The rule was no cell phone use allowed. If a teacher saw a phone used, we could take it, write up student and turn phone into office. Was a zero tolerance by admin. The student would get in school detention. At my school admin was very supportive of teachers.

I was teaching mostly seniors. Many only taking a couple or few classes and then going immediately to work. So they had legitimate reasons to have phones. Short of total ban with a check at the doors, students will have phones. They are very adept at using them wo being seen. Plus teacher is outnumbered 20+ to 1

I believe I did a good job of preventing use but not perfect. I caught one student using phone on a test. But this is what I was alluding to. Photomath is easy to use. Just snap a photo of the problem and receive answer. Chatgpt would be harder to use since you need to type in question. But add a scan and ocr front end (very simple to do, I suspect some has already done so) and voilà.

Short of outright campus wide ban, it is a real challenge. Parents don’t make it easy either.

I was teaching HS math until Nov (had to leave for minor surgery, still considering return). So hat off to MrsSpud.

The tech likely already all exists. Snap image of problem, ocr, feed to chatgpt to id and extract equation(s), pass to photomath to solve and return answer.

At most minor tweaks to algorithms to extract the math out of the words. But applying known and developed “laws” and “theory” is in the wheel house for AI. For efficiency, they could focus the backend material to math and science sources.

The ability to solve the equations already exists in photomath (and similar) apps. Just need AI to extract the math from the words and formulate the equations.

Why would anyone use ChatGPT for that instead of RTA software that already exists?

FWIW, for the last year or so all our product descriptions and marketing materials (in the beverage industry) have been AI created from a Google sheet where we store all relevant information and a couple tags. Since last fall, we even pick product names and can designs from a list of AI suggestions.

For many years, cell phones were banned from live play for poker. However, no room these days would consider such a rule, it would kill their business ....

My example was in a live tournament blind on blind situation.

I think the main use will be by weaker players.

In another thread, I posted about Nick Vertucci not knowing why Mariano waits for players in front of him to act pre (in the HCL cash games), before looking at his hole cards.
So let's just say that NV becomes aware of ChatGPT, well then the next time there is something in game that he doesn't understand, he may well use ChatGPT to seek out the answer/explanation. (when he is away from the table, or perhaps after the session)

Even something like using ChatGPT to interrogate "how to deal with poker tilt" or "how to stop going on tilt in poker" could produce a useful tip sheet that a rec player can
stick on the wall above their computer screen when playing online. Tips about avoiding tilt are already out there of course, but a lot of players will be too lazy to search for them
and compile them, but if a bot can produce all of the best tips summarised on to one or two pages, then one would be stupid not to take advantage of this.

I consider myself to be a bit better than the category of player above, but even so, there are some elements of PLO cash games that I know I am not as strong at as others.

One example is thin value betting the river. So if I ask 5, 10, or 20 slightly differently phrased versions of questions about that topic I will likely get some useful output.
You can play around with it so much, even ask, how does a named player bet for value on the river in PLO. There is definitely a lot of trial and error in achieving useful outputs,
but the rewards could be great.

I appreciate that the quality, relevance and accuracy of the output results will be variable, but this should improve over time.

Regarding RTA itself, well if someone is planning on using it, then there's a good chance that asking ChatGPT how to most effectively use RTA, might produce some good information.

Even the best poker players in the world might learn some new things using ChatGPT, simply because it is obviously impossible for any human to read everything ever published (since 2020) that is on the internet, as well as all of the new things that will be published going forwards.

ChatGPT, I predict will also produce far more child geniuses than ever before, because gifted children will be able to read and learn about anything and everything way faster than ever before.

Which is exactly what you could use RTA for. Fortunately, that's not allowed and getting help from ChatGPT also wouldn't be allowed during hands.

What is allowed to use during hands and magnitudes more useful than what ChatGPT can do right now are charts etc on your phone. We had a thread about that not too long ago:
https://forumserver.twoplustwo.com/2...-live-1808780/

So you are not allowed to read strategy articles in between hands during a poker tournament?

That is what I am talking about, condensed down strategy advice, created by ChatGPT, not the actual RTA solution.

I said specifically during hands. In between hands you can read strategy stuff and look at solutions. That's what the linked thread is about.

Frankly, I am excited for everyone sitting at the table who benefits more from reading a strategy article than studying charts / solutions.

As long as we have players who benefit from reading strategy articles the game is in good shape.

The answer is clearly no, chatgpt wont, but someone smart and it heavy out there probably already is working on training a similar model only with GTO solutions, or already did. Then it would, but then it also wouldnt be different to the rta that is already out there...

ChatGPT was trained with publicly avaiable data, they basically scraped the whole internet for all available information. I highly doubt that all the GTO solutions are publicly avaialable, as they are all behind a pay wall.

So no.

The solutions aren't the same against every type of opponent that is in the BB when it gets folded around to us in the small.

I've never looked at a push/fold chart in my life, but I know that I'm open shoving about J4o+ against a tight player at or <12BBs effective, but I'm tightening up my
open shove range if I've profiled the BB as an online MTT crusher as they will call me down much lighter (assuming they have me down as just the average of the population),
they are calling down with ~QTo+. (Sometimes ICM dependant of course)

So there is ICM to possibly consider for both me and them. Also possible live tells, such as if we have we picked up that the BB just doesn't look interested in their hand and we are sure that they are not giving off a deliberate reverse tell.

Even ICM is not always an exactly calculable factor, because you could have two completely identical comps, stack sizes, clock, pay structure, everything.
But in parallel universe comp 1, most of the table are playing tight and conservatively, but in parallel universe comp 2, the medium and big stacks are going
to war with each other.

My ranges aren't going to agree with a solver, but I am not interested in looking at or following a solver religiously anyway, because of all of the other factors I mentioned.

If I was multi tabling lots of online MTTs then yes, it would be foolish not to learn all of the solver suggested ranges, but live, no, not for me.

Maybe my piece above will be picked up by ChatGPT one day! And this is what I am talking about; assuming my piece has some value, even if it isn't completely accurate but in parts is good advice, then a machine learning bot could amalgamate parts of it with other pieces posted by anyone, anywhere else on the internet, and come up with a valuable resource for players that is based on a consensus of opinion.

"I'm not convinced buddy".

Lol, lol, lol, wp.

-------------------------------------------

There is of course a paradox happening here, because ChatGPT is displaying a very high degree of artificial intelligence in the act
of telling us that we have nothing to be scared about regarding it's high level of artificial intelligence!

It's amusing how chatgpt often seems to give highly conservative answers on topics, while at the same time coming off as someone who is confidently wrong.

I tested it a bunch, it can not even solve simple logic problems, let alone super complicated poker spots. I have no knowledge of coding or ai, but to me it looks nothing like artificial intelligence but rather some complex coding program fed with all the information on the internet. Sure it writes you fantasy stories if you give it just some words to base the story around, all very interesting stuff. But if it fails to be able to use logic/critical thinking skills, has nothing to do with artificial inteligence to me.

Try asking it some poker questions. I look forward to the guys following the advice at the tables

Math and logic definitely aren't its strong suits, and to answer the question in the title, poker is safe (for now).

It is scary good at some other things, though. I do digital marketing as a part of my day job, and ChatGPT is already pretty good at things like SEO and copywriting. Any human worth their salt is still better, but AI is closing the gap, fast.

Hello,

I am studying for an Cloud AI certification so i am interested in this.

Not worried about ChatGPT or any other sort of AI for poker... Because it already exists. How many years on this site have people been complaining about bot infested poker websites.

Besides, hasnt Doug Polk already beat AI?

That being said... There should be a WSOP event for AI developed applications only w/ shot clock.

Much like there is an ELO rating and even chess tournaments for chess engines, would be interesting to see what developers, investment groups, universities, etc...would enter their bot(s) into said competition.

Something like that would never happen anytime soon... But would be interesting to see multiple poker engines play super deep stacked tourneys against each other with developers/et all with skin in the game.

It's quite a useful tool for looking up how to code something. I use it for that instead of asking stack exchange now, (mostly). I use it to look up healthy recipe's if I wanna cook something new. It's useful if you use it with extreme skepticism.

I've asked it to describe some mathematical proofs for me, basic undergraduate stuff, and it's pretty hit and miss. It comes off as downright stupid sometimes. But that should be expected from something that thinks in terms of statistics.

This begs the question of how time zones impact these kind of answers. Maybe tomorrow's the day?

"No, ChatGPT is an AI language model and does not have the capability to kill or impact the popularity of poker. Poker is a popular game and hobby enjoyed by many people, and its popularity is unlikely to be affected by the existence of AI language models like ChatGPT."

-ChatGPT

One of my high school teachers brought a signal jammer to work so nobody's cell phone would have service but he got in trouble and had to get rid of it lol

Chat GPT is cool but no, it is no threat to poker

Apparently, anything created by ChatGPT is public domain because it does not include "human" authorship which is a prerequisite for copyright. People can just copy each other for academic or commercial reasons without legal worries.

So basically just like your standard 2p2 poster on any topic.

Sure, except some posters tend to give great answers far more often, and some get things terribly wrong far more often. Once you've been around for a while, you start to learn which posters fit into each category. ChatGPT is like all those posters mashed together, and you don't know whose answers you're getting for any particular question.

This confirms that it can't kill poker. Not yet anyways.

Solvers being used in real time would be way more dangerous than chatGPT cause they are actually meant to "solve" poker. I know GTO wizard isn't technically a solver, but what's stopping people from firing up a GTO wizard solution in a tough river spot while time banking all the way down (on a separate device like a laptop with their phone's internet).

Chat GPT doesn't understand advanced gambling concepts.

Doesn't really understand most Advantage Plays in a casino.

I wouldn't expect it to have decent knowledge of poker and the effects of flop sizing/perceived range vs CB sizing.

I'm sure it doesn't understand smaller bets or over-bets.

D.