and am agreement tied to it. The agreement said consult an attorney for review (I am not at a level that needed that kind of oversight) and my attorney brother said have AI review it. The output from the AI spat out the same concerns about the agreement I would have if I were in a technical role related to my job in terms of concerns about the exit requirements. It wasn't relevant to me specifically, but would be relevant if I had more proprietary aspects of my job. Think about what a computer system would have said 5 years ago relative to what it says today. Extrapolate that to 5 years from now.

My dad just got diagnosed with early stages of Alzheimer's. My mom and I used AI to describe to protocol of the medication a neurologist suggested for his treatment. After an MRI showed a resulting minor stroke from one monitoring MRI to the next and the implications of it without having to wait on a discussion with the neurologist to understand the implications of it. It is useful today to provide expanded understanding they could acquire with hours of googling in a contracted time.

I got diagnosed with rectal cancer in January 2024 and it has helped explain treatments, risks, results and errors on reports from imaging as well as standard protocols/therapies after the fact in line with what the doctor confirmed before I had the chance to discuss it with them. Those are technical, discrete situations that are branching answers to questions with accurate answers, correcting radiologist assessments of scans. Radiologist report said I had a portion of my liver removed, I did not. AI's error, humans do as well.

It's like people looking at the internet in 1998 and saying "what does it actually do to impact society?" And in that form it was "not much" but fast forward 10 years at it become a large portion and ever growing portion of the economy. Everything is impacted by the internet now. Everything will be impacted by AI as it improves over time. I can today use it to expand my abilities to complete tasks, it won't do it flawlessly, but with it I can improve my own skills by asking it for help and fixing where it fails.

People underestimate the rate of improvement of technology over time. Five years ago a Tesla would make absolutely stupid decisions in driving scenarios and would scare the **** out of an average person. Today it drives me everywhere, point to point with extremely rare. Basically always navigational corrections and not safety issues. It's very early.

ETA: All of the medical information was later validated by my oncologist, colorectal surgeon, dermatologist (had maculopapular eruption from chemo treatments, easily the second worst consequence of cancer treatment behind radiation burns/digestive issues from lower GI radiation), and neurologist (good friend is a high level neurologist). I wouldn't rely solely on AI yet, but the accuracy of information has been within reasonable alignment of their interpretations of scans/symptoms verified by doctors.

nai06 said:

---

@NFLPlayerProps said:

---

So far I am using it to increase productivity by automating repetitive tasks and increasing creativity, communication, and employee engagement/development. Also using it to decrease my rate of making mistakes from 1-2 a year to non-existent (so far).

I've also used it to enhance my options trading, specifically selling 30-40 DTE covered calls and balancing the Greeks to roll at the best possible time. About to turn it loose on the family budget, but I'm not sure I want my wife to know how much I'm spending on my M3 and 4Runner

---

Do you think it does a good job of this?

I ask because I see AI generated communication and creative writing on a daily basis and it's always dog *****

---

Back to my point, it all sounds exactly the same. AI consistently uses the same overall sentence structure and organization. I can immediately point it out every time if it uses its default structure. I think it's great for technical writing (eg job description) and tweaking your own comms for clean up and conciseness and have used it for such.

But it stands out like a sore thumb for these current event, political, creative, making a declarative point type of posts you see on social media or LinkedIn. Immediate turn off for me.

That's a tough run. Sorry you are going through all of that.

I generally agree with you on AI though. It's going to get better. It's already been a game changer for me in just making spreadsheets better which I like to dabble in. It's allowed me to do things I did not have the skill set to accomplish on my own without spending money on a course. Instead, I spend an hour or two just prompting AI and tweaking the results until I get the outcome I'm looking for. So, it works great for Excel spreadsheets and coding that are rules based and have a defined path to follow.

I also used it for healthcare recently too though and it was highly helpful. This doesn't sound like a big deal, but I used it to build me a pain medication schedule based on my diet and activity post surgery last month. And then I just tweaked it as needed as I progressed through recovery. I knew the parameters I needed to stay in based on doctor's rules but the doc doesn't always give you a solid plan to follow.

It still has huge issues when you are using it for basic internet search because it combs the web and will pull back whatever it finds. No rules there at all - just an information spigot.

nai06 said:

---

@NFLPlayerProps said:

---

So far I am using it to increase productivity by automating repetitive tasks and increasing creativity, communication, and employee engagement/development. Also using it to decrease my rate of making mistakes from 1-2 a year to non-existent (so far).

I've also used it to enhance my options trading, specifically selling 30-40 DTE covered calls and balancing the Greeks to roll at the best possible time. About to turn it loose on the family budget, but I'm not sure I want my wife to know how much I'm spending on my M3 and 4Runner

---

Do you think it does a good job of this?

I ask because I see AI generated communication and creative writing on a daily basis and it's always dog *****

---

I do, and that's why I use it. I don't usually implement tools that I think do a bad job.

I think the biggest issue with AI is in its marketing. There are too many lay-persons who seem to think it should be like talking to a god. We are still years away from that form of AI.

It's simply a tool you have to learn to use, just like anything. It's uses can be as simple or as involved as learning a program like Excel. Sure, most people just make basic tables. But those who truly know how to use Excel can do some absolutely amazing things with it.

Right now, I feel like i'm listening to people who make small tables in excel complaining it's not that useful, whereas there are a couple on this thread who are making pivot tables and creating macros to solve all kinds of business issues.

The issue isn't Excel, it's the end-user.

It's a ******ed chatbot, basically, that contradicts itself with every prompt.

The link it provides to UT Athletics, as the reference to it's answer, clearly says that we lost.



Fool's gold. Generative AI is worse today than google search was 2 decades ago.

What dataset is the AI studying that gets something so easily provable, wrong? We will never know. It's a black box. Trust me bro.

And that's the danger. Where did it come up with this answer when the result was binary and we 100% lost?

If it cant get basic facts correct, that are binary and empirically provable, all of its output should be considered complete garbage.

Grok correctly answered January 23, 2002. You get what you pay for.

As I've already posted, and everyone should know by now, Gemini is the worst LLM out there because it is trained on Google's bull***** Google doesn't care about truth or facts.

It probably told you that because it recognized you're an Aggie and it didn't want to perpetrate a micro-aggression against you. So it made up a lie to make you feel better.

I'm not even simply making a joke. Gemini is trained on hyper-leftist ideology because Google builds that into their algorithm.

Lathspell said:

---

As I've already posted, and everyone should know by now, Gemini is the worst LLM out there because it is trained on Google's bull***** Google doesn't care about truth or facts.

It probably told you that because it recognized you're an Aggie and it didn't want to perpetrate a micro-aggression against you. So it made up a lie to make you feel better.

I'm not even simply making a joke. Gemini is trained on hyper-leftist ideology because Google builds that into their algorithm.

---

Fair, but that's also the crux of the point -- where in the LLM did Google "learn" to make up a patently false premise? It didn't learn from reddit users who insisted that we won by 10 points instead of lost by 10 points. It didnt "learn" anything.

I guess that's also your point, which is correct. Gemini is willing to LIE to make its users feel "good" or "better".

Generative AI, as a source, predicates and guesses "the next best word" based on a supposedly infinite "input". None of these inputs would have had the aggies winning the game in December of 2019. So, where did it look to achieve it's output that the Aggies did win 60-50?

Gemini does suck donkey dick, as referenced in this thread I started awhile ago: Here: https://texags.com/forums/30/topics/3518332

There should be zero sources for the Aggies winning that game...so...where did the output come from?

Shut it down, it's slop output.

Agreed... Gemini is trash. I don't understand your point on continuing to yell at the clouds about it.

I've said it. You've said it. No one has contradicted either of us regarding the statement: Gemini is trash. However, there were two other posters who cited two separate other LLM's that answered correctly.

I use AI daily, and though it's definitely not perfect, I don't think it's as bad as some of you seem to make it out to be. But if you feel that way, the fine. Don't use it, lol. I truly don't understand your overall point.

I use co-pilot daily and I find that it takes more time to verify that the output is legitimate than it would take me to do the same myself, manually.

It's great at meeting recaps, though, so it's not all bad.

But I still dont see the benefit if it patently lies to you for no reason. The truth should always be the truth. And if AI is too dumbassed to figure that out, it is completely worthless.

hph6203 said:

---

Grok correctly answered January 23, 2002. You get what you pay for.

---

Well, I guess Grok is wrong now, since the date is now 1/17/26.

You'll probably roll your eyes, but copilot is nearly as bad as Gemini.

Everything Microsoft releases is basically ***** They just give it away so everyone uses it. Explain to men how a company that develops EVERY PIECE of software they do, including having to work on their video and calling platforms to compete with the leaders in those spaces, also has the resources to compete with the best LLM's?

Quick answer for you: they can't.

Everything Microsoft releases is a worse product than you can get from other providers. Case and point: all the people who use Teams for video and calling.

Teams sucks. The other 3 top UCaaS providers in the market are infinitely better. But why do so many still use Teams? Because it has always been included in M365. Not because it's the best but because it's the most convenient. That's just one of their branches where they suck, and I could do that same breakdown with everything else they offer.

The only reason people will ever use copilot is because it will be directly built into their environment, giving it access to an organizations entire Microsoft ecosystem. However, I'm sure they'll have to open that up eventually.

Lathspell said:

---

You'll probably roll your eyes, but copilot is nearly as bad as Gemini.

Everything Microsoft releases is basically ***** They just give it away so everyone uses it. Explain to men how a company that develops EVERY PIECE of software they do, including having to work on their video and calling platforms to compete with the leaders in those spaces, also has the resources to compete with the best LLM's?

Quick answer for you: they can't.

Everything Microsoft releases is a worse product than you can get from other providers. Case and point: all the people who use Teams for video and calling.

Teams sucks. The other 3 top UCaaS providers in the market are infinitely better. But why do so many still use Teams? Because it has always been included in M365. Not because it's the best but because it's the most convenient. That's just one of their branches where they suck, and I could do that same breakdown with everything else they offer.

The only reason people will ever use copilot is because it will be directly built into their environment, giving it access to an organizations entire Microsoft ecosystem. However, I'm sure they'll have to open that up eventually.

---

I don't roll my eyes at all, I'm just calling it out.

It doesn't matter what LLM is being trained, AI is basically a ******ed chatbot.

The danger is that people are treating it as some sort of other-worldly wisdom.

Lathspell said:

---

Agreed... Gemini is trash. I don't understand your point on continuing to yell at the clouds about it.

I've said it. You've said it. No one has contradicted either of us regarding the statement: Gemini is trash. However, there were two other posters who cited two separate other LLM's that answered correctly.

I use AI daily, and though it's definitely not perfect, I don't think it's as bad as some of you seem to make it out to be. But if you feel that way, the fine. Don't use it, lol. I truly don't understand your overall point.

---

Quote:

---

Agreed... Gemini is trash. I don't understand your point on continuing to yell at the clouds about it.

---

Umm, that's the premise of the original post, so....

Your solution seems to be "oh, Microsoft and Google suck at AI" ...and then nothing else.

It's all ****, my dude.

Lathspell said:

---

It probably told you that because it recognized you're an Aggie and it didn't want to perpetrate a micro-aggression against you. So it made up a lie to make you feel better.

---

Uhhh, yeah. I know, that's the point.

Gemini just correctly responded with today's game and the 2002 game as the previous win in Austin and mentioned the last meeting in Austin as a 70-69 loss.

Don't think they're embedding their best stuff in Google search.

hph6203 said:

---

Gemini just correctly responded with today's game and the 2002 game as the previous win in Austin and mentioned the last meeting in Austin as a 70-69 loss.

Don't think they're embedding their best stuff in Google search.

---

That's all fine and dandy, but it responded incorrectly a few hours prior with completely made up bull*****

This is a problem.

It's probably a different, lesser model, it's like saying computers suck, because a Commodore 64 can't run Photoshop. LLMs are expensive to run, more expensive than a Google search so Google running a lesser model to add some incremental functionality that sometimes works and sometimes doesn't isn't indicative of the capabilities of LLMs as a whole.

gemini.google.com if you want a variable LLM that can adapt to the query. I'll also point out that your phrasing in your search will return relevant webpages that you can comb through for the answer, but it's not great when trying to get an accurate answer out of an LLM. Same with the integrated Copilot products. They're not putting their best stuff out there for free.

It doesn't take some sort of hyper-intelligence to state facts, or it shouldn't, anyway.

If an LLM produces an absolutely incorrect value, which is what I demonstrated, it should never be trusted for anything, ever.

A failure humans commit all the time.

So, I saw ChatGPT is going to start using ads. They say they won't influence the answers ChatGPT gives, but I wonder how long Sam Altman will be able to resist putting ads directly into the results. I don't think it's going to be that long. I think OpenAI could be in for a lot of trouble, seeing how their competitors (mainly Google) can sustain losses much longer. Also, it seems like switching out models is fairly trivial, so they are going to have a hard time raising their prices to cover their costs. It's more of a topic for the business board, but I'm betting Microsoft ends up buying them out right eventually.