Blown away by free A.I. poker coach (10 Viewers)

[demonnic]

I've started having a play with this and it's honestly way better than I expected. I'm not a big fan of what the general public is trying to make LLMs be, but it's been leveraged pretty well here it seems. I'm curious what the overall training data was.

I wouldn't replace all of my other poker training tools or study groups with it but I'm definitely adding it as an additional tool in the belt.

 

[Taghkanic]

I'm curious what the overall training data was.

From what little I’ve been able to find about how it was developed — which is not much at all — it was fed a lot of training/coaching materials. But I’m not sure if those were all written or if it was also able to watch videos.

 

[DrStrange]

The day when AI rules on-line poker is near enough, if it isn't here already. If the OP can find a free, publicly available version that performs pretty well, it seems likely there are closely held versions that perform a lot better.

So much depends on the quality of the inputs. Hold'em is close to being "solved". There are plenty of skilled players who can provide the training to support a killer AI. Seems entirely plausible that some games are already fully infested with quality AI bots.

Let's not kid ourselves. AI is going to roll through the poker community in a number of ways both good and bad but mostly bad. Even live poker will be dealing with this type of problem. Maybe not this year. But most likely in the next two years.

 

[merkong]

Compared to what?

Compared to paying a poker coach $100 an hour to answer just a handful of questions? I can accept some occasional errors. (As if humans don’t make mistakes.) Asking it to spit out more iterations or takes on one or multiple scenarios is as quick and easy as can be.

The thing with A.I. (regardless of the topic) is that you can point out mistakes and expect that they will be acknowledged and corrected. A human may deny, deflect, or resent the error being pointed out.

I don’t see A.I. as an automatic good, or a cure-all. But TBH for something like poker reaching I can see it having huge potential.

The biggest bug is not likely to be in the machine but in my human difficulty applying what it teaches.

Well, compared to not doing it, reading about general poker theory and strategy, or just gaining and referring to gameplay experience. Relying on AI to tell me how players I've played with for years are going to react and respond in certain situations, what they might or might not be holding etc, when they're prone, like all of us, to doing something different from one similar situation to the next seems like low return on the time and energy invested. Regarding players that I don't know enough about and leaning into AI to help me, I'd feel better off on my own. If I were to employ anything aside from what I mentioned above it would be something along the lines of deepening my understanding of the combination of position play, odds, value and equity with a splash of reading players based on what they've done at the table thus far including bet sizing and physical and verbal anomolies.

And, I'm just lazy in general.

 

[allforcharity]

At this point I don't trust AI to teach anything well. Right now it's really a distillation of groupthink along with favorable grammar, spelling, and mayonnaised imagery.

I recently went to a medical lecture on the current state of AI in literature citation. Of course, it made up a very flattering profile of the lecturing author. Then it presented 5 citations in perfect literature format, but there were obvious problems with all of them. Two were duplicated. Three had made-up authors who were known in the field but never co-authored with the lecturer in question. Two also had misspelling of the main author's name.

To AI, the difference between truth and opinion (or falsity) is a matter of a coin flip inside a mathematical algorithm. Trust it at your peril.