This article http://spectrum.ieee.org/automaton/robotics/artificial-intel...
says 120000 hands were played over the course of more than two weeks. How much of a factor was sheer boredom?
AI Beats Four Top Poker Players
141–150 of 234 posts
Re: AI Beats Four Top Poker Players
#142Earlier quoted context omitted.
I hire the guy that read the papers and know who to hire :) In an orchestra, the conductor is not replaceable but the instrument players are. Similarly, someone who allocates his or other's capital cannot be easily replaced very much like the worker's. Your engineer leaves there are 100 others ready to take the place. Business owner leaves you'd have to find a buyer to keep everyone going. This is different from a CE…
> Your top sales guy makes $9 for every $1 spent on him (+$8) where as your top engineer costs $3 for every $0 he generates (-$3) In this case you should immediately fire all your engineers, your CFO, and your COO. Because why would you want to employ a team of engineers who are costing you money but bringing you no value? You should also fire your cleaning staff, who have never closed a deal in their lives, plus the…
I'm glad you posted it because it shows how little people understand their engineer's financial importance.
Granted, engineers are required to create products and maintain it. You need product to sell.
But at the marginal level for every dollar your sales person makes your engineer is not. You might argue but the product is generating revenues but that's not what drives a sale. A sale is a function of value derived from the product and the price paid for it. Engineers aren't driving the sale unless the product itself is a developer tool. But even in that case the top level business controllers will always have the final say. In the end, buying such developer centric tool is about minimizing cost expenditure.
There is critical proprietary knowledge and experience that is formed from a sales person that is driving revenues. Businesses do not want to lose the hen that lays the golden egg as they are hard and expensive to replace.
Engineers on the other hand are easily replaceable and are expensive because they cannot generate revenues. They are not efficient when pulled away to a sales meeting from the work they are doing. Their knowledge and skill costs businesses time and money with no way to get it back directly. It must be sold, money collected and redistributed to all the payrolls. That's the main risk a business owner or a CEO is dealing with where as the engineer collects a cheque at the end of the month with little to no concern or exposure to that risk.
Society rewards risk takers disproportionately at the corporate level. Even if it's tough to measure to the Board and investors losing a C-level executive is always going to be more impactful than a senior engineer who has far more workers to replace him as a result of being cost intensive.
Find me an engineer that can code and sell, now that is a truly rare hybrid, a mewtwo, but it will still be more expensive than a guy who just sells (and does well).
A poor salesman understands this better than anyone, his weight is worth the revenues he generates. For engineers it's the efficiency / dollar or output / dollar that they are competing against which is always headed towards commoditization and any business would replace them with an AI that can code if they could if it cost less.
Re: AI Beats Four Top Poker Players
#143Earlier quoted context omitted.
And yet the people who made fortunes in the auto industry were all car makers? What an odd coincidence. But let's indulge your point of view for the sake of the argument. So you've got a black box AI that does something. You know what inputs it needs and what outputs it produces. How do you arbitrage it? What creates the price discrepancy that allows you to arbitrage at all, given that your incapacity to understand t…
The arbitrage is on the labor. Find someone who can build x for y and then sell it for z. The arbitrage opportunity comes from creating or obtaining a market you control. That market can be everything from a platform(Apple) to a collection of sales channels(Oracle) to a monopoly IP position(Microsoft). You do not need find people because they come to you because you control the market. > In that case, why not trade a…
You nailed it. AI advancement leads to arbitrage opportunities in white collar labor as did mechanical automation lead to arbitrage opportunities in blue collar labor which leads to group of giant firms that collude to keep wages low as a result of the Nash Equilibrium. Apple won't pay workers more than Google does and vice versa. Starting a tit for tat would drive wages high which would increase uncertainty and risk for market monopolies.
Re: AI Beats Four Top Poker Players
#144Earlier quoted context omitted.
That's only true if the humans made decisions about what computations to do after each day's play. The computations could have been determined in advance, eg. always run some algorithm to analyze the opponent's play for weaknesses. That's not cheating because the human players get to do the same thing. It's impossible to enforce a "no thinking outside the game room" rule for a multi-day tournament.
I dont think it's "cheating", but it's not just AI that's beating the humans. If operators are allowed to change/tune the AI's algorithm, the human opponent should get to replace himself with another player. or conversely, operators shouldn't be allowed to tweak the AI logic, but if they program the AI to tweak itself , like review its moves and adjust its algorithm automatically , that would be reasonable, then the…
I don't see why you call it "operator tweaking" even if the computations were determined in advance. You're asking for the poker bot to be scripted to log into the Pittsburgh Supercomputing Centre on its own and run overnight jobs there. That idea sounds extremely brittle and the challenge of making that script reliable has nothing to do with poker.
Re: AI Beats Four Top Poker Players
#145Earlier quoted context omitted.
I'm surprised that Andrew Ng made this claim. The strategy that was built for Libratus' predecessor did not do sophisticated modeling of the opponents, or use new algorithmic principles. Poker is solved using a very large game tree, just as with the other games. The structure of the tree is modified to support the notion of hidden state, but beyond that it is essentially the same as the other games. The structure for…
> Poker is solved using a very large game tree You mean Libratus' strategy used a very large game tree. That is not the only strategy. Take a look at research from the University of Alberta [0]. Also, I'm not certain Libratus' strategy can be simplified to "very large game tree" as I haven't seen the paper, yet. While finding a Nash equilibrium means no other player can beat you, it doesn't mean you're going to make…
A Nash equilibrium (or very close approximation thereof) does sound like it would reliably accumulate a chip lead over a very large number of hands against four top level human players though. They'll tend not to have readily exploitable weaknesses and won't be equilibrium players either.
Re: AI Beats Four Top Poker Players
#146Earlier quoted context omitted.
> Poker is solved using a very large game tree You mean Libratus' strategy used a very large game tree. That is not the only strategy. Take a look at research from the University of Alberta [0]. Also, I'm not certain Libratus' strategy can be simplified to "very large game tree" as I haven't seen the paper, yet. While finding a Nash equilibrium means no other player can beat you, it doesn't mean you're going to make…
Key example of that Nash point is Rock Paper Scissors. You can't exploit anybody playing 100% random, yet that's the Nash Equilibrium. RPSAI competitions tend to have 100% random players perform quite poorly
Do you mean to say that 100% random RPSAIs have a lower winrate vs humans than RPSAIs that learn and exploit human patterns? Surely a 100% random RPSAI doesn't have a poor win rate against any other RPSAI?
Re: AI Beats Four Top Poker Players
#147"Each night after the play ended, the Pittsburgh Supercomputing Centre added computations to sharpen the AI's strategy." This sounds more like an "advanced chess" setup, where a human teams up with an AI to play. The title of the article should really be "amateur poker players + AI defeat professional poker players". The real test would be if the AI self-corrected over the length of the tournament, without human inte…
"He added that the professionals had been sharing notes and tips in an effort to find weaknesses in the AI's game-play." This doesn't sound like they had the goal of conducting a fully controlled experiment here, but it's still interesting none the less.
Especially given that the professionals were live streaming and kibitzing with chatters on Twitch during the event.
Re: AI Beats Four Top Poker Players
#148This article http://spectrum.ieee.org/automaton/robotics/artificial-intel... says 120000 hands were played over the course of more than two weeks. How much of a factor was sheer boredom?
It was a factor for sure, especially as the AI would tank (stop and think) quite a bit at predictable spots and so the "Pros" would occasionally use strategies to try to knock out the AI before it would get to that point (they were shooting for a targeted number of hands, so maximizing the hands per hour is to the Pro's benefit). That said, they still had both money and pride on the line, and at least one Pro insisted that he was genuinely doing everything he could to try to win, despite the boredom and desire to be done with it.
Re: AI Beats Four Top Poker Players
#149Put it against Negreanu. "Oh you have AJ don't you?"
I'd watch that Twitch stream. (I've mentioned it elsewhere on here but I still find it strange that they were livestreaming and allowing the pros to interact with the viewing audience during the experiment.)
Re: AI Beats Four Top Poker Players
#150Earlier quoted context omitted.
Funny, I know more people who are asking about machine learning than people who are experts. So who is in high supply?
Machine learning experts are in higher supply, your anecdotal experience offers little weight. High demand creates a high supply which drives wages to zero.
For the foreseeable future, machine learning experts will be in sufficient supply for them (i) to have wages which are considerably above zero and (ii) have the ability to raise capital from people that don't think they're commodity labour
(it cuts the other way too: if an oversupply of labour in a field as difficult-to-learn as ML does arise, the demand shortage causing wages to fall is almost certainly because ML techniques aren't giving companies and traders as much of a financial edge as they hoped for...)
Added bonus for the ML experts: in a lot of the possible areas they can choose to work in, their individual marginal contribution to the company's profitability is at least as quantifiable as that of a salesperson. In many areas of finance you can quantify the impact of an individual line of code!