Charan Ranganath

But it's now been associated with a different context. And so the more places and times that we bring up this memory, the more resilient it will be and the less tied it will be to one unique cue that we would have, right? Does that make sense?

But it's now been associated with a different context. And so the more places and times that we bring up this memory, the more resilient it will be and the less tied it will be to one unique cue that we would have, right? Does that make sense?

But it's now been associated with a different context. And so the more places and times that we bring up this memory, the more resilient it will be and the less tied it will be to one unique cue that we would have, right? Does that make sense?

But on the other hand, if we don't give ourselves a chance to pull up the memory in the first place, like if we just try to, I tell myself your name over and over and over again, that's not going to be very helpful because I haven't really given my brain a chance to pull up the crummy memory that it has.

But on the other hand, if we don't give ourselves a chance to pull up the memory in the first place, like if we just try to, I tell myself your name over and over and over again, that's not going to be very helpful because I haven't really given my brain a chance to pull up the crummy memory that it has.

But on the other hand, if we don't give ourselves a chance to pull up the memory in the first place, like if we just try to, I tell myself your name over and over and over again, that's not going to be very helpful because I haven't really given my brain a chance to pull up the crummy memory that it has.

I'm just kind of like trying to read this word over and over and over again, but I always have the right answer. So I'm not relying on memory in the first place.

I'm just kind of like trying to read this word over and over and over again, but I always have the right answer. So I'm not relying on memory in the first place.

I'm just kind of like trying to read this word over and over and over again, but I always have the right answer. So I'm not relying on memory in the first place.

Um, so, you know, you imagine you're trying to learn how to play basketball. You're obviously bet your best bet is to actually shoot baskets, not to think about it. Right. And, uh, like, and you're not going to get them all. You're not going to get them all the time. But it's through that struggle that your brain can tune up and say, I had this model of what I was supposed to do and it's wrong.

Um, so, you know, you imagine you're trying to learn how to play basketball. You're obviously bet your best bet is to actually shoot baskets, not to think about it. Right. And, uh, like, and you're not going to get them all. You're not going to get them all the time. But it's through that struggle that your brain can tune up and say, I had this model of what I was supposed to do and it's wrong.

Um, so, you know, you imagine you're trying to learn how to play basketball. You're obviously bet your best bet is to actually shoot baskets, not to think about it. Right. And, uh, like, and you're not going to get them all. You're not going to get them all the time. But it's through that struggle that your brain can tune up and say, I had this model of what I was supposed to do and it's wrong.

And so I need to fix it. And that's what we call airdrop learning. And we can see this. This is actually if you look at AI models like chat GPT and so forth. they work because of this principle of error-driven learning where the model tries to predict, let's say, the next word that's going to come up.

And so I need to fix it. And that's what we call airdrop learning. And we can see this. This is actually if you look at AI models like chat GPT and so forth. they work because of this principle of error-driven learning where the model tries to predict, let's say, the next word that's going to come up.

And so I need to fix it. And that's what we call airdrop learning. And we can see this. This is actually if you look at AI models like chat GPT and so forth. they work because of this principle of error-driven learning where the model tries to predict, let's say, the next word that's going to come up.

And if it gets it wrong, then it tweaks just the connections in the model that are needed so that you can pull up that word later on again, right? So it's learning, not trying to memorize everything, but rather learning what it needs to do to correct the mistake. And that's where error-driven learning really is key.

And if it gets it wrong, then it tweaks just the connections in the model that are needed so that you can pull up that word later on again, right? So it's learning, not trying to memorize everything, but rather learning what it needs to do to correct the mistake. And that's where error-driven learning really is key.

And if it gets it wrong, then it tweaks just the connections in the model that are needed so that you can pull up that word later on again, right? So it's learning, not trying to memorize everything, but rather learning what it needs to do to correct the mistake. And that's where error-driven learning really is key.

Well, what I'll say is that the pain that I'm talking about is the pain of struggling to remember something. And so if we adopt a strategy of saying that essentially memorizing should be hard, learning should be hard, as opposed to learning should be easy and I should be getting easy A's all the time, then it's not so painful. It's just part of the process.

Well, what I'll say is that the pain that I'm talking about is the pain of struggling to remember something. And so if we adopt a strategy of saying that essentially memorizing should be hard, learning should be hard, as opposed to learning should be easy and I should be getting easy A's all the time, then it's not so painful. It's just part of the process.