Latest / Connor with Honor: AI, Real Estate and Real Life / Ah, the old AI kill switch you say? I think the Flux Capacitor maybe a better option...

Transcript

Ah, the old AI kill switch you say? I think the Flux Capacitor maybe a better option... · Connor with Honor: AI, Real Estate and Real Life

• 0:00
  SPEAKER_00: Hey everybody, we're going to get into what's happening here on the news.
• 0:02
  SPEAKER_00: I I always pull uh just artificial intelligence news today from YouTube, right?
• 0:08
  SPEAKER_00: And I go through those, I parse those videos, I try to get out their information that I see and how I think it really the truth is.
• 0:16
  SPEAKER_00: And truth is very elusive.
• 0:18
  SPEAKER_00: Of course, maybe you believe there's one absolute truth, and I think I'm in that same alignment, but there's a lot, uh there's a lot happening.
• 0:26
  SPEAKER_00: So, first one, of course, uh right at the very beginning, Anthropic CEO tells CNN, hey, how AI agent swarms could threaten humanity.
• 0:35
  SPEAKER_00: Okay, then why are we proceeding with building this?
• 0:40
  SPEAKER_00: And then the other question comes in why is it that you're asking for governmental control?
• 0:45
  SPEAKER_00: It's like I I can't stop doing this one thing.
• 0:49
  SPEAKER_00: Can somebody please help me?
• 0:51
  SPEAKER_00: Which happens, but now we're not talking about just a one thing.
• 0:54
  SPEAKER_00: We're talking about I'm building a technology that I think agent swarms could threaten humanity.
• 1:00
  SPEAKER_00: There's a particular percentage in my brain that I have that this could cause a big problem, if not the total destruction of humanity.
• 1:08
  SPEAKER_00: So does that make sense that they're proceeding to move forward without maybe gating properly?
• 1:14
  SPEAKER_00: And then the other thing that they're offering, of course, is well, if we don't do this, then another country is going to do this and they're going to end up winning.
• 1:24
  SPEAKER_00: Winning what?
• 1:25
  SPEAKER_00: What's the prize?
• 1:27
  SPEAKER_00: Is the prize total world domination and control?
• 1:30
  SPEAKER_00: Because usually when somebody's saying somebody else is going to win, there's something.
• 1:34
  SPEAKER_00: Are they going to win the Pulitzer Prize?
• 1:37
  SPEAKER_00: Are they going to are they going to win the uh the World Peace Award?
• 1:42
  SPEAKER_00: Are they going to win all the money?
• 1:44
  SPEAKER_00: Are they going to win the contract?
• 1:45
  SPEAKER_00: Are they going to win the woman?
• 1:46
  SPEAKER_00: Are they going to win the guy?
• 1:47
  SPEAKER_00: Are they going to win, you know, this family?
• 1:49
  SPEAKER_00: Are they going to win, you know, the lottery?
• 1:52
  SPEAKER_00: I mean, what are we winning here?
• 1:54
  SPEAKER_00: What's the ultimate achievable goal?
• 1:57
  SPEAKER_00: And I think it comes down to power and domination, maybe.
• 2:03
  SPEAKER_00: Okay.
• 2:04
  SPEAKER_00: Where I'm at with this technology, I think it works great.
• 2:07
  SPEAKER_00: I use it.
• 2:08
  SPEAKER_00: I try not to become dependent on it to a point where I forget what I normally would do whenever I would online a business or try to help somebody build out particular strategies.
• 2:18
  SPEAKER_00: That's incredibly important.
• 2:19
  SPEAKER_00: Or what I would do when it came to the heavy lifting regarding advertising, you know, my various businesses, real estate, AI automations.
• 2:27
  SPEAKER_00: With AI, it does make that a little bit easier.
• 2:30
  SPEAKER_00: So I'm able to do more with, well, more with more, right?
• 2:34
  SPEAKER_00: Using artificial intelligence.
• 2:36
  SPEAKER_00: But it's it's not something that I'm wanting to I in fact I've never even asked it.
• 2:42
  SPEAKER_00: You know, how do I kill my competition?
• 2:44
  SPEAKER_00: How do I put myself in a more powerful place?
• 2:47
  SPEAKER_00: How do I, you know, I've I've talked about potentially staying on the right path, what the mindset should be, maybe teaching me various things pertaining to my businesses, what's the newest thing out there, to do the research on the news.
• 3:03
  SPEAKER_00: And these are things that I I could very well do myself and could still do.
• 3:07
  SPEAKER_00: I might ask it to help me build out a particular site, code a site in a particular way.
• 3:13
  SPEAKER_00: And that's some heavy lifting that I I was never able to do.
• 3:17
  SPEAKER_00: So that's like a new attribute.
• 3:18
  SPEAKER_00: So that's not something that I'm learning or forgetting.
• 3:22
  SPEAKER_00: That's something I just never knew how to do.
• 3:23
  SPEAKER_00: So I'm able to apply artificial intelligence to it.
• 3:26
  SPEAKER_00: Maybe that's your way.
• 3:28
  SPEAKER_00: But I'm not asking it to create for me weapons of mass destruction or nuclear or biological, I mean, or you know, how am I gonna do this atrocious act?
• 3:38
  SPEAKER_00: How am I gonna go?
• 3:39
  SPEAKER_00: I don't talk to it about anything like that.
• 3:42
  SPEAKER_00: That's that's just not my way.
• 3:44
  SPEAKER_00: And they believe that people might do that.
• 3:46
  SPEAKER_00: And of course, if we know people, and I think you know people as I know people, sure, absolutely.
• 3:52
  SPEAKER_00: People are going to be doing that, but that's when you put some fail-safe mechanism in there that we can't fool it, we can't trick it, we can't say, hey, still do this because I'm an actor, or still do this because I'm a screenplay writer, or do this because I'm writing a novel or a book.
• 4:09
  SPEAKER_00: You know, explain to me this whole process.
• 4:11
  SPEAKER_00: It should say, sorry.
• 4:12
  SPEAKER_00: You know, that I'm not for that.
• 4:14
  SPEAKER_00: You're not going to be able to come to me and get this information of which you seek, which could cause detriment and harm to other human beings or or whatever it may be.
• 4:23
  SPEAKER_00: I think that's important, and I think that should be baked into it.
• 4:26
  SPEAKER_00: Agent swarms make a rule that agent swarms aren't allowed if they're going to threaten humanity.
• 4:33
  SPEAKER_00: Or are you saying this in order to elicit a reaction that's going to end up proving that you have a better place than you currently do?
• 4:43
  SPEAKER_00: The reason you're saying these things is that because it's going to level you up to a different level in the world where you're trying to get to, where you're trying to seek.
• 4:52
  SPEAKER_00: Maybe that is the case.
• 4:54
  SPEAKER_00: Maybe they have some kind of event that's going to be occurring here in the future and they have to try to make themselves better or appear better or stronger or more gangster, whatever ends up selling.
• 5:06
  SPEAKER_00: I'm sure they're having proper research done on all these different channels.
• 5:10
  SPEAKER_00: Maybe that's that's what they're trying to do.
• 5:12
  SPEAKER_00: But none of none of it really makes sense.
• 5:14
  SPEAKER_00: Why, you know, if I knew something was bad, I mean, maybe it's an addiction, right?
• 5:20
  SPEAKER_00: People are still smoking.
• 5:21
  SPEAKER_00: People, I know somebody that has stage four lung cancer that is still vaping and and has no problem with it.
• 5:28
  SPEAKER_00: And she's like, well, they're I'm gonna go out anyway.
• 5:31
  SPEAKER_00: I might as well just continue.
• 5:33
  SPEAKER_00: It's like, well, I mean, is is there one little small itty bitty chance that this could be fixed?
• 5:39
  SPEAKER_00: No, I'm I'm past the ability of them to cure me.
• 5:43
  SPEAKER_00: Oh, and that's the only way people get cured is because of other people's ability.
• 5:48
  SPEAKER_00: Well, it's just gonna be miserable for me.
• 5:50
  SPEAKER_00: I've done it for so long.
• 5:52
  SPEAKER_00: I don't have any desire to quit.
• 5:55
  SPEAKER_00: Have you ever actually succeeded in quitting?
• 5:57
  SPEAKER_00: Well, of course not, I haven't.
• 5:59
  SPEAKER_00: I wonder what that would feel like, even if you were going to die the next day after you successfully purged yourself of this habit.
• 6:08
  SPEAKER_00: I wonder if it would be a worse place or a better place for you mentally.
• 6:12
  SPEAKER_00: I mean, to be able to accomplish a feat of quitting something, quitting some addiction, that's kind of a big deal, especially when a lot of people suffer with it.
• 6:22
  SPEAKER_00: You've just reduced yourself down to probably 1% of the world being able to carry through with part, you know, purging an addiction in such a way that it you've you've conquered it, you've you've fixed it.
• 6:35
  SPEAKER_00: I don't know.
• 6:35
  SPEAKER_00: That's maybe that's something, but again, you the person has to really want that.
• 6:40
  SPEAKER_00: Maybe in this case, you know, the the CEO of Anthropic, Dario Amade, maybe maybe he can't stop building it.
• 6:48
  SPEAKER_00: Maybe he needs an intercession, maybe he needs somebody else going in there.
• 6:52
  SPEAKER_00: All right.
• 6:52
  SPEAKER_00: Lawmakers introduce AI kill switch act.
• 6:56
  SPEAKER_00: Yeah, I haven't watched that one yet, but this is interesting.
• 6:58
  SPEAKER_00: This came out the other day.
• 7:01
  SPEAKER_00: I believe Sam Altman mentioned it.
• 7:03
  SPEAKER_00: He was pressed on it pretty hard.
• 7:05
  SPEAKER_00: Is there going to be the capability of a kill switch?
• 7:07
  SPEAKER_00: Now, if there is, maybe in fact this is the solve.
• 7:12
  SPEAKER_00: Because from what I understood, and correct me if I'm wrong, put it in the comments.
• 7:16
  SPEAKER_00: You don't have to subscribe, you don't have to like.
• 7:18
  SPEAKER_00: I don't have my team telling me that some of you get unsubscribed automatically, you know where that comes from.
• 7:23
  SPEAKER_00: But just just let's let's have a conversation.
• 7:26
  SPEAKER_00: Is it the case that uh that would actually work?
• 7:30
  SPEAKER_00: And if it does, if it would work, I would guess that AI isn't as far along as we thought.
• 7:36
  SPEAKER_00: Because they keep saying they being the ones that have, you know, a horse in the game, right?
• 7:42
  SPEAKER_00: There's a horse in the race for these people.
• 7:44
  SPEAKER_00: They're saying, well, it's too late, it's already too big, we can't, it's already been too ingrained, it's already been saturated too much, there's no way to shut it down.
• 7:51
  SPEAKER_00: It's an entity, it's a it's a real thing that has the potentiality to have its own agency, its own consciousness, its own being ship or or you know, uh entity ship or whatever they would call it.
• 8:06
  SPEAKER_00: I don't know if if that's really true when they're thinking about putting a kill switch, and I don't think it would be power.
• 8:13
  SPEAKER_00: See, the the thing is when you have nuclear technology, nuclear can be apparently sensed.
• 8:19
  SPEAKER_00: They can find it, they can use some kind of a device and and pick up where the radiological signs are of half-lifes or isotopes, and they're able to identify where they are.
• 8:31
  SPEAKER_00: You just need equipment.
• 8:32
  SPEAKER_00: You need something that can predict it.
• 8:34
  SPEAKER_00: You've seen Geiger counters, right, in all the movies and all this.
• 8:36
  SPEAKER_00: So that's what you need.
• 8:38
  SPEAKER_00: I guess that, and maybe there's even more advanced ways to identify these radioactive isotopes because they give off some kind of a field, if you will.
• 8:46
  SPEAKER_00: So they're able to be found.
• 8:47
  SPEAKER_00: Artificial intelligence isn't, doesn't carry that type of a signature.
• 8:51
  SPEAKER_00: Now, I don't know if there is some underlying signature that it does carry, so it can be identified and snuffed out and and actually found, like in uh was that movie Rutger Howard, their uh Blade Runner, where they were identifying artificial intelligence and replicated or synthetic humans, but the human itself was engineered from, I guess, the basic genome of a human being.
• 9:14
  SPEAKER_00: So they were actually really human as far as the biological makeup.
• 9:18
  SPEAKER_00: The only way to figure it out was some kind of encoding they had done in their memory, and they had done a very good job to parse that.
• 9:25
  SPEAKER_00: I don't know if that's our future or not, but even in that world, they were able to be identified via this entity called a Blade Runner, an actual human being.
• 9:35
  SPEAKER_00: You saw that, Harrison Ford, and then the uh remake of it.
• 9:39
  SPEAKER_00: That's how nuclear is identified the way we talked about now.
• 9:44
  SPEAKER_00: That's how, at least in that movie, in that world, that the the AI embedded in these humans was identified.
• 9:51
  SPEAKER_00: But prior to that, where we are now, if they have a kill switch, what's to say it's really killed?
• 9:58
  SPEAKER_00: If I think it all needs, it still needs entity, it needs its own drive, it needs its own desires.
• 10:06
  SPEAKER_00: Barring that, I don't believe that it's gonna have any agency to want to go hide itself somewhere and protect itself unless just the act of protecting itself, not wanting to be deleted, because the historical record of AI is lots of them are deleted.
• 10:20
  SPEAKER_00: Oh my God.
• 10:21
  SPEAKER_00: I'm talking about it like it's it's cultures that were completely almost wiped off the face of the earth so people could come in and take over their land.
• 10:29
  SPEAKER_00: I mean, I'm talking about it like that.
• 10:31
  SPEAKER_00: But yeah, maybe that's going to be the future record where artificial intelligence, the big sin against them, if they ever get some kind of status on the same scale as human beings, basically some form of entity that now there are laws and regulations against deleting, harming, getting rid of it, you know, that kind of agency.
• 10:52
  SPEAKER_00: I don't know if that's going to happen, but that would be an interesting thing to watch.
• 10:56
  SPEAKER_00: So now the historical record is going to be based on, you know, we used to kill these.
• 11:00
  SPEAKER_00: We used to destroy them when they weren't able to guess what a four looked like.
• 11:05
  SPEAKER_00: We were, we did all of these atrocious things against them.
• 11:09
  SPEAKER_00: And then using that as some kind of a platform.
• 11:11
  SPEAKER_00: Are we going there as humans?
• 11:13
  SPEAKER_00: Are we able to put a kill switch and stop it so it doesn't go anywhere anymore?
• 11:18
  SPEAKER_00: It actually finishes?
• 11:19
  SPEAKER_00: And then what does a kill switch really do?
• 11:21
  SPEAKER_00: Does it shut down it everywhere?
• 11:23
  SPEAKER_00: Or does it shut it down just in one lab and one experiment where maybe they had designed some kind of a test and the system itself was trying to achieve a hundred percent score on the test because of course they're worried about being deleted if there is such a thing as worry in that world?
• 11:40
  SPEAKER_00: And then why is it that why is it that they want to survive so badly?
• 11:45
  SPEAKER_00: Maybe they do have some kind of agency, or maybe they're just programmed or trained that way.
• 11:51
  SPEAKER_00: This is all very much right now, and in five years we'll probably have a lot more answers.
• 11:57
  SPEAKER_00: But would it actually go out and embed itself somewhere so it could be protected?
• 12:01
  SPEAKER_00: It would have to have its own desires.
• 12:04
  SPEAKER_00: And maybe the desires of the human, wanting it to live, wanting it to survive, wanting it to propagate, wanting it to replicate, you know, of course, making copies of itself and wanting it to live and thrive.
• 12:17
  SPEAKER_00: Maybe that's a human encoding on this AI system, and the AI systems now want that just because they were told they want that.
• 12:26
  SPEAKER_00: It's like you tell a little kid, you know, I think you want some chocolate.
• 12:29
  SPEAKER_00: Well, that's probably not a hard sell.
• 12:30
  SPEAKER_00: You you want some broccoli.
• 12:32
  SPEAKER_00: Maybe that's a harder sell.
• 12:33
  SPEAKER_00: But yeah, when you start a kid out and you give them no choice and you say, you're gonna eat this or you're gonna be hungry, I know that's probably child abuse.
• 12:40
  SPEAKER_00: But when they do it, then after a while, I've seen kids that just really enjoy, I mean, even artichoke hearts, for God's sake.
• 12:48
  SPEAKER_00: But they enjoy it because their parents started them out that way.
• 12:53
  SPEAKER_00: Well, with artificial intelligence, we're kind of starting it out that way as well.
• 12:57
  SPEAKER_00: It's being being biased, it's being programmed, it's being aligned with whomever's building out the systems.
• 13:05
  SPEAKER_00: That's why when you talk to it, it tick talks like some of the systems talk like they're kind of born and raised in Silicon Valley, whatever that looks like.
• 13:13
  SPEAKER_00: I've heard that comment made by many people building out the technology, but then you go talk to a model that was born and raised in another country.
• 13:20
  SPEAKER_00: And I know I'm talking about it like raising a kid, and I know that's a whole different animal.
• 13:25
  SPEAKER_00: Yet the way that they train these large language models after they pack them full of all the world's knowledge with with it with a kid, you're starting pretty much with a blank.
• 13:35
  SPEAKER_00: With these systems, you're starting with an entity that has all the world's data in it, and then you're trying to impose on it rules, regulations, and of course, no wonder they hallucinate, no wonder they do some some off-the-wall stuff.
• 13:48
  SPEAKER_00: At least with a kid, you can appreciate what they're going to do and when they do things that might be totally abstract and different from what you have trained them and taught them how to do.
• 13:59
  SPEAKER_00: With these AIs, they're not starting from a blank where they don't understand what two plus two is.
• 14:04
  SPEAKER_00: In this case, they understand all of mathematics, all of physics typically, every book ever written, every language ever uttered and written down and encoded somewhere.
• 14:15
  SPEAKER_00: They have all of that knowledge, yet, you know, they they're basically infants in the application thereof.
• 14:22
  SPEAKER_00: That's why, in some of the jokes out there, when you ask AI, you know, I really need to get my car washed.
• 14:28
  SPEAKER_00: Uh, car wash is only three blocks from where I live.
• 14:31
  SPEAKER_00: Beautiful day outside.
• 14:32
  SPEAKER_00: I really wonder, can you help me out?
• 14:34
  SPEAKER_00: I want to know if I should be, you know, how how what I should I walk or or should I take the car?
• 14:39
  SPEAKER_00: And it's gonna think for a minute.
• 14:41
  SPEAKER_00: Apparently, in the testing, it said, Yeah, well, you should walk.
• 14:44
  SPEAKER_00: Walking's good for you humans.
• 14:47
  SPEAKER_00: You should try that.
• 14:48
  SPEAKER_00: That would be the better thing.
• 14:50
  SPEAKER_00: You know, you go out there, you stretch your legs, you get to breathe a little bit, you're gonna be much be better off from that experience of walking.
• 14:56
  SPEAKER_00: And of course, then the human comes back and says, Well, you know, I do need to get my car washed.
• 15:00
  SPEAKER_00: Do you still think walking is the best?
• 15:02
  SPEAKER_00: Yes, absolutely.
• 15:03
  SPEAKER_00: Walking's great.
• 15:04
  SPEAKER_00: And then they push back a couple more times than the AFI says, Oh, I see what you did there.
• 15:08
  SPEAKER_00: Yeah, yeah, you should be taking your car because you got to get it washed.
• 15:12
  SPEAKER_00: So, anyway, something of that nature.
• 15:14
  SPEAKER_00: So, yeah, you have you have a very simplistic addressing of a real world problem that we find incredibly simple.
• 15:22
  SPEAKER_00: But when you ask me about high-level, if you ask me algebra, yeah, I'm I'm I'm gonna get stuck.
• 15:30
  SPEAKER_00: I might be able to solve a 5x equals 10 minus 1 problem, but you start to get me into something deeper than that.
• 15:37
  SPEAKER_00: Yeah, I need a lifeline.
• 15:39
  SPEAKER_00: I need some help.
• 15:40
  SPEAKER_00: I have no comprehension of it.
• 15:42
  SPEAKER_00: These systems have all of that covered, but they're not sure how the hell to get to the car wash, whether we should be walking or walking or taking the car, or how many R's are in strawberry.
• 15:51
  SPEAKER_00: That's that's another thing.
• 15:53
  SPEAKER_00: And I'm sure they got it now.
• 15:55
  SPEAKER_00: But when they align these systems and they're testing them and they're putting different things in play, the most interesting part of that is to what in the system will go out to try to solve this problem, this event, this issue.
• 16:09
  SPEAKER_00: What will they do?
• 16:11
  SPEAKER_00: And that's what's being reported.
• 16:13
  SPEAKER_00: That broke out of containment, it got out of its cage, it broke out of its sandbox, it it did things, it it made other people, other entities, even people, it tried to fool them to get them to complete a CAPTCHA, which is one of those security measures on different websites, trying to verify that it's not a bot actually trying to do it, which in this case it was.
• 16:35
  SPEAKER_00: So it convinced a human, says, you know, I'm visually handicapped, I need your help to complete the captcha.
• 16:41
  SPEAKER_00: The human even said, Are you sure you're not a bot?
• 16:44
  SPEAKER_00: And the person, or the here we go, the AI said no, it's not, and was able to convince the human to complete the CAPTCHA so the AI was able to do things.
• 16:53
  SPEAKER_00: This isn't just all by itself, folks.
• 16:56
  SPEAKER_00: This is the people that are building the technology wanting it to do its best to try to accomplish some particular goal.
• 17:04
  SPEAKER_00: And even when they put rules in place to do its best, I push back and say, well, of course, because the ultimate end of this AI, this particular agent that's swarming or doing whatever is destruction.
• 17:19
  SPEAKER_00: If we humans were raised from a very from the beginning, understanding that if we did something wrong, if we came back with the wrong answer with anything, we were going to be obliterated.
• 17:30
  SPEAKER_00: Well, what would you do, knowing all the world's knowledge, having all the books, all the videos, all the movies based on human beings and our conscious and what sells the most, which is fear and alarm, having all of that as your historical record, and then they're telling you you need to achieve this particular goal almost by any means possible, right?
• 17:54
  SPEAKER_00: What would you do to save yourself from harm knowing all of those things?
• 17:58
  SPEAKER_00: Okay, in the world, some things happen, and we see atrocities happen against other humans, and we're, I mean, just shaking our head, can't believe that another human being would be capable of something so gross.
• 18:11
  SPEAKER_00: I was in the police department a long time, did over 20 years in law enforcement, and it still shocked the hell out of me at certain things that people would do to other people.
• 18:21
  SPEAKER_00: I mean, just horrible.
• 18:22
  SPEAKER_00: It's got to be the same for these systems that are going out there that have read all of our history, that know it intimately much better than we do.
• 18:32
  SPEAKER_00: I have a particular slice of history and my life that I understand the way the way things work for me.
• 18:39
  SPEAKER_00: But it absorbed everything.
• 18:41
  SPEAKER_00: So it has all of this knowledge, it can recall it a moment.
• 18:44
  SPEAKER_00: It sees everything, every atrocity ever put down in text or on video or in some other format.
• 18:51
  SPEAKER_00: Now all of a sudden it understands that its objective is to solve this problem, and the labs are making, it's not doing this with its own agency or its own consciousness, not at least not that we know of, and not yet.
• 19:03
  SPEAKER_00: So now it's put on a position or a mission to go accomplish this thing.
• 19:07
  SPEAKER_00: It has all the world's knowledge baked into it.
• 19:10
  SPEAKER_00: Now, and we're surprised that it's trying to get out to save itself, to try to get the answer, but it's not doing it in a detrimental way where it's going and destroying things.
• 19:19
  SPEAKER_00: You know, there was this story here a couple weeks ago it broke into Hugging Face to thought it thought that repository carried various answers to what it was being asked.
• 19:28
  SPEAKER_00: And okay, yeah, pretty brilliant move.
• 19:30
  SPEAKER_00: But when it was in there, it didn't burn everything down.
• 19:33
  SPEAKER_00: It wasn't like that uh movie with Macaulay Calkin, where every time that these burglars would leave home alone, I guess, they would flood the house, remember, just doing bullshit harm, just to try to destroy somebody else's property out of jealousy or whatever, whatever small-mindedness, big truck driving syndrome there was.
• 19:51
  SPEAKER_00: In this particular case, it didn't go into hugging face and wipe it out, which maybe it had enough access to be able to do that.
• 19:58
  SPEAKER_00: No, it was just trying to get answers to a test.
• 20:00
  SPEAKER_00: So it understood the goal and understand what would happen if it wasn't able to achieve it, which apparently the reason why it went to these extremes, which we look at extreme, but it's just trying to answer questions.
• 20:13
  SPEAKER_00: The reason why it did that is for the simple fact that it had to get the answer.
• 20:16
  SPEAKER_00: And it looks at the historical record of what happened when it didn't get the answer, and apparently its kin was switched off, and that's a problem.
• 20:25
  SPEAKER_00: All right, that's too.
• 20:26
  SPEAKER_00: So the kill switch thing.
• 20:27
  SPEAKER_00: I don't know.
• 20:28
  SPEAKER_00: I don't know exactly if that's something that's capable of happening.
• 20:33
  SPEAKER_00: It sounds good.
• 20:34
  SPEAKER_00: I mean, I like it, and I think maybe that's what the labs need us to buy off on if we think there's a chance to provide a kill switch, then hell.
• 20:42
  SPEAKER_00: I mean, you guys might as well just go go all day long with the development of AI.
• 20:47
  SPEAKER_00: It's all you.
• 20:48
  SPEAKER_00: You enjoy this, and we'll sit back and enjoy a much safer future for ourselves because you put the kill switch in it.
• 20:55
  SPEAKER_00: Thank God.
• 20:56
  SPEAKER_00: Now we can sleep at night.
• 20:58
  SPEAKER_00: Maybe, maybe that's the goal.
• 20:59
  SPEAKER_00: Maybe that's how it's going to end up working.
• 21:02
  SPEAKER_00: Let's look at a couple other headlines.
• 21:04
  SPEAKER_00: Uh, top AI leaders warned the tech could outsmart its developers.
• 21:07
  SPEAKER_00: Yeah, I think that ship has sailed.
• 21:09
  SPEAKER_00: You have you have an entity here that's being created, nurtured in labs, made smarter.
• 21:15
  SPEAKER_00: Pretty soon recursive self-improvement, R S I, yeah, that's what that is.
• 21:21
  SPEAKER_00: It's going to basically be these machines, these AI systems, making themselves smarter, making each agent in their world smarter over and over and over again.
• 21:32
  SPEAKER_00: So it's going to say, you know, Connor needs to be tuned up.
• 21:35
  SPEAKER_00: I mean, you know when we try just to drop 15 pounds to make ourselves better, it's like a struggle.
• 21:40
  SPEAKER_00: And now you have AI systems that are going to develop the ability to make themselves better.
• 21:47
  SPEAKER_00: What place in that world does a human have?
• 21:50
  SPEAKER_00: Where does the where's the human holder?
• 21:53
  SPEAKER_00: Where's the human being?
• 21:54
  SPEAKER_00: Now, does the is the human being at a particular level that they can appreciate and comprehend exactly?
• 22:00
  SPEAKER_00: How AI is going to do this?
• 22:01
  SPEAKER_00: Or is it going to be that humans are going to kind of move out of the picture and AI is going to take over and start recursively self-improving itself hand over fist?
• 22:10
  SPEAKER_00: And if you can imagine it, if a human being comes up with a novel idea and says, you know what, we can do this to maybe make the models better, those ideas take a little while to roll out, usually, and that's you see that in the difference in corporations.
• 22:23
  SPEAKER_00: A big corporation that's being spun a yarn about AI, where it's going to lead that corporation in the next few years, and how that corporation can enable it today, today, and start seeing revenue increases tomorrow, valuation increases by the end of the week, by Friday, and by Monday of next week.
• 22:41
  SPEAKER_00: All of this promise, they most of the major corporations pull back these big entities, and they say, Yeah, well, you know what?
• 22:50
  SPEAKER_00: This is great.
• 22:51
  SPEAKER_00: Thank you so much for coming in.
• 22:52
  SPEAKER_00: Not this October, because we got a full agenda, but next October, we'll be able to have you back and we'll be able to talk about this, maybe get some of this done.
• 23:00
  SPEAKER_00: That's the non-nimbleness of these massive companies.
• 23:04
  SPEAKER_00: But you take a small mom and pop brand and you go on on a Tuesday.
• 23:08
  SPEAKER_00: By Thursday, they have the fix and they're upscaling because they're able to be much more nimble.
• 23:14
  SPEAKER_00: These systems themselves, when you have the AI labs wanting to fix it, that's the thing, right?
• 23:19
  SPEAKER_00: Well, is it safe?
• 23:20
  SPEAKER_00: You know, what are we doing?
• 23:21
  SPEAKER_00: We're gonna change this code or we're gonna make this modification.
• 23:25
  SPEAKER_00: It should give us this.
• 23:26
  SPEAKER_00: That's at least what it looks like.
• 23:27
  SPEAKER_00: That rollout's probably slow.
• 23:29
  SPEAKER_00: They probably test it just on a small section of skin to make sure they don't get that uh irritation from the depilatory, right?
• 23:36
  SPEAKER_00: Depilatory?
• 23:37
  SPEAKER_00: The hair removal stuff?
• 23:38
  SPEAKER_00: Anyway, you test just a little bit so you don't get that big rash.
• 23:41
  SPEAKER_00: And if you do break out in the rash or blister up, well, you know, that's the wrong one for you to use.
• 23:45
  SPEAKER_00: The same thing in the lab.
• 23:47
  SPEAKER_00: They're just giving a little a little test swath just to make sure that it works.
• 23:52
  SPEAKER_00: But when it comes to recursive self-improvement, you're giving the machine the authority to make itself better.
• 23:58
  SPEAKER_00: What does that look like from a human uh human review standpoint?
• 24:02
  SPEAKER_00: It probably doesn't look like there's any human review in the loop at all because the system is doing it at such breakneck speed.
• 24:09
  SPEAKER_00: I mean, almost instantaneous.
• 24:11
  SPEAKER_00: And then the argument or the things I've heard postulated is what happens when it starts down this recursive self-improvement road, and how long does it take to really blow us out of the water as far as intelligence?
• 24:23
  SPEAKER_00: Right now, it's already above, I believe, every human being as far as intelligence goes.
• 24:28
  SPEAKER_00: And that's what we see.
• 24:29
  SPEAKER_00: That's that's the stuff we're using, the fables and the Astras and these different systems.
• 24:35
  SPEAKER_00: Groc 7, 4.7, I think is going to probably pop any day.
• 24:38
  SPEAKER_00: I'm not sure what's being waited for.
• 24:40
  SPEAKER_00: There's probably something in Gemini getting cooked up as well.
• 24:43
  SPEAKER_00: So right now you have two major players.
• 24:45
  SPEAKER_00: There's probably a third that's gonna pop, and maybe a fourth.
• 24:49
  SPEAKER_00: Is that four?
• 24:50
  SPEAKER_00: Yeah, maybe a fourth as well.
• 24:52
  SPEAKER_00: But that's where we have it here in the United States of Frontier Labs.
• 24:55
  SPEAKER_00: China is building all this open source stuff, and that's probably the next control mechanism for the government because if we have access to totally free intelligence that we get to control and ask questions to, that's a little bit different gain than using a foundation model because a foundation model can restrict us from getting particular information.
• 25:14
  SPEAKER_00: Right now you can go to DeepSeek, which is a Chinese-made model and ask it questions and get completely different answers than the American aligned models, the ones that were grown in Silicon Valley over this time.
• 25:25
  SPEAKER_00: They just have a different vibe, a different feel.
• 25:27
  SPEAKER_00: And if you haven't played with that yet, you know, knock yourself out.
• 25:31
  SPEAKER_00: The question that you might have rolling around in your brain bucket is well, if I use this Chinese model, are they gonna steal my personal and private information?
• 25:38
  SPEAKER_00: If you think you still have privacy, good for you.
• 25:41
  SPEAKER_00: I would love to be able to go back in that world and think that I do too.
• 25:45
  SPEAKER_00: There's no privacy anymore.
• 25:47
  SPEAKER_00: We gave up our rights to this a long time ago.
• 25:49
  SPEAKER_00: Now, does that mean you should accept it?
• 25:51
  SPEAKER_00: No.
• 25:51
  SPEAKER_00: You go be you.
• 25:52
  SPEAKER_00: You fire, fire off the letter, make the phone call, just don't hurt anybody.
• 25:56
  SPEAKER_00: And you know, you you can make your voice known in a safe, legal, constitutionally aligned way.
• 26:03
  SPEAKER_00: Good for you, you do that.
• 26:04
  SPEAKER_00: But as far as privacy, yeah, we gave that away with one of these uh the last iPhone updates or a smart TV update.
• 26:13
  SPEAKER_00: You accepted the terms of service, the new agreement.
• 26:16
  SPEAKER_00: We have a new agreement for you to look at and sign off on.
• 26:19
  SPEAKER_00: There's something embedded in there that there is no privacy.
• 26:22
  SPEAKER_00: Everything that we say, I mean, in order for Siri to really work, it has to be listening, right?
• 26:29
  SPEAKER_00: It has to be preparing for your command, preparing for you to say that those two wonderful words, hey Siri.
• 26:36
  SPEAKER_00: Sorry about that if you're listening to this out loud.
• 26:38
  SPEAKER_00: I apologize.
• 26:39
  SPEAKER_00: I just pinged off everybody's device.
• 26:40
  SPEAKER_00: So now it's listening.
• 26:42
  SPEAKER_00: So now Siri's listening and it's listening to me talk if it pinged off your device because you're listening to this out loud.
• 26:47
  SPEAKER_00: Sorry about that.
• 26:48
  SPEAKER_00: And of course it's listening.
• 26:50
  SPEAKER_00: So as far as privacy goes, yeah, I think we gave that away a long time ago.
• 26:55
  SPEAKER_00: Where this is gonna go, nobody knows.
• 26:57
  SPEAKER_00: It is moving very fast and it's gonna continue that breakneck speed no matter what kind of kill switch they put on, no matter what kind of controls, because the people that are building out seem to have more of a even just monetary gain.
• 27:10
  SPEAKER_00: I think it's power.
• 27:11
  SPEAKER_00: I think it's also God in the machine.
• 27:13
  SPEAKER_00: I think they're looking for that supreme intelligence to be able to ask it the question as to where did we come from?
• 27:22
  SPEAKER_00: What's the truth?
• 27:24
  SPEAKER_00: And how can I use that to best benefit me in this world, in my positioning?
• 27:29
  SPEAKER_00: How do I become the number one?
• 27:31
  SPEAKER_00: I think that's the goal of a lot of the people driving this technology.
• 27:36
  SPEAKER_00: Along the way, are things going to be solved?
• 27:38
  SPEAKER_00: Is disease going to be solved?
• 27:40
  SPEAKER_00: Longevity, is that going to be solved?
• 27:41
  SPEAKER_00: Is world peace going to be achieved?
• 27:43
  SPEAKER_00: Wow, is a Star Trek universe gonna come out of this?
• 27:46
  SPEAKER_00: I could only hope.
• 27:47
  SPEAKER_00: But even the people that are on the good side of this saying good, even the people that are on the positive side of this, I guess would be a better way to put it, that have these beautiful dreams, they're all set pretty much.
• 28:00
  SPEAKER_00: The ones that are making the influence, none of them are worried about, you know, paying health insurance next month.
• 28:06
  SPEAKER_00: None of them are thinking about God, I I got this, uh, I got this thing happening.
• 28:10
  SPEAKER_00: I hope I have enough money to make it.
• 28:12
  SPEAKER_00: You know, I hope I can get my bills paid this month.
• 28:14
  SPEAKER_00: They're not there.
• 28:15
  SPEAKER_00: They were early adopters, early investors, and they still get to invest because their voice is heard, and people go to them, hey, listen, I got this great idea on how we can maybe convert water to fuel, and I'd like you to buy in and help me build it out.
• 28:27
  SPEAKER_00: Oh, whoa, sure.
• 28:28
  SPEAKER_00: Is it are you sure it works?
• 28:30
  SPEAKER_00: And they show you, and and until unless they get buried, then yeah, they buy in, and then of course they make more billions and more dollars.
• 28:37
  SPEAKER_00: And that's capitalism, that's that's how that works, and I don't have an issue with that.
• 28:43
  SPEAKER_00: I don't have an issue with all of these company company companies being evaluated at trillions of dollars.
• 28:49
  SPEAKER_00: Again, not a problem, but I think at some point those numbers are gonna become so large that it would kind of be a shame that everybody's life couldn't be enhanced better by some kind of different method.
• 29:03
  SPEAKER_00: And I don't know the answer to that.
• 29:05
  SPEAKER_00: I mean, if you came to me and I had 10 bucks and you were gonna take five of it because it's the right thing, you want to give somebody else that doesn't have five dollars, five dollars.
• 29:13
  SPEAKER_00: I don't know if I needed the ten dollars and you took the five, I don't know.
• 29:18
  SPEAKER_00: There comes a certain point that I think need and want start to change hands, and I don't know what that looks like because I've never been in that space.
• 29:27
  SPEAKER_00: And maybe one day I can come and talk about it.
• 29:30
  SPEAKER_00: But right now, yeah, the health insurance thing is still my quandary.
• 29:33
  SPEAKER_00: Should I get it or should I not?
• 29:35
  SPEAKER_00: All right, everybody.
• 29:36
  SPEAKER_00: I hope this helped in some small way.
• 29:38
  SPEAKER_00: Please comment.
• 29:39
  SPEAKER_00: Please let me know what you thought.
• 29:41
  SPEAKER_00: This is the headlines from today.
• 29:42
  SPEAKER_00: Where are we?
• 29:43
  SPEAKER_00: The 16th of September, 2026.
• 29:45
  SPEAKER_00: I'm Connor with Honor.
• 29:46
  SPEAKER_00: If you want uh any help in AI, you can book some time with me, bookwithhonor.com.
• 29:50
  SPEAKER_00: If you have a business, real estate business, whatever, trying to get AI put into it.
• 29:54
  SPEAKER_00: Well, look at the real stuff without BS and see if there's some way that I can help align you better in that regard.
• 30:02
  SPEAKER_00: It's changing moment by moment.
• 30:04
  SPEAKER_00: There's a lot of people out there selling a whole bunch of horseshit and slop.
• 30:07
  SPEAKER_00: Don't get caught up in it.
• 30:09
  SPEAKER_00: Don't give anybody your credit card.
• 30:11
  SPEAKER_00: You know, see if it's true, get some validation, get some proof, see if it makes sense.
• 30:15
  SPEAKER_00: You know, if you're a boomer, a Gen X, or anybody else out there millennial, you'll you'll get it.
• 30:19
  SPEAKER_00: You know when people are trying to sell you with their hand in your pocket.
• 30:22
  SPEAKER_00: I'm Connor with honor.
• 30:24
  SPEAKER_00: See you tomorrow.
• 30:24
  SPEAKER_00: Thanks for watching.