Sadly I don't necessarily find the accusations of sexual harassment all that surprising in EA circles. IMO the attitude that your circle is particularly virtuous in some related (too rational, too holy, too woke and aware) is a risk factor. Indeed, I think one of the worst things we do regarding sexual assault and harassment is to present it (as in Epstein) as something only someone truly awful would do and thereby make it that much harder for anyone to even wonder if they or their friends might be doing it.
Having said that, I'd quibble about framing the problem in terms of our tools not being sufficient to handle predatory behavior. I mean it's only a quibble since most of what the piece says seems right and I certainly agree predatory behavior is a very hard problem. One of the things that makes it so very hard is that many 'solutions' can themselves be abused or even make things worse. For instance trying to bar relationships between TAs and students or boss and subordinate often give the predatory individual greater power (you promised to keep us a secret, I could lose my job). But framing it as our tools are inadequate presupposes we should expect them to be better. Is that right? Or is it a very hard problem and we should be surprised we are doing as well as we are?
Personally, I tend to think that the better frame is: given that sexual harassment and assault seem unfortunately common how can we help recognize and avoid doing or enabling that behavior rather than framing it as what tools do us good people need to stop those bad guys. Not to mention the risks that trying to create those tools will make things worse.
> Present-day AIs aren’t just “next-token predictors”: they can do useful work and have undergone reinforcement learning so that they answer questions usefully [The Argument]. The “AIs are just stochastic parrots” thing is highbrow misinformation
I would describe the phrase "AI are just next-token predictors" as itself neither correct or incorrect; it's more that it's at the wrong level of abstraction. The interface to our AI systems is that you feed them a bunch of tokens and they output more tokens. In some sense that means they are next-token predictors and in some sense that's a big part of how they are trained.
The main word I object to here is "just" because you're describing the algorithm that forms the basis for them, and then saying it can't account for human-like thought, because humans [??????].
We have only some vague guesses as to how human cognition really works, like variants of Hebbian learning. Would it make sense to say that humans can't learn because they're "just" neuron co-wirers?
Fundamentally we don't know what algorithms are sufficient to produce what levels of intelligence. Personally I do think LLMs are unlikely to blow past humans (mostly for lack of good training data) but I definitely would not stake my bet on the algorithm being the limiting factor.
I don't know about anyone else, but when IT HAPPENS my plan is to drive around my city with the windows down while my speakers play such assorted tracks as "Ding Dong the Witch Is Dead", "Yub Nub", and other songs of SIMILAR SUBJECT MATTER. Some may call this behavior ANTI-SOCIAL, but I say that compared to instigating VARIOUS RECENT EVENTS my planned response is POSITIVELY MILD. I should probably assemble a playlist now to be ready when IT HAPPENS.
To answer the tweet: because humans are evolved to find love stories particularly compelling. Other things being equal a love story is just going to be more interesting to most people and when you make it a competition you look for every advantage. Not to mention that it's much easier to communicate a love story within the limited range of ice skating.
I dunno; people find violent conflicts pretty compelling also. I for one would love to see a heist skate. I've seen a lot of (non-ice) dance performances that simulate other kinds of conflict and they're often pretty great.
ohh sure I agree violent conflict is good too. But the problem is that you don't have much to work with on the ice and it's hard to look like you are fighting while you engage in intimate lifts and twirls unless it's romantic conflict.
And if you do want to display violent conflict how do you convey the motives to the audience? Miming theft is hard. A great motive for violence that is easy to convey is romantic disagreement/rejection/etc
But also violence tends to make us dislike one of the participants -- even pretend violence. If you want to win a medal you want judges to be filled with positive emotional energy toward you.
I recognize the awful lawlessness of Trump!ICE and the actual neighborliness of Minnesota.
And yet I see that the Minnesotans have actually achieved what is only a theory and a aspiration for me, and at the same time their success in deciding who is permitted to be their neighbors would seem to make it more difficult for me to do the same.
And I look to the future of whether my side can use that form of solidarity when the inevitable political reversal comes.
would you be willing to say more about your expected political reversal? is there a different government agency (or outside group?) you expect to be patrolling the streets while masked, pulling people out of their cars, deploying lethal force, etc?
> At such establishments, a customer could order coffee and cigarettes to be delivered, with the added feature that the courier was sometimes a woman who would provide sexual services upon arrival.
That sounds... rather more expensive than coffee and cigarettes...
> Oliver Sacks’s early books (up until The Man Who Mistook His Wife For A Hat) were exaggerated to the point of being fraudulent
Going by the article, it sounds like this *includes* The Man Who Mistook His Wife For A Hat? The way you phrased it, I read it as excluding it.
A question about Roblox that article didn't answer: If everything is user-made, what did it *start* with? Wikipedia seems to mostly answer this question; it launched in 2006 with some games made by Roblox themselves, the ability for users to make games (while always the idea) wasn't ready until later that year, and Roblox kept adding games of their own until 2008. I am still curious though just how many games it initially launched with and what they were.
Sadly I don't necessarily find the accusations of sexual harassment all that surprising in EA circles. IMO the attitude that your circle is particularly virtuous in some related (too rational, too holy, too woke and aware) is a risk factor. Indeed, I think one of the worst things we do regarding sexual assault and harassment is to present it (as in Epstein) as something only someone truly awful would do and thereby make it that much harder for anyone to even wonder if they or their friends might be doing it.
Having said that, I'd quibble about framing the problem in terms of our tools not being sufficient to handle predatory behavior. I mean it's only a quibble since most of what the piece says seems right and I certainly agree predatory behavior is a very hard problem. One of the things that makes it so very hard is that many 'solutions' can themselves be abused or even make things worse. For instance trying to bar relationships between TAs and students or boss and subordinate often give the predatory individual greater power (you promised to keep us a secret, I could lose my job). But framing it as our tools are inadequate presupposes we should expect them to be better. Is that right? Or is it a very hard problem and we should be surprised we are doing as well as we are?
Personally, I tend to think that the better frame is: given that sexual harassment and assault seem unfortunately common how can we help recognize and avoid doing or enabling that behavior rather than framing it as what tools do us good people need to stop those bad guys. Not to mention the risks that trying to create those tools will make things worse.
> Present-day AIs aren’t just “next-token predictors”: they can do useful work and have undergone reinforcement learning so that they answer questions usefully [The Argument]. The “AIs are just stochastic parrots” thing is highbrow misinformation
I would describe the phrase "AI are just next-token predictors" as itself neither correct or incorrect; it's more that it's at the wrong level of abstraction. The interface to our AI systems is that you feed them a bunch of tokens and they output more tokens. In some sense that means they are next-token predictors and in some sense that's a big part of how they are trained.
The main word I object to here is "just" because you're describing the algorithm that forms the basis for them, and then saying it can't account for human-like thought, because humans [??????].
We have only some vague guesses as to how human cognition really works, like variants of Hebbian learning. Would it make sense to say that humans can't learn because they're "just" neuron co-wirers?
Fundamentally we don't know what algorithms are sufficient to produce what levels of intelligence. Personally I do think LLMs are unlikely to blow past humans (mostly for lack of good training data) but I definitely would not stake my bet on the algorithm being the limiting factor.
I don't know about anyone else, but when IT HAPPENS my plan is to drive around my city with the windows down while my speakers play such assorted tracks as "Ding Dong the Witch Is Dead", "Yub Nub", and other songs of SIMILAR SUBJECT MATTER. Some may call this behavior ANTI-SOCIAL, but I say that compared to instigating VARIOUS RECENT EVENTS my planned response is POSITIVELY MILD. I should probably assemble a playlist now to be ready when IT HAPPENS.
Really struck that you see coercion and evil in the Kanakia story. Write more about it?
I would also love to hear more about this. The guy in the story doesn't sound coercive or evil to me.
To answer the tweet: because humans are evolved to find love stories particularly compelling. Other things being equal a love story is just going to be more interesting to most people and when you make it a competition you look for every advantage. Not to mention that it's much easier to communicate a love story within the limited range of ice skating.
I dunno; people find violent conflicts pretty compelling also. I for one would love to see a heist skate. I've seen a lot of (non-ice) dance performances that simulate other kinds of conflict and they're often pretty great.
ohh sure I agree violent conflict is good too. But the problem is that you don't have much to work with on the ice and it's hard to look like you are fighting while you engage in intimate lifts and twirls unless it's romantic conflict.
And if you do want to display violent conflict how do you convey the motives to the audience? Miming theft is hard. A great motive for violence that is easy to convey is romantic disagreement/rejection/etc
But also violence tends to make us dislike one of the participants -- even pretend violence. If you want to win a medal you want judges to be filled with positive emotional energy toward you.
I recognize the awful lawlessness of Trump!ICE and the actual neighborliness of Minnesota.
And yet I see that the Minnesotans have actually achieved what is only a theory and a aspiration for me, and at the same time their success in deciding who is permitted to be their neighbors would seem to make it more difficult for me to do the same.
And I look to the future of whether my side can use that form of solidarity when the inevitable political reversal comes.
would you be willing to say more about your expected political reversal? is there a different government agency (or outside group?) you expect to be patrolling the streets while masked, pulling people out of their cars, deploying lethal force, etc?
> At such establishments, a customer could order coffee and cigarettes to be delivered, with the added feature that the courier was sometimes a woman who would provide sexual services upon arrival.
That sounds... rather more expensive than coffee and cigarettes...
> Oliver Sacks’s early books (up until The Man Who Mistook His Wife For A Hat) were exaggerated to the point of being fraudulent
Going by the article, it sounds like this *includes* The Man Who Mistook His Wife For A Hat? The way you phrased it, I read it as excluding it.
A question about Roblox that article didn't answer: If everything is user-made, what did it *start* with? Wikipedia seems to mostly answer this question; it launched in 2006 with some games made by Roblox themselves, the ability for users to make games (while always the idea) wasn't ready until later that year, and Roblox kept adding games of their own until 2008. I am still curious though just how many games it initially launched with and what they were.