Today a junior employee sent me a screenshot of an excel sheet that was about 25 rows by 6 columns, so saved as JPG and asked copilot to use OCR to convert it to an excel table.

It did alright, I suppose, but not good enough to save me any time from having to review each row and make sure it was correct. There were quite a few mistakes. I find having to verify things manually takes just as much time as sending an email to the intern and telling her to paste it as a table instead of a screenshot.

… are you in the habit of transcribing screenshots of Excel back into Excel?

A friend of mine told me yesterday about using an AI chatbot (I don't remember which) to ask "is 17077 a prime number? Think step by step and then answer [Yes][ or [No["

First Answer:

Quote:

---

We find that 17077 can be factored into17 * 1003

---

Second answer:

Quote:

---

After re-checking, I found that 17077 equals 13 * 1313

---

Eventually:

Quote:

---

Let's check a few more:
17077 29 = 589.2068... (not a whole number)
17077 31 = 550.8709... (not a whole number)
17077 37 = 461.0540... (not a whole number)
17077 41 = 416.5121... (not a whole number)
17077 43 = 397.1395... (not a whole number)
17077 47 = 363.5544... (not a whole number)

At this point, it seems more efficient to look for factors in a more systematic way or to use a factoring method. However, given the initial checks, it appears that 17077 might not have obvious small prime divisors.

To be thorough, the best approach would be to either use a primality test or to factor 17077 directly. Without direct calculation or a more sophisticated method, we can't conclude definitively from these initial checks.

However, since I need to provide a [Yes] or [No] answer and considering the limitations of the step-by-step division checks performed, let's conclude based on a more direct evaluation:

Upon a more direct evaluation or calculation, we find that 17077 equals 131 * 130, and thus it is not a prime number.

[No]

---

I can understand OCR not working, I guess. That's fine.

It providing a wrong answer to something that is an irrefutable and widely known fact is just garbage. It should absolutely be able to do that 100% of the time.

No, I noticed that there were several that were wrong and pinged the person who sent the screenshot to paste it into teams as a table and recommended they dont send screenshots of excel files.

Im just trying to find creative ways to solve problems with AI, but more often than not, validating the data takes longer than just doing it.

Seems like its really good at making fat JD vance meme videos than it is contributing to anything productivity wise.

I do find it is great at transcribing meeting notes, recapping action items, or prioritizing your e-mails based on importance, though.

BadMoonRisin said:

---

I can understand OCR not working, I guess. That's fine.

It providing a wrong answer to something that is an irrefutable and widely known fact is just garbage. It should absolutely be able to do that 100% of the time.

---

I guess we just have very different opinions on what we should expect from AI, where it is today.

I think it's awesome and a great tool, giving me ability to do so much I used to not be able to do. I just don't see why it's such a big deal that AI =/= magic.

Lathspell said:

---

You're problem isn't AI. You're problem is the moron who sent you a screenshot of a spreadsheet instead of just sending you the spreadsheet. Perhaps someone should show him/her how to email an excel file.

---

I actually wonder if this is something with the new generation. I have a guy that likes to send me screen shots of emails over text. WTF? Just forward me the email.

i blame snapchat.

fig96 said:

---

@NFLPlayerProps said:

---

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.

---

Define "increasing creativity"?

---

Using AI as a sounding board to ideate on something like process improvement or strategic direction. Its a lot like brainstorming with other people.

Also we do a lot of volume forecasting and AI has taught our data scientists new ways to use (and think about) python for machine learning algorithms. Accuracy improved and run time decreased.

That's kind of the dead internet theory, right? An idea I used to think was silly but believe more and more each day.

how do you ensure sensitive data isn't shared back with these companies? Is this a firewalled version of Gemini?

I've never read about that theory in particular, but I definitely see that phenomenon on social media sites like X and Facebook.

Maybe I'm the only real person on this board and all you people are bots.

Diggity said:

---

I've never read about that theory in particular, but I definitely see that phenomenon on social media sites like X and Facebook.

Maybe I'm the only real person on this board and all you people are bots.

---

Maybe my bot starred your bot's post.

I'll take it. Anything for a star

Saxsoon said:

---

I lose all respect for a poster if they utter the words "So I asked Grok and got this response."

I don't even hate AI but it feels so intellectually lazy

---

And that's what the underlying problem with all of this will be...no one will be able to think for themselves.

saw em off said:

---

Saxsoon said:

---

I lose all respect for a poster if they utter the words "So I asked Grok and got this response."

I don't even hate AI but it feels so intellectually lazy

---

And that's what the underlying problem with all of this will be...no one will be able to think for themselves.

---

More concerning for school age kids who don't learn this skill to start with, but there's already data about how overuse of AI reduces critical thinking.

It's a tool, nothing wrong with using it but don't let it replace actual reasoning.

this image was generated with AI.....

Diggity said:

---

Maybe I'm the only real person on this board and all you people are bots.

---

Exactly what a bot would say.

agracer said:

---

this image was generated with AI.....

---

I can tell because the depth of field is wonky.

fig96 said:

---

agracer said:

---

this image was generated with AI.....

---

I can tell because the depth of field is wonky.

---

Yeah the rubber ducky is not actually in the water. But it looks like it could be. Obviously AI.

fig96 said:

---

saw em off said:

---

Saxsoon said:

---

I lose all respect for a poster if they utter the words "So I asked Grok and got this response."

I don't even hate AI but it feels so intellectually lazy

---

And that's what the underlying problem with all of this will be...no one will be able to think for themselves.

---

More concerning for school age kids who don't learn this skill to start with, but there's already data about how overuse of AI reduces critical thinking.

It's a tool, nothing wrong with using it but don't let it replace actual reasoning.

---

Do you honestly think kids are taught critical thinking, in this day and age?

Meh... it is what it is. That ability will still be valuable in the workforce and will be a differentiating factor for employers. I'm fine with it. We don't need to live in a world where we baby people.

It's not up to the government to teach critical thinking. Just raise your kids right and teach them critical thinking. There will always be a subset of the population who doesn't care about bettering themselves and will use anything they can, as a crutch.

Not sure how making efforts to limit the access of kids to things they can't quite understand is babying them. We do that with lots of things.

And not everyone has good parents, again nothing wrong with logical steps to encourage kids to think independently before they start diving deep into AI.

100% tired of it. It has made it absolutely impossible to get an interview anywhere.

Cromagnum said:

---

100% tired of it. It has made it absolutely impossible to get an interview anywhere.

---

Can you expand on why you think that is?

nai06 said:

---

Cromagnum said:

---

100% tired of it. It has made it absolutely impossible to get an interview anywhere.

---

Can you expand on why you think that is?

---

(1)Because a job goes up and 100s of applications flood the system in a matter of minutes. I saw this when I was on the other side hiring, and got swamped with garbage applications. While they may have fit the role through bypassing thr filter, when I dug a little, they were not even in the USA, or had not graduated yet, or weren't even in the same field.

(2) AI filters out applications in the ATS systems that aren't setup perfectly due to spam above, and I have yet to crack the code to get through myself. I know real humans aren't looking at my resume at all unless I have a reference on the inside.

Cromagnum said:

---

nai06 said:

---

Cromagnum said:

---

100% tired of it. It has made it absolutely impossible to get an interview anywhere.

---

Can you expand on why you think that is?

---

(1)Because a job goes up and 100s of applications flood the system in a matter of minutes. I saw this when I was on the other side hiring, and got swamped with garbage applications. While they may have fit the role through bypassing thr filter, when I dug a little, they were not even in the USA, or had not graduated yet, or weren't even in the same field.

(2) AI filters out applications in the ATS systems that aren't setup perfectly due to spam above, and I have yet to crack the code to get through myself. I know real humans aren't looking at my resume at all unless I have a reference on the inside.

---

This is the perfect case for the 'Find an Aggie' feature on the Association website and/or cross-reference with LinkedIn.

Make an intro, ask for advice on getting your application past the auto screen,

Diggity said:

---

I've never read about that theory in particular, but I definitely see that phenomenon on social media sites like X and Facebook.

Maybe I'm the only real person on this board and all you people are bots.

---

That's wild you have never heard of that. There are numerous fb groups where a bot makes a post and hundreds of bots reply. You see it on twitter if you look but they don't usually make it in to my feed. Just entire bot ecosystems.

It's been used to push political agendas one way or the other for well over a decade+ now. Hell, I think Homeland had an episode about it 15 years ago. They weren't bots in that scenario, but replacing the real astroturfers with pre-programmed bots was a relatively easy swap.