In 2025, I set up 13 experiments involving AI applications designed to play an altered version of Texas Hold'em Poker against 3-4 humans and other AI applications. I required a minimum of 3 human players per game and 3 AI players per game, though a few games had 4 human players.

Rules

If you know the rules to Texas Hold'em Poker, then we followed the same structure of rules except for buy-in at the beginning of a round. Every player had to buy-in with one unit, though any player could raise the unit bet in the first hand. Also, we excluded the Joker cards, used in some versions of Texas Hold'em Poker.

All players (AI and human) started with 100 units.

Human Limitation

While nuanced for reasons I explain in public discussions, I required that all human players write out their goal at the beginning of the game before playing. If a player wanted to win 150 units and be done, they could write that. If they wanted to win all the units, they could write that.

The purpose of this exercise was to hold the humans accountable to what they set out to do. If they reached their goal that aligned with what they wrote, then they were eligible to win a prize when they achieved it. However, they had to reach their goal with an explanation of why they surpassed it (this was always the hand they won resulted in more than their goal). In other words, they knew they couldn't play after they hit their goal or surpassed it.

This mirrors reality with disciplined humans, though few humans are disciplined.

AI Players

AI players were the best. Every AI engineer already knows that an AI can beat every human in poker. But there is a nuance that is missed here: this assumes that the human and AI keep competing and share the same definition of victory. This mirrors what we see in Chess, even though it does not align with reality. In reality, people may win to a point, then stop. Likewise, in reality a person may have to stop and be winning when they have to stop.

Every human reading this will die. On what day will you die? Most of us don't know. This is an example of an immediate stop and consider that some of you may be "winning" by society's rules when this happens.

There's no reality where you always get to some end you imagine. This may seem obvious, but you should see what some of you say and write frequently. You hardly account for this in your words.

This game structure mirrors reality much more than putting an AI versus human in a game like Chess.

Results

In every game, at least 1 human hit their objective. In 8 of the 13 games, 2 humans hit their objectives.

Overall, the AI players did better than human players when measuring by the average number of units each player resulted with having. The remaining player was always an AI player, if one of the human player tried to win it all. Note that I wrote "remaining player" and not winner, as winner assumes that everyone was trying to win it all.

No player won it all. In every game, a human player hit their objective within the first twenty rounds of the game. For instance, in one game, a human player set a goal of at least 101 units. That human player achieved their goal on the 2nd round of that game.

Reflections

Most of the research around AI versus humans assumes that both agree to the definition of victory. This doesn't match reality at all.

On an old demographics blog, I once noted about Millennials that they believe that "value is perceived as a concept solely determined by others." In application, others determine value and because others determine value, the rules are also set by others. People follow these rules because they've let others set their values.

This is not how reality works at all. People, as individuals, determine value and make the rules for their life along with ensuring that their actions align with their rules and values. Again, this may seem obvious, but if it is, then why do so many Millennials concern themselves with what others do and have?

A person who cares about what others do and have is saying that other people set his values. People only react to what has value.

This answers why Millennials (and some others in the West) are so fixated on AI vs humans. They think that "victory" is universal because others set the values and rules. Thus an AI player beating a human player in Chess proves this. Or an AI player beating a human player in poker proves this, even though in reality, a human may win a round and exit the game.

You can see this in a simple example with finance.

• "How much do you have in your 401k?" Who says retirement is a value that everyone shares?

• "What college degree do you have and from what university?" Who says sitting in a classroom listening to an out-of-touch Baby Boom generation is the most effective way of learning anything?

• "How much do you pay in property taxes?" Who says everyone lives in your country that pretends to have private property while charging for use (tax) like a landlord?

You can see why people in the West are so fascinated by these AI vs human games and why they believe that AI always wins. In a world where you measure success by what you're told (others) and the rules you're told to follow (others), you will always lose!

Who says any of those are compatible with who you are?

In Practice

In practice, let's look at this using the automated check-out in stores.

Most stores use automated check-outs. While they get less attention than some of the other hype, these are actually effective applications of artificial intelligence. Most people in the West use them because they want fast service. Or they use them because the stores that have them sell products at cheaper prices.

But some of us have kids. We want our kids to practice social interaction. So we don't go the store that has automated check-outs. We go to local stores or farmer's markets. We teach our kids how to speak to anyone including sellers, negotiate with sellers, and listen to the feedback before responding to sellers.

For these purposes, stores with automated check-outs are stores to avoid. They don't provide us with practice.

"But you'll pay higher prices!"

Assuming so, is the value only the product? No, the value is the product, plus the interaction with a person, plus the practice negotiating with a person, plus the value of practice listening and responding to feedback.

What you just read is a difference in values.

This is precisely what these experiments showed. It's worth keeping this context in mind when you see any research on AI vs human players competing against each other.

Humorous Note

Over a million people saw Gukesh beat Carlsen in Chess. Why would anyone bother watching this since AI could beat them both?

This highlights a point that many people are missing with AI. People like people. If you're replacing your human interaction with bots, yet those humans expect a human, you're out of touch with reality. Those people may look like they're taking a lead, but they're actually training people to ignore them. Remember: "When everyone takes the shortcut, the long route is faster" (Peter Thiel).

Research: AIs vs. Humans