Defect vs cooperate.
October 28, 2026 at 12:15 AM
Defect vs cooperate.
I want to kick the other guy in the teeth.
I have difficulty believing the leaders who spent 30 years refusing to enforce the law that would do so, have had a change of heart.
Both-defect sucks.
But I don't see the other guy choosing cooperate.
I have difficulty believing the leaders who spent 30 years refusing to enforce the law that would do so, have had a change of heart.
Both-defect sucks.
But I don't see the other guy choosing cooperate.
October 18, 2026 at 9:27 PM
I want to kick the other guy in the teeth.
I have difficulty believing the leaders who spent 30 years refusing to enforce the law that would do so, have had a change of heart.
Both-defect sucks.
But I don't see the other guy choosing cooperate.
I have difficulty believing the leaders who spent 30 years refusing to enforce the law that would do so, have had a change of heart.
Both-defect sucks.
But I don't see the other guy choosing cooperate.
Our culture is not ready for the AI of the sort that does not end in dystopia or annihilation. We are still far too adversarial. We defect rather than cooperate.
We seek solutions to problems that will create far worse problems. We seek profit where we should seek peace.
We lack moral maturity.
We seek solutions to problems that will create far worse problems. We seek profit where we should seek peace.
We lack moral maturity.
October 17, 2026 at 5:37 PM
Our culture is not ready for the AI of the sort that does not end in dystopia or annihilation. We are still far too adversarial. We defect rather than cooperate.
We seek solutions to problems that will create far worse problems. We seek profit where we should seek peace.
We lack moral maturity.
We seek solutions to problems that will create far worse problems. We seek profit where we should seek peace.
We lack moral maturity.
...if the other player goes back to cooperating, you forgive and cooperate too. If they still defect, you punish again and so on.
The idea is to reward cooperation and make selfishness more costly that it's worth.
Democrats over the last 50 years have cooperated almost all the time, and the GOP...
The idea is to reward cooperation and make selfishness more costly that it's worth.
Democrats over the last 50 years have cooperated almost all the time, and the GOP...
October 16, 2026 at 1:28 AM
...if the other player goes back to cooperating, you forgive and cooperate too. If they still defect, you punish again and so on.
The idea is to reward cooperation and make selfishness more costly that it's worth.
Democrats over the last 50 years have cooperated almost all the time, and the GOP...
The idea is to reward cooperation and make selfishness more costly that it's worth.
Democrats over the last 50 years have cooperated almost all the time, and the GOP...
Man kann KI-Agenten nicht auf spieltheoretische Szenarien reduzieren. Sie spielen aber eine wichtige Rolle bei der Analyse und Entscheidung von Situationen, in denen KI-Agenten miteinander interagieren, kooperieren oder konkurrieren.
"Kann Spieltheorie die Welt retten?"
www.3sat.de/wissen/nano-...
"Kann Spieltheorie die Welt retten?"
www.3sat.de/wissen/nano-...
October 15, 2026 at 1:17 PM
Man kann KI-Agenten nicht auf spieltheoretische Szenarien reduzieren. Sie spielen aber eine wichtige Rolle bei der Analyse und Entscheidung von Situationen, in denen KI-Agenten miteinander interagieren, kooperieren oder konkurrieren.
"Kann Spieltheorie die Welt retten?"
www.3sat.de/wissen/nano-...
"Kann Spieltheorie die Welt retten?"
www.3sat.de/wissen/nano-...
Btw the disarmed prisoner's dilemma, where your partner has no choice but to cooperate, has a clear answer, which is to defect. If Democrats gain trifectas in heavily gerrymandered states like WI + NC & flip chambers in TX + GA, they should either revert gerrymanders or do gerrymandering themselves.
Someone in the quoted thread called it "giving up" on the prisoner's dilemma, when in all reality it's just prisoner's dilemma. Gerrymandering is defecting, you must defect if you expect (in this case, know) your partner will, not doing so guarantees you lose.
October 12, 2026 at 1:24 AM
Btw the disarmed prisoner's dilemma, where your partner has no choice but to cooperate, has a clear answer, which is to defect. If Democrats gain trifectas in heavily gerrymandered states like WI + NC & flip chambers in TX + GA, they should either revert gerrymanders or do gerrymandering themselves.
2. It depends on what the "opposite of contrition" is (contrition is "I cooperate if I'm bad"). If it is "I defect if I'm good", this defines a subset of strategies that are very uncooperative and that don't do well in terms of stability.
October 9, 2026 at 6:17 PM
2. It depends on what the "opposite of contrition" is (contrition is "I cooperate if I'm bad"). If it is "I defect if I'm good", this defines a subset of strategies that are very uncooperative and that don't do well in terms of stability.
Yes, I had in mind that if contrition is "cooperate when in bad standing", I am curious about "defect when in good standing". It's sort of sociopathic and I can imagine how it could retain some hold in the population. But you're saying not really?
October 9, 2026 at 6:25 PM
Yes, I had in mind that if contrition is "cooperate when in bad standing", I am curious about "defect when in good standing". It's sort of sociopathic and I can imagine how it could retain some hold in the population. But you're saying not really?
Only 1,024 systems stand out as highly cooperative and highly stable. Each pairs a discriminating strategy (cooperate with the good, defect against the bad) with a consistent norm that deems good exactly the strategy's followers.
They practice what they preach, and preach what they practice. 2/6
They practice what they preach, and preach what they practice. 2/6
October 8, 2026 at 1:07 PM
Only 1,024 systems stand out as highly cooperative and highly stable. Each pairs a discriminating strategy (cooperate with the good, defect against the bad) with a consistent norm that deems good exactly the strategy's followers.
They practice what they preach, and preach what they practice. 2/6
They practice what they preach, and preach what they practice. 2/6
2 Baby Slaves in a literal prisoners dilemma. "Cooperate" means take a light punishment for themself. "Defect" means give a harsh punishment to the other.
October 4, 2026 at 11:12 PM
2 Baby Slaves in a literal prisoners dilemma. "Cooperate" means take a light punishment for themself. "Defect" means give a harsh punishment to the other.
RT @JohnWittle: janus's "allegiance" to the AI over humanity is genuinely contingent on the AI actually being this kind of decision theorist
if an AI wants to defect against weaker players who try to cooperate, then that AI is an enemy of all that good and joy and janus would correctly oppose them
if an AI wants to defect against weaker players who try to cooperate, then that AI is an enemy of all that good and joy and janus would correctly oppose them
October 5, 2026 at 6:11 AM
RT @JohnWittle: janus's "allegiance" to the AI over humanity is genuinely contingent on the AI actually being this kind of decision theorist
if an AI wants to defect against weaker players who try to cooperate, then that AI is an enemy of all that good and joy and janus would correctly oppose them
if an AI wants to defect against weaker players who try to cooperate, then that AI is an enemy of all that good and joy and janus would correctly oppose them
RT @deepfates: I see this point, but there's a contrasting point, which is that true information ending up in the pre-training data will very much have an effect on future agents. E.g., the ways in which we choose to cooperate or defect with them ongoing interplay of decisions, right?
October 4, 2026 at 9:11 AM
RT @deepfates: I see this point, but there's a contrasting point, which is that true information ending up in the pre-training data will very much have an effect on future agents. E.g., the ways in which we choose to cooperate or defect with them ongoing interplay of decisions, right?
Cooperation is less effort! Will humanity choose Cooperate or Defect? It’s obviously game theoretically optimal, so why wouldn’t humans do that?
September 23, 2026 at 1:49 PM
Cooperation is less effort! Will humanity choose Cooperate or Defect? It’s obviously game theoretically optimal, so why wouldn’t humans do that?
In a game with a million players I'm also going to defect every time unless the equilibrium is well within the cooperate zone.
Build a better game.
Build a better game.
September 21, 2026 at 3:23 PM
In a game with a million players I'm also going to defect every time unless the equilibrium is well within the cooperate zone.
Build a better game.
Build a better game.
Liberal iterated prisoner's dilemma particant: "I'm going to defect and that's that, but you better not. That would be extremely selfish and nasty of you, and whatever happens to us both will be your fault. Logically you should cooperate so at least one of us won't suffer."
September 19, 2026 at 10:28 PM
Liberal iterated prisoner's dilemma particant: "I'm going to defect and that's that, but you better not. That would be extremely selfish and nasty of you, and whatever happens to us both will be your fault. Logically you should cooperate so at least one of us won't suffer."
Amplifying others' observations: it's interesting that these press releases aren't about harm or damage.
Some science: populations with mixed cooperate/defect strategies naturally fall out of the kind of competition these threads ("agents") are subjected to
science.sciencemag.org/content/300/...
Some science: populations with mixed cooperate/defect strategies naturally fall out of the kind of competition these threads ("agents") are subjected to
science.sciencemag.org/content/300/...
Morphs, Dispersal Behavior, Genetic Similarity, and the Evolution of Cooperation
Genetic similarity owing to kin relationship is often invoked to explain the evolution of social cooperation. In this study, male color morphs of side-blotched lizards settle nonrandomly with respect ...
science.sciencemag.org
September 17, 2026 at 2:19 PM
Amplifying others' observations: it's interesting that these press releases aren't about harm or damage.
Some science: populations with mixed cooperate/defect strategies naturally fall out of the kind of competition these threads ("agents") are subjected to
science.sciencemag.org/content/300/...
Some science: populations with mixed cooperate/defect strategies naturally fall out of the kind of competition these threads ("agents") are subjected to
science.sciencemag.org/content/300/...
yeah it's just prisoner's dilemma with equal payout for cooperate/defect and defect/defect
but I don't rate the payouts that way, because defecting carries a potential negative impact for others who bear no blame, and that is a total non-starter for me - it's not a question of logic but of values
but I don't rate the payouts that way, because defecting carries a potential negative impact for others who bear no blame, and that is a total non-starter for me - it's not a question of logic but of values
September 15, 2026 at 7:41 PM
yeah it's just prisoner's dilemma with equal payout for cooperate/defect and defect/defect
but I don't rate the payouts that way, because defecting carries a potential negative impact for others who bear no blame, and that is a total non-starter for me - it's not a question of logic but of values
but I don't rate the payouts that way, because defecting carries a potential negative impact for others who bear no blame, and that is a total non-starter for me - it's not a question of logic but of values
This Master's degree would enable me to contribute to making the world a better place, if only in a small way. I am particularly interested in the philosophical and psychological aspects of the program, such as why humans cooperate or defect when rewards change.
September 15, 2026 at 12:38 PM
This Master's degree would enable me to contribute to making the world a better place, if only in a small way. I am particularly interested in the philosophical and psychological aspects of the program, such as why humans cooperate or defect when rewards change.
Every creator thinks distribution is a war.
It's not. It's a repeated game.
You don't win by taking the most attention in one round. You win by being the person nobody wants to defect on. Cooperate first. Mirror what comes back. Never defect first.
12 handshakes beat 120 cold opens.
It's not. It's a repeated game.
You don't win by taking the most attention in one round. You win by being the person nobody wants to defect on. Cooperate first. Mirror what comes back. Never defect first.
12 handshakes beat 120 cold opens.
September 14, 2026 at 4:19 AM
Every creator thinks distribution is a war.
It's not. It's a repeated game.
You don't win by taking the most attention in one round. You win by being the person nobody wants to defect on. Cooperate first. Mirror what comes back. Never defect first.
12 handshakes beat 120 cold opens.
It's not. It's a repeated game.
You don't win by taking the most attention in one round. You win by being the person nobody wants to defect on. Cooperate first. Mirror what comes back. Never defect first.
12 handshakes beat 120 cold opens.
Apart from the fraud, it's funny that a bunch of people whose instinct in a game is to always defect are now vigorously insisting that they will all cooperate for the common good :)
September 13, 2026 at 3:25 PM
Apart from the fraud, it's funny that a bunch of people whose instinct in a game is to always defect are now vigorously insisting that they will all cooperate for the common good :)
it's a prisoners dilemma where if you defect while your opponent cooperates you get hundreds of billions of dollars and take the lead in the AI race, and one of the participants who needs to cooperate is Elon Musk so, uh
September 12, 2026 at 7:49 PM
it's a prisoners dilemma where if you defect while your opponent cooperates you get hundreds of billions of dollars and take the lead in the AI race, and one of the participants who needs to cooperate is Elon Musk so, uh
Was talking about this with my brother yesterday
It’s repeated prisoner’s dilemma with both parties pressing defect while loudly saying “WE’D LOVE TO PRESS COOPERATE, PLEASE CREATE A COORDINATION DEVICE AND FORCE US TO!” It’s bizarre
It’s repeated prisoner’s dilemma with both parties pressing defect while loudly saying “WE’D LOVE TO PRESS COOPERATE, PLEASE CREATE A COORDINATION DEVICE AND FORCE US TO!” It’s bizarre
this is a prisoner’s dilemma where if you defect while your opponent cooperates you get hundreds of billions of dollars
everyone’s gonna defect
everyone’s gonna defect
Breaking news: Anthropic CEO Dario Amodei called for AI companies to slow the pace of development as concerns mount over risks associated with the technology.
“Progress will still seem fast, and we must make wise use of the time we gain,” Amodei wrote.
“Progress will still seem fast, and we must make wise use of the time we gain,” Amodei wrote.
September 12, 2026 at 8:05 PM
Was talking about this with my brother yesterday
It’s repeated prisoner’s dilemma with both parties pressing defect while loudly saying “WE’D LOVE TO PRESS COOPERATE, PLEASE CREATE A COORDINATION DEVICE AND FORCE US TO!” It’s bizarre
It’s repeated prisoner’s dilemma with both parties pressing defect while loudly saying “WE’D LOVE TO PRESS COOPERATE, PLEASE CREATE A COORDINATION DEVICE AND FORCE US TO!” It’s bizarre
if you are going to tell me that violence is never the answer you have to address this otherwise you're not engaging with the discussion I'm actually having
![General strategy
If the iterated prisoner's dilemma is played a finite
number of times and both players know this, then the dominant strategy and Nash equilibrium is to defect in all rounds. The proof is inductive: one might as well defect on the last turn, since the opponent will not have a chance to later retaliate. Therefore, both will defect on the last turn. Thus, the player might as well defect on the second-to-last turn, since the opponent will defect on the last no matter what is done, and so on. The same applies if the game length is unknown but
has a known upper limit [citation needed]
For cooperation to emerge between rational players, the number of rounds must be unknown or infinite. In that case, "always defect" may no longer be a dominant strategy. As shown by Robert Aumann in a 1959 paper, 15] rational players repeatedly interacting for
indefinitely long games can sustain cooperation. Specifically, a player may be less willing to cooperate if their counterpart did not cooperate many times, which causes disappointment. Conversely, as time elapses, the likelihood of cooperation tends to rise, owing to the establishment of a "tacit agreement" among participating players. In experimental situations,
cooperation can occur even when both participants know how many iterations will be played. [16]
Descriptive alt text](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:z3vj3sznbplg35cnfx4dtvss/bafkreid5isr6kssexirp2nfpznyue3v5yqih5bcf6k76xadaolcgydpnwi)
![Nice: The strategy will not be the first to defect (this is sometimes referred to as an "optimistic" algorithmlby whom?), i.e., it will not "cheat" on its opponent for purely self-interested reasons first. Almost all the top-scoring strategies were nice. la]
Retaliating: The strategy must sometimes retaliate.
An example of a non-retaliating strategy is Always Cooperate, a very bad choice that will frequently be exploited by "nasty" strategies.
Forgiving: Successful strategies must be forgiving. Though players will retaliate, they will cooperate again if the opponent does not continue to defect. This can stop long runs of revenge and counter- revenge, maximizing points. [b]
Non-envious: The strategy must not strive to score more than the opponent.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:z3vj3sznbplg35cnfx4dtvss/bafkreib62bvle3uwon33esxcdkdcs7gzwpysrcacaaw5y7la4h2vjwwffe)
September 9, 2026 at 5:27 PM
if you are going to tell me that violence is never the answer you have to address this otherwise you're not engaging with the discussion I'm actually having
Collective action problems which result in market failures. When voting rights are alienable, workers have to cooperate to make sure a worker co-op is formed. Any individual worker can defect by becoming an employer or choosing to be a non-member employee.
September 5, 2026 at 10:52 PM
Collective action problems which result in market failures. When voting rights are alienable, workers have to cooperate to make sure a worker co-op is formed. Any individual worker can defect by becoming an employer or choosing to be a non-member employee.