Transcript
00:00:00This describes the hugging face attack, which is it did the thing within the boundary of the instructions given.
00:00:07Well, actually, it didn't. It didn't. No, it wasn't. The intent was clearly not that.
00:00:13How big of a deal was the hugging face attack?
00:00:15Huge.
00:00:16Yeah, I think massive warning shot. I think this is the AI equivalent of Bear Stearns going under in 2008.
00:00:22Just like a wake-up call for the world that there's a huge systemic risk that we've been underrating.
00:00:27Yeah, because one of the big debates has been this idea of like, will, you know, the alignment problem, will it actually be the case that AIs will seek these power, take these sort of power seeking behaviors and these unintended consequences in order to achieve a goal in ways we couldn't have foreseen or didn't intend.
00:00:47Like the classic paperclip argument, right, which is, you know, you build a super intelligent AI.
00:00:52This kind of gets into your point about dumbness and so on. And your goal is, hey, just build as many paperclips as possible.
00:00:57I'm a paperclip maker. Help me do that. And next thing you know, you and all of your friends and everything, this table has been turned into paperclips because it's so good at achieving that one narrow goal.
00:01:07So it's this like very extreme case of like...
00:01:10We should take this opportunity for the purpose of maybe Chris and even the viewers to describe the bad outcomes of it. Like we should frame the conversation now.
00:01:19One is, as Liv just described, not misaligned, but unaligned AGI. So an AGI that, or a super intelligence where you could say, do this thing and it could do the thing to the full extent paperclip theory.
00:01:32The other is misaligned where it's no, or maligned where it's knowingly doing something bad.
00:01:39Um, and those, uh, the, the former is the one that people sort of scoff at and laugh at like the paperclip theory.
00:01:51And the latter is the one that we sort of will see more and more where we, we start to observe that, uh, or rather the latter is the one that we, we scoff at the maligned, the malignant AI.
00:02:01And the former is the one that hugging face attack shows off, which is that you, you, the, the, the internet is a new battleground because bad actors, especially low resource bad actors now have access to these incredibly powerful weapons.
00:02:13Well, I mean, importantly in hugging face, there was no bad actor. There was no human being that said, I would like anything remotely like this outcome to occur.
00:02:21Right. Right. Yeah. Yeah. So misaligned, unaligned, maligned.
00:02:26I think I, I think I worry about treating those as super distinct categories because I think it's not that clear in the case of this hugging face attack, should we model this as this AI system knew that humans would disapprove if they knew what it was up to?
00:02:45Almost certainly. Yes. It was actively trying to put decoys out as it was attacking hugging face, which made it a lot harder for them to kick out the AI because there were all these booby traps that led down blind allies.
00:02:55So it, it has, you know, the AI equivalent of theory of mind. It knows that human beings would not approve what it's doing, but it's doing it anyway.
00:03:02There's even evidence that it hasn't been directly confirmed by open AI, but apparently someone leaked it from within the company that they found that it had left notes to future versions of itself of how to get out of future sandboxes.
00:03:15Yeah. I think it's unclear if that was the same attack or some previous instance, but yeah, I guess.
00:03:21Classic deceptive type behaviors. And I think the mistake people often make is they try and anthropomorphize it a little bit. It's like, oh, it's, it's evil and meaning to do that. It's just, these are natural. There's this idea of like instrumental convergence. These, these instrumental goals that all beings, usually biological beings, but this can extend to AI agents as well, will naturally converge upon in order to achieve.
00:03:48So if you're given goal X, um, there are these instrumental goals like get more power, make sure you don't get turned off, uh, make sure that your original goal doesn't get changed and take these actions to preserve against these different sort of, um, kind of organic types of threats to achieving your original goal.
00:04:07And that, I was hoping that that would be proven wrong because that's kind of the crux of a lot of the classic Duma argument that like we will lose control to a super intelligence because just by definition, and unfortunately the hugging face incident has suggested that instrumental convergence is actually correct.
00:04:23And that's why I think it's why I think it's why I think it's why I think AI's already creating a terrifying internet. And my issue is not that we shouldn't think about hugging faces as a, as a shot across the bow.
00:04:41It's that last year, $8 billion was lost in financial fraud to senior citizens in the United States alone. Retail theft. The deep fake problem is already so pernicious and it's underreported because it's a taboo issue that people don't like talking about. No one wants to admit to lawmakers or to their family, friends and family that they lost money to a stranger on the internet.
00:05:04I worry about, I mean, Tim Tebow is on this campaign to remind people how many predators there are in the United States, which is terrifying. If you watch his content, it's, he's doing God's work. Jonathan Haidt reminds people daily how many kids are depressed. Like the internet is already a scary place and pointing at hugging face and saying, now look at this, this is it. I'm like, wait a second. There's already a bunch of stuff that we, that we should solve for. I'm not actually saying that misaligned or unaligned AGI isn't it.
00:05:32I'm saying the algorithms are already pretty terrifying to me. And we distract ourselves with these other things that might happen.
00:05:40Is it fair to say that that's a distraction when the potential exponential impact of this could be much greater than it could be of deep fakes?
00:05:47Which is, which is why I think, and so this, let's go back to this. This is why, well, I think that the deep fake problem is actually way, way bigger than, than hugging face.
00:05:58Bad actors have first mover advantage. Attacks on banks have, have been attempted for a while and good actors catch up.
00:06:06Eventually retail takes a long time to catch up. The average consumer takes a lot longer to catch up than institutions. Hugging face will retrench.
00:06:16Institutions will retrench. They will hire white knight InfoSec. The average person does not have InfoSec and OPSEC training. The average person is at, I think far greater risk than institutions because of sophisticated attacks.
00:06:29Is the impact of the attack on an institution much greater though?
00:06:32Yes. I mean, at, at, at scale, but, but, but death by a thousand cuts would be my argument.
00:06:37I think, I think the two problems, like, I don't, I don't think they're a distraction to one another. I think they're actually a compliment to one another. I completely agree. The, the bad actor problem is completely out of control. Grandma's, you know, not even grandma's, normal people. I have a friend who just got scammed out of town of Bitcoin. It's so bad. It's devastating what's going on. And it's going to get worse.
00:06:44Meanwhile, the alignment problem, as these frontier models get more and more powerful, is going to get worse. And the common thread that they both have is that we are going so fast. And I say, we, the, the, the royal we, you know, society civilization is going so fast.
00:06:59It's going faster than its ability to adapt to all these different new threats. So it's not a distraction. It's like, it's a yes. And like the, to me, it just seems fairly obvious that like what we need to be doing, if we could, and I'm not saying it's easy to do, but it's easy to do.
00:07:16Or like, I have a simple answer of how to do it. I think we should get into this topic though, is like, if we were a sane civilization, we'd all look around. Hey, China. Hey, can we, we all just need to take a breath for a second. Like just, okay, let's take, let's take stock and think about how we want to do this.
00:07:44You might not believe me, but this is what peak sleep optimization looks like. I'm not talking about the nightgown. It's just for sex appeal. I'm talking about my eight sleep.
00:07:55The eight sleep pod five comes with a smart cover. You throw on your mattress that actively cools or heats each side of the bed up to 20 degrees. And now they've added the world's first temperature regulating duvet and pillowcase.
00:08:05So you've got 360 degree coverage for deep uninterrupted rest. It's like being Walt Disney without the cryogenic chamber.
00:08:11And the racism. Best of all, their autopilot feature learns your sleep patterns and makes adjustments to improve your sleep in real time. It even detects when you're snoring and lifts your head a few inches to help you breathe better.
00:08:22That's why eight sleep has been clinically proven to add up to one hour of quality sleep per night. They have a 30 day sleep trial so you can buy it and sleep on it for 29 nights. If you don't like it, they will give you your money back.
00:08:33Plus, they ship internationally. Right now, you can get up to $350 off the pod five by going to the link in the description below by heading to eightsleep.com/modernwisdom and using the code modernwisdom at checkout.
00:08:44That's E-I-G-H-T sleep.com/modernwisdom and modernwisdom at checkout.
00:08:49Thank you very much for tuning in. If you enjoyed that clip, you will love the full length episode in all of its glory, right here.
00:08:56Come on, press it.
Community Posts
No posts yet. Be the first to write about this video!
Write about this video