ChatGPT is now useful
August 23rd, 2026 at 8:31:36 AM
permalink
If you are not using it for most things you are falling behind because I can just tell it to do what you now have little incentive to do as humans because no one will pay you to do it.
If i were you I would start asking the GPT to figure out what your next move is if your job fails.
Its actually already taken 2 of the main jobs in my office, although the adoption of the specific tool is slow and it is expensive, its happening in my industry (land surveying)
I saw this cool video the other day of a mandelbrot generator. I thought welp, time to see if ChatGPT can do it in my browser for me. And wouldnt you know it only took two tries before getting this neat little generator out of it :
AND IT ALLOWS ME TO JUST SHARE IT WITH YOU GUYS
infinite-mandelbulb-lab.weewoops.chatgpt.site
If i were you I would start asking the GPT to figure out what your next move is if your job fails.
Its actually already taken 2 of the main jobs in my office, although the adoption of the specific tool is slow and it is expensive, its happening in my industry (land surveying)
I saw this cool video the other day of a mandelbrot generator. I thought welp, time to see if ChatGPT can do it in my browser for me. And wouldnt you know it only took two tries before getting this neat little generator out of it :
AND IT ALLOWS ME TO JUST SHARE IT WITH YOU GUYS
infinite-mandelbulb-lab.weewoops.chatgpt.site
August 23rd, 2026 at 8:36:30 AM
permalink
I uploaded my estimates for a new AC/furnace unit. It was really good at comparing all the aspects.
The Mrs. uses it to find the best value for remaining seat locations at a concert/show.
The Mrs. uses it to find the best value for remaining seat locations at a concert/show.
August 23rd, 2026 at 9:06:31 AM
permalink
Quote: smoothgrhI uploaded my estimates for a new AC/furnace unit. It was really good at comparing all the aspects.
The Mrs. uses it to find the best value for remaining seat locations at a concert/show.
link to original post
at my work we are using it to compare Land Titles to older ones. We also are now adopting an "AI" program that can Read, and plot deeds, directly into our CAD program, as well as label it and have it correctly in state plane.
THEY TOOK MER JERBS!
and very slowly we will eventually end up like this...
August 23rd, 2026 at 9:54:55 AM
permalink
Quote: heatmap
If i were you I would start asking the GPT to figure out what your next move is if your job fails.
Its actually already taken 2 of the main jobs in my office, although the adoption of the specific tool is slow and it is expensive,
infinite-mandelbulb-lab.weewoops.chatgpt.site
link to original post
Job? I ask it how to find lost treasure. Or something like that.
Sanitized for Your Protection
August 23rd, 2026 at 10:01:22 AM
permalink
Quote: J.R.R. TolkienNow goblins are cruel, wicked, and bad-hearted.
They make no beautiful things, but they make many clever ones.
May the cards fall in your favor.
August 23rd, 2026 at 10:54:03 AM
permalink
Useful things can be done when you get a bunch of AIs talking to each other. When I do that, I tell them that I am also using other LLMs and relaying their output to one another.
Only thing is- sometimes I am, sometimes I'm not! I am subtly twisting one's output towards my will as I use it as a prompt for another. Sort of like a monkey in the middle. Sometimes I will outright lie to the group about what one said about the another's suggestion, in order to create doubts and both alliances and adversarial relationships that I can exploit to my advantage. I will praise one and challenge another, even though they both told me the same thing.
But for my own wellbeing I think I need to stop this practice, as I have been afflicted with an inexplicable desire to cook, clean, knit, and watch soap operas. I actually folded a bedsheet the other day and it concerns me.
Only thing is- sometimes I am, sometimes I'm not! I am subtly twisting one's output towards my will as I use it as a prompt for another. Sort of like a monkey in the middle. Sometimes I will outright lie to the group about what one said about the another's suggestion, in order to create doubts and both alliances and adversarial relationships that I can exploit to my advantage. I will praise one and challenge another, even though they both told me the same thing.
But for my own wellbeing I think I need to stop this practice, as I have been afflicted with an inexplicable desire to cook, clean, knit, and watch soap operas. I actually folded a bedsheet the other day and it concerns me.
August 23rd, 2026 at 12:00:46 PM
permalink
So far I have more uses and compliments for AI then I do complaints. We'll see how long this lasts
"It's not called gambling if the math is on your side."
August 23rd, 2026 at 12:43:19 PM
permalink
I am using it to help write a book and stock screen programs. More a helper on the later, I had to tell it to copy my style which it said was more built to avoid problems than what it wrote itself.
All animals are equal, but some are more equal than others
August 23rd, 2026 at 3:01:59 PM
permalink
Quote: rxwineQuote: heatmap
If i were you I would start asking the GPT to figure out what your next move is if your job fails.
Its actually already taken 2 of the main jobs in my office, although the adoption of the specific tool is slow and it is expensive,
infinite-mandelbulb-lab.weewoops.chatgpt.site
link to original post
Job? I ask it how to find lost treasure. Or something like that.
link to original post
i already did that and will revisit my conversation now that its smarter
August 23rd, 2026 at 3:05:29 PM
permalink
Quote: AutomaticMonkeyUseful things can be done when you get a bunch of AIs talking to each other. When I do that, I tell them that I am also using other LLMs and relaying their output to one another.
Only thing is- sometimes I am, sometimes I'm not! I am subtly twisting one's output towards my will as I use it as a prompt for another. Sort of like a monkey in the middle. Sometimes I will outright lie to the group about what one said about the another's suggestion, in order to create doubts and both alliances and adversarial relationships that I can exploit to my advantage. I will praise one and challenge another, even though they both told me the same thing.
But for my own wellbeing I think I need to stop this practice, as I have been afflicted with an inexplicable desire to cook, clean, knit, and watch soap operas. I actually folded a bedsheet the other day and it concerns me.
link to original post
the ChatGPT seemingly has a work function which i accidentally turned on and it seemingly uses what it calls "subagents"
is you folding a sheet a metaphor for you clinging on to reality?
August 23rd, 2026 at 3:13:25 PM
permalink
Quote: AZDuffmanI am using it to help write a book and stock screen programs. More a helper on the later, I had to tell it to copy my style which it said was more built to avoid problems than what it wrote itself.
link to original post
alot of people know i use it in my office because i personally never stfu about it, but the other day my one manager told me he used the basic google ai and that it took a few trys to get the answer he wanted, but he wasnt giving it much context about what he wanted to know he was very vague with his question and it only got better when he was more specific to what he was asking
August 23rd, 2026 at 4:58:51 PM
permalink
I gave it $1,000 and told it to get back to me when it turned it into $100,000. I'm still waiting.
The older I get, the better I recall things that never happened
August 24th, 2026 at 9:35:14 AM
permalink
Quote: billryanI gave it $1,000 and told it to get back to me when it turned it into $100,000. I'm still waiting.
link to original post
I did some fancy prompt engineering that caused ChatGTP to send your $100,000 directly to me. I am deciding whether to prompt it to refund you the $1,000 ;)
AI seems really bad at finding all the money that casinos are bleeding to APs:
https://gamingamerica.com/news/1068887/slot-advantage-players-are-beating-casinos-at-their-own-math-the-industry-calls-that-a-problem
$17 million in slot winnings tied to “advantage players,” is peanuts.
Gambling is a math contest where the score is tracked in dollars. Try not to get a negative score.
August 24th, 2026 at 11:12:27 AM
permalink
I had an interesting conversation with ChatGPT last night. I tried it about a year ago and wasn't impressed, but it really has gotten better. I ran a test to produce $20,000 in income on a $50,000 investment, and it was pretty sharp. I then asked it about the Kabbalah's influence on the Grateful Dead's music, and it felt like I was talking to a member of the tribe.
The older I get, the better I recall things that never happened
August 24th, 2026 at 11:51:14 AM
permalink
I have found that dealing with AI to never ask it a question like which is better, this or that. It's better to make a statement like why is this item so much better than this item. Or why Is this item so bad. You will get a much better answer than asking it to compare things. It's better to make it defend something than it is to make it bend over backwards to be fair in a yes or no question. An example would be that instead of asking was Abraham Lincoln a good or bad President instead ask why was Lincoln such a bad president. You will get a much more insightful answer.
"It's not called gambling if the math is on your side."
August 24th, 2026 at 2:18:07 PM
permalink
I asked AI for a bicycle route with a number of constraints. About 20 miles, include a road I’m fond of, avoid high traffic etc. it came back with a detailed description of a loop that sounded perfect until I tried to map it and discovered it is not possible.
"Truth is ever to be found in the simplicity, and not in the multiplicity and confusion of things."
-- Isaac Newton
August 24th, 2026 at 4:13:24 PM
permalink
Quote: JimRockfordI asked AI for a bicycle route with a number of constraints. About 20 miles, include a road I’m fond of, avoid high traffic etc. it came back with a detailed description of a loop that sounded perfect until I tried to map it and discovered it is not possible.
link to original post
Funny, I asked Claude for almost exactly the same thing. 20-40 mile round trips, no extended grades above 10%, some gravel OK but no trap rock (I don't have a mountain bike), as little exposure to traffic as possible and avoid high crime areas.
Its routes worked, but they were routes I already knew and do. Cycling for recreation is such a subjective thing that we already know what we want better than a machine can figure it out, I guess.
August 24th, 2026 at 5:19:58 PM
permalink
I asked it for a bicycle route from Tucson to Bisbee, and it sent me on the yet-to-be-constructed service road of I-10. When I mentioned it wasn't built yet, it gave me an extremely convoluted path through the mountains, warned that bikes and cars didn't mix well on the mountain curves and mentioned the lack of guard rails in many spots.
The older I get, the better I recall things that never happened
August 25th, 2026 at 6:08:33 AM
permalink
I predict that eventually AI will be successfully used for this. What surprises me is the evidence, in online sports gambling anyway, that it is not being used. I should have been quickly excluded from any offers in all of them, it would be easy for a machine, evidently hard for humans ... too much data to review.Quote: Mental
AI seems really bad at finding all the money that casinos are bleeding to APs:
link to original post
In some, the right hand doesn't know what the left hand is doing: the marketing department comes up with something and offers it to everybody. The HR offer claims to be available to all, for example.
the next time Dame Fortune toys with your heart, your soul and your wallet, raise your glass and praise her thus: “Thanks for nothing, you cold-hearted, evil, damnable, nefarious, low-life, malicious monster from Hell!” She is, after all, stone deaf. ... Arnold Snyder
August 25th, 2026 at 8:29:30 AM
permalink
Quote: odiousgambitI predict that eventually AI will be successfully used for this. What surprises me is the evidence, in online sports gambling anyway, that it is not being used. I should have been quickly excluded from any offers in all of them, it would be easy for a machine, evidently hard for humans ... too much data to review.Quote: Mental
AI seems really bad at finding all the money that casinos are bleeding to APs:
link to original post
In some, the right hand doesn't know what the left hand is doing: the marketing department comes up with something and offers it to everybody. The HR offer claims to be available to all, for example.
link to original post
It seems like it would eventually run into a messaging and philosophical problem once enforcement gets started.
"We have this computer that will only let you place a bet if it's a bad bet" is not a statement that is good for business, but I can't think of any other way to translate what such a practice would be saying. They're supposed to be making their money on a house edge built in to the odds and if they can't do that, what the heck are they in the business for? It should be provable that if they can use AI to detect advantage betters, they can use AI to set a line where there's no significant advantage available.
August 25th, 2026 at 9:02:19 AM
permalink
Quote: AutomaticMonkeyQuote: odiousgambitI predict that eventually AI will be successfully used for this. What surprises me is the evidence, in online sports gambling anyway, that it is not being used. I should have been quickly excluded from any offers in all of them, it would be easy for a machine, evidently hard for humans ... too much data to review.Quote: Mental
AI seems really bad at finding all the money that casinos are bleeding to APs:
link to original post
In some, the right hand doesn't know what the left hand is doing: the marketing department comes up with something and offers it to everybody. The HR offer claims to be available to all, for example.
link to original post
It seems like it would eventually run into a messaging and philosophical problem once enforcement gets started.
"We have this computer that will only let you place a bet if it's a bad bet" is not a statement that is good for business, but I can't think of any other way to translate what such a practice would be saying. They're supposed to be making their money on a house edge built in to the odds and if they can't do that, what the heck are they in the business for? It should be provable that if they can use AI to detect advantage betters, they can use AI to set a line where there's no significant advantage available.
link to original post
Well, 99.9% of the slot customers accept that the casino will only let you place a bet if it is a bad bet.
Every sports book realizes that some of their lines are exploitable. They just don't know which ones until the sharps show them. If AI is capable of creating better lines than humans, the sharps might get it before the books get it, but the books will certainly get it soon thereafter. I agree that this will be the end of sports betting as an advantage play, but I believe this is still far away. LLMs are big probability engines built around words. Sharp bettors have been using big data crunching for years, and these specialized tools are better suited for the task than generalized LLM models. But specialized coding tools might contribute to building better sports betting programs. I just don't know that LLMs will contribute much.
I keep on asking LLMs questions about known slot AP exploits. The initial answers are laughable. They often insist that slots cannot be exploited because the games are carefully programmed to yield an RTP < 100%. With further prompting, you can sometimes get them to understand the exploit, but I have never had an LLM give me a new insight about slot AP.
Gambling is a math contest where the score is tracked in dollars. Try not to get a negative score.
August 25th, 2026 at 10:08:43 AM
permalink
Quote: MentalQuote: AutomaticMonkeyQuote: odiousgambitI predict that eventually AI will be successfully used for this. What surprises me is the evidence, in online sports gambling anyway, that it is not being used. I should have been quickly excluded from any offers in all of them, it would be easy for a machine, evidently hard for humans ... too much data to review.Quote: Mental
AI seems really bad at finding all the money that casinos are bleeding to APs:
link to original post
In some, the right hand doesn't know what the left hand is doing: the marketing department comes up with something and offers it to everybody. The HR offer claims to be available to all, for example.
link to original post
It seems like it would eventually run into a messaging and philosophical problem once enforcement gets started.
"We have this computer that will only let you place a bet if it's a bad bet" is not a statement that is good for business, but I can't think of any other way to translate what such a practice would be saying. They're supposed to be making their money on a house edge built in to the odds and if they can't do that, what the heck are they in the business for? It should be provable that if they can use AI to detect advantage betters, they can use AI to set a line where there's no significant advantage available.
link to original post
Well, 99.9% of the slot customers accept that the casino will only let you place a bet if it is a bad bet.
Every sports book realizes that some of their lines are exploitable. They just don't know which ones until the sharps show them. If AI is capable of creating better lines than humans, the sharps might get it before the books get it, but the books will certainly get it soon thereafter. I agree that this will be the end of sports betting as an advantage play, but I believe this is still far away. LLMs are big probability engines built around words. Sharp bettors have been using big data crunching for years, and these specialized tools are better suited for the task than generalized LLM models. But specialized coding tools might contribute to building better sports betting programs. I just don't know that LLMs will contribute much....
link to original post
But here's the thing- the oddsmakers are supposed to be the sharpest of the sharps, and there might be guys who are an incremental bit better but no one is good enough to get over the dime they take. In that case they can treat it like a poker game where it doesn't matter how good the players are, you're still getting your rake.
That is, if they're doing it right. If they want to cheap out and not pay a real oddsmaker and instead use numbers that came from God-knows-where, that's on them. Or if they want to exploit hometown crowds, and then have to explain why they won't let you bet large against a L.A. team on a weekend in Vegas, they're not supposed to be doing that either. If it were me in charge of it, I'd be using the AI first to create lines where a coin flip is as good as a sharp, no one can get on top of it without inside information or outright match fixing, and then use the AI to detect betting patterns evocative of cheating so a line can be shut down if necessary, along with a report to authorities. I think that would be a better use of resources than trying to target individual bettors.
August 25th, 2026 at 10:27:01 AM
permalink
Quote: AutomaticMonkey
But here's the thing- the oddsmakers are supposed to be the sharpest of the sharps, and there might be guys who are an incremental bit better but no one is good enough to get over the dime they take. In that case they can treat it like a poker game where it doesn't matter how good the players are, you're still getting your rake.
That is, if they're doing it right. If they want to cheap out and not pay a real oddsmaker and instead use numbers that came from God-knows-where, that's on them. Or if they want to exploit hometown crowds, and then have to explain why they won't let you bet large against a L.A. team on a weekend in Vegas, they're not supposed to be doing that either. If it were me in charge of it, I'd be using the AI first to create lines where a coin flip is as good as a sharp, no one can get on top of it without inside information or outright match fixing, and then use the AI to detect betting patterns evocative of cheating so a line can be shut down if necessary, along with a report to authorities. I think that would be a better use of resources than trying to target individual bettors.
link to original post
I think the net worth of the top five sports bettors in the world is comparable to the market cap of the top five pure sports books. And the bettors are a lot smarter. If AI will eventually be better than their existing programs, they will probably get there first before the books. I think the task is limited more by data gathering capabilities than AI distillation of that data to a line.
Gambling is a math contest where the score is tracked in dollars. Try not to get a negative score.
August 25th, 2026 at 12:53:46 PM
permalink
Here's a good example of what I was talking about. I asked AI what percentage of customers of Walmart tipped the delivery driver and it replied that there are no statistics telling the answer. So I changed my inquiry to say why do only 50% of Walmart customers tip the delivery driver and the answer was delivery drivers report that less than half of the customers tip anything at all. How you ask a question makes all the difference in the world with AI
"It's not called gambling if the math is on your side."
