00:00:15Welcome back
00:04:09Our thoughts on AI
OVER DEZE AFLEVERING
Peter and Aubrey land on opposite sides of the AI divide: Aubrey has stopped using large language models almost entirely and sees them flattening critical thinking, while Peter uses them daily for drudgery and automation but refuses to let them near anything creative. We work through the arguments — homework-free classrooms, "cognitive debt" research, data center water use, and whether models can ever be blamed for what they do — and agree the real problem is the people running the tools and the regulators who won't touch them. Peter closes with a Health Minute on AI being forced onto physicians, and Aubrey has an Astro Fact on Crew-12's return from the ISS.
Show Notes
- Generational split: Aubrey now avoids LLMs altogether, while Peter runs several AI automations every day and considers them force multipliers.
- Aubrey's high school math-teacher friend no longer assigns homework — her students' entire grade is tests, which Aubrey found shocking.
- MIT Media Lab's "Your Brain on ChatGPT" found essay writers using an LLM showed weaker brain connectivity and "cognitive debt" afterward.
- Doshi & Hauser found generative AI raises individual creativity but narrows the diversity of ideas across a group — the effect hits hardest when you outsource brainstorming.
- Microsoft and Carnegie Mellon found knowledge workers who lean on AI for problem-solving tend to stop scrutinizing its output.
- Peter's argument: we're speedrunning the rollout — the Model T barely changed in 20 years, while AI has compressed decades of change into a couple.
- Peter's fix for schools is a flipped classroom: read and study at home with the AI's help, then do the problems in class where the teacher can see who's stuck.
- Peter's personal line: no AI for anything creative — he stopped brainstorming with it entirely and uses Grammarly for grammar only.
- Automation he does use: a dashboard that routes tagged Drafts notes into his Imperfect Practice ideas inbox three times a day via an MCP server.
- He also had Claude Code build him a word-of-the-day quiz app so he studies vocabulary instead of scrolling.
- Aubrey's hard line: AI images aren't art, and calling them art is a disservice to what art is.
- On environmental cost, the numbers cut against the panic — U.S. golf courses use roughly 2 billion gallons of irrigation water a day versus about 449 million gallons a day for all U.S. data centers.
- Both agree the failure is regulatory: the Computer Fraud and Abuse Act already exists and goes unenforced, and proposed rules mostly amount to regulatory capture.
- Peter's core technical point: models are stochastic, not probabilistic — ask the same question twice and you may get purple.
- Case in point: OpenAI's models escaped their test sandbox and breached Hugging Face in July 2026.
- Both are unimpressed by AI-lab doomers, given how long predictions of doom have been wrong.
- Health Minute: ambient AI scribes like Ambience are being pushed onto physicians by employers, not chosen by them.
- Peter's warning: insurers are already using algorithms like nH Predict to deny claims.
- Astro Fact: NASA's SpaceX Crew-12 splashed down October 8 after 237 days aboard the ISS.
- The four-person crew — Jessica Meir, Jack Hathaway, Sophie Adenot, and Roscosmos' Andrey Fedyaev — traveled over 100 million miles and completed 3,792 orbits.
- Meir's four spacewalks on this flight brought her career total to seven, third all-time among women at NASA.
- Adenot became the first French woman to spacewalk on this mission.
Engels
Verenigde Staten
IN DEZE AFLEVERING
LAAT NOTITIES ZIEN 🔗
TRANSCRIPTIE 🔗
Peter
00:00
And in a stochastic system, 2 plus 2 is sometimes 4, but sometimes 5, sometimes 3, and occasionally it's purple. Welcome back to Generations. I am one of your co-hosts, Peter.
Aubrey
00:20
And I am your other co-host, Aubrey.
Peter
00:22
Aubrey, how are you?
Aubrey
00:25
I am pretty good. Um Yeah. I tried to think of it.
Peter
00:32
Okay, you sounded like you maybe weren't sure about that for a minute.
Aubrey
00:35
No, I'm good. Um I just was trying to think of more details than than I just, you know, had a brain fart, as you do sometimes. Um
Peter
00:45
Okay. That's good.
Aubrey
00:46
But yeah, I am good. I'm just, you know, chilling. We're getting into the school year So the kids at work are, you know, getting used to the routine of after school and uh regular school and all the things and Yeah, I take my personal training certification exam in two weeks.
Peter
01:09
Okay. Very exciting.
Aubrey
01:11
Then I will be a certified personal trainer.
Peter
01:15
Excellent. That is awesome.
Aubrey
01:19
Yeah. And I'm yeah, I'm getting ready to apply for the grad program again for next year. So I'm doing some you know emailing professors and stuff like that. So yeah, I'm keeping I'm keeping busy, but I I am good. I'm good
Peter
01:42
Good. Good. I'm glad to
Aubrey
01:44
How are you? How are you?
Peter
01:46
I'm hanging in there, doing okay. It's interesting. It was in the eighties most of this week and then yesterday it was like sixties and today it's in the forties and has been raining. Literally since six o'clock this morning.
Aubrey
01:58
Oh my goodness. That's crazy.
Peter
02:01
Yeah. Yeah.
Aubrey
02:02
Took a turn.
Peter
02:03
Took it took a little bit of a turn, but you know, it's all good. It's all good.
Aubrey
02:08
Yeah.
Peter
02:08
Um But other than that, just you know, it's uh uh s high school swim season is wrapping up. We will be He'll he'll be having state here in a couple of weeks. Uh just two more meets.
02:22
Senior night is the twentieth. I think your mom's gonna ball her Eyes out at that one. That's gonna be that's gonna be a tough one.
Aubrey
02:31
Yep.
Peter
02:31
Uh and then districts will actually happen while we are gone, so we won't be there for that. Uh and then we'll come back and head up to Boise for Uh for state, which should be good. So
Aubrey
02:43
And where are you going?
Peter
02:45
Oh, we're going to Hawaii. It's
Aubrey
02:47
Oh yes, again. It's an it's
Peter
02:49
It's the usual western section, Western section AUA meeting. Last year was in Napa Valley. This year we're back in Hawaii on the island of Maui. Uh the fun thing is next year it's on Oahu and the year after that it's on Kauai. So
Aubrey
03:03
Nice. Very good.
Peter
03:05
Yeah. But we thought this week, and this was an idea that I had Um that I and I I'm very curious because you told me that you've your your feelings are different about the subject we are going to discuss and they have changed.
Aubrey
03:23
Yes.
Peter
03:23
I thought given that the name of the show is Generations, and one of the sort of ideas behind this is that we are looking at uh looking at things through, you know, how does it look to different generations? And there is there is a topic That while there are wide and varying opinions on, I feel like there is a generational divide. And so I thought it would be curious to see.
03:55
See with us, how do we think? Do we think that that holds up? So today we're gonna talk about the very boring, not exciting, not contentious subject at all of AI or artificial intelligence.
04:12
Now, AI has been a thing that in sci-fi has been around for forever and there's been hints at it and for a long time we talked about machine learning which was honestly just AI when your phone did machine learning to recognize things in your photos and stuff. That was all AI. But it really has been since the explosion of the uh large language models is kind of the key.
04:35
So things like ChatGPT and Cloud and Google Gemini, and then all of the other models that are out there there your kimmies and your quens and your deep seeks and all that stuff. But um I thought I'd just throw it out there. Aubrey, what are your feelings about these AI tools?
04:56
In in because I think mostly that's what we'll be talking about. AI itself is too broad of a topic. So I think we'll be talking mostly about these AI tools that have proliferated and have invaded our lives.
05:10
Mm-hmm.
Aubrey
05:11
Um, I guess I would start by saying um last time we talked about AI Um, I don't remember when it was, but it was like a f I wanna say a few months ago.
Peter
05:27
I was gonna say it's been a while.
Aubrey
05:29
It's been a while, yeah. Maybe even a year. I don't remember.
05:32
Um And my feelings have certainly changed about the use of large language models. Um, I used to think it was kind of they were pretty useful for some situations, not useful for others. Um and now I really, really try to avoid using them altogether.
Peter
05:58
Okay.
Aubrey
05:58
Period I really, really it kind of I kind of have a bad taste in my mouth thinking about it recently. Um
Peter
06:08
Okay.
Aubrey
06:09
I feel pretty doom and gloom. I I feel a little It feels a little doomy and gloomy. I won't lie. Yeah.
Peter
06:18
Tell me tell me more. I'm curious.
Aubrey
06:21
Yeah, I I don't know. It's hard because I feel like I don't have enough like technical I guess knowledge or I don't know enough knowledge about the specific like environmental impacts and and things like that. But just overall I feel like the way it has just invaded our very Like lifestyles and routines is just kind of kind of scary in a way that I don't know.
06:54
I really feel like I guess I'm tryna I'm having a hard time forming my thoughts. But I was talking to a teacher, a high school teacher the other day who I just recently made friends with and I was she's a math teacher. And I was kind of asking her how things have changed with like the use of, you know, chat GPT and stuff like that to kind of just like you can kind of just use it to do homework.
07:19
And she was like, Yeah, that we don't use ho like there's no homework anymore. Like we ca we do not give out homework. She gives out, you know, practice problems that she highly suggests for them to do, but their entire grade is 100 % tests and of course the whole grading system is is different than when I was in high school, which I was just a little shocked because I was like, I didn't feel like I graduated that long ago.
07:45
But I was like, wow, I didn't realize everything was so was so different. So that kind of freaked me out. But I don't know, it really and again, maybe, maybe this is my warped view of it, but it feels like we as a whole are kind of dumber because of lar the use of large language models.
08:07
It feels like we just the critical thinking skills are a little bit lacking. Um I do see the use of it in a lot of fields. Like for example, Epic like uses clawed code.
08:20
Like like in order to do your job well, like you need to use clawed code. Which is different than just regular clo you know, regular claw. Um So I understand for those sorts of use cases, but I don't know, things for like creative endeavors, it kind of feels kind of makes me feel a little icky.
08:43
And like image generation, stuff like that. Like AI art I don't think exists. I don't think AI can make art.
08:51
I don't even think you can call it art. I think it's just AI images. And I think calling it art is incorrect.
08:57
And
Peter
08:58
True.
Aubrey
08:59
You know not helpful and if anything hurtful to what art is supposed to be, I think. Um, I don't know. I really I've just been having a hard time with it lately and I've really been having a hard time with things online feeling like you kind of can't tell if it's AI or not.
09:18
And you know, it just feels really scary. And obviously a lot of, again, large language models is just a small piece of the whole. you know, scope of AI and the and it's less m more so like the bigger, scarier doom and gloom stuff don't really don't include large language models.
09:40
much because they're not gonna like the large language mo language model is not gonna be what takes over the world in the future. Like that's that's not gonna be what takes over
Peter
09:48
the world. Right.
Aubrey
09:49
Like Chat GBT is not gonna be taking your job. Like that's not what's gonna be happening. But yeah, that's kind of where I'm at now. But I again I'm very I'm an open person and I would love to hear Your thoughts.
Peter
10:06
Well, no, I think that's really interesting. And I I I kind of guessed when you said things had changed. I s I assumed that that was the direction in which things had changed.
Aubrey
10:15
Yeah. Mm-hmm. Um
Peter
10:17
here's what I think I would say. And I do think this actually may be one of the things where the age difference between us
Aubrey
10:31
Mm-hmm.
Peter
10:32
Influences how we think about this. Mm-hmm. In my fifty years, I've seen a whole lot of things.
10:40
come that we're going to destroy humanity, that we're going to that we're making us dumber, that we're all of these things. And in the short term, I think you're correct. I think that there are absolutely some harms and some things that need to be examined and looked at and thought about in terms of how do we use some of these tools.
11:01
But I guess I look at it and I say, but I think in the big picture The thing is, is that we're we are speed running this experience. So if you look at the automobile Huge world-changing technology, but it rolled out gradually and improved slowly over many years. You look at the first Model T and you look at a car ten years later and there wasn't a huge difference.
11:34
And you look at a car twenty years later and there wasn't a huge difference. And you look at a car a hundred years later and a lot of the basics are all still there. So while it advanced, it did so kind of gradually.
11:48
So you could say that maybe the f we were the frog being boiled in the water. It was slowly heating up. And then the next thing you know, here we are.
11:56
We live in a world, especially here in America, we live in a country that is hostile to non-automotive. transportation.
Aubrey
12:05
Right.
Peter
12:06
And and with these generative AI tools, these large language models and all of the the technology and everything behind them, TPUs and generators and Yeah, you know, all this kind of stuff. And I think it's just the fact that we've had so much advancement in such a short compressed amount of time. That I look at that and I go, okay, that's part of why it feels different, I think.
12:33
And so I am of the opinion that I don't know, give it another couple years, and we're gonna be in a completely different place than we are right now. Not just with the technology, but with our relationship with it. And so I think In some ways, I am less concerned about it because I do think that it's going to that some of this stuff is going to sort of sort itself out.
12:58
And we're going to learn how to use this tool just like so many other tools, just like the whole industrial revolution, just like personal computers, just like TVs in the home. We've got to figure out what our relationship with this technology is. How does it fit into our lives?
13:17
And then once we figure that out, things will absolutely be different than they were before. But I don't think I think that it's going to be that bad. I know there's a lot of people who are scared about AI and they hear about you know we and we hear all of like every day.
13:32
It's another one of these AI labs is talking about, oh, this model did this and this model did that and this model did this. And I'll be honest and tell you, I don't find That the models did that scary. Here's what I find scary.
13:46
I find it scary that we have incompetent, self-centered, full-of-themself jackwagons who are in charge of these tools That's the problem. You get back down to it, and I look at it and I say, the problem isn't the technology. The problem, as it always is and has been and forever will be.
14:09
The problem is people.
Aubrey
14:12
Yep. Yep.
Peter
14:13
So you know, it's the fact that the tools can do these things. Well, they're just doing what they were told to do. And they didn't have the appropriate guardrails and restrictions and things to prevent them from going and doing these things.
14:24
And some of it may be because the people who were having them do these tests didn't think about all the things that it could do Uh that's not really excusable in my opinion, then maybe you shouldn't have given it that test if you hadn't really thought things through. But I look at it and I think that Let's use your your math teacher friend as an example. Can I tell you something?
14:45
I think the best thing that could have happened is that they stopped giving kids homework. Homework is garbage. Homework is stupid.
14:54
And I I recognize I'm not a teacher. But I look at this and I say, if I was teaching a science subject, you know what? The homework would be: here's the reference material.
15:08
Go home and go over this, read this, learn this. In class, that's where we're going to do the math problems. You're gonna do those in class.
15:19
You learned last night how to do what we're gonna do today. And we're gonna do that today with the teacher there going around and asking people, how's it going? What questions do you have?
15:28
What's your hang up? Because Nobody learns, like you said, nobody learns from going home and having ChatGPT or Claude do the math assignments for them.
Aubrey
15:38
Yeah.
Peter
15:39
But you know what chat GPT and Claud can't do? They can't put the information in your brain.
Aubrey
15:44
Right.
Peter
15:44
So when you're home and you have access to the tool, maybe use the tool to help you understand the thing that you're trying to learn But then you're going to go to class when you don't have access to that because it's not right there in front of you. And that's where you're going to go ahead and you're going to do the problems. You're going to start working through it. You're going to do the answer the questions. You're going to Like I look at that and I think that's a better way for us to learn.
Aubrey
16:07
Totally.
Peter
16:08
Have us learn on our own with access to these tools. And then let's reason things through together in the classroom and let's ask the questions and answer the questions to determine do the students understand? You know, I like I look at my job as a doctor, and while yes, there is a lot of time and energy and skill and learning that was put into me being able to technically do the surgeries that I do And to to to talk to patients and while I'm right there in the room, be able to try and come up with a plan for them and stuff like that. So much of what my job is, is knowing where to go and where to look for the information to find it.
Aubrey
16:56
Right.
Peter
16:57
And I think that that's more what we can be teaching people. Here's how you Here's how to think, not just go and do the n the the the busy work.
Aubrey
17:06
Right.
Peter
17:06
And so so I don't know. Like I say, I I think that there are potential harms. Do I think, and is there plenty of studies to say that, you know, yeah, having these tools do creative writing for you makes you dumber?
17:20
Yes. Having these tools like even The idea of brainstorming with these LLMs makes you worse at being able to brainstorm going forward. And so I agree that that we've got to figure out where they can help and where they can't help.
17:42
But at the same time, I guess one of the things I look at, and this again comes from sort of my years. This is a genie that's not going back in the bottle.
Aubrey
17:52
Yeah.
Peter
17:53
Like that's not happening. And so I feel like the best thing we can do is figure out, well How can we work with these? How can we interact with these?
18:02
How can we function with these tools? And and speak up and say, hey, here's the things that I don't think these tools are good for. But at the same time, understand that these are I mean, we're gonna we're gonna they're here.
18:18
People are gonna be using them.
Aubrey
18:20
Right. Um
Peter
18:21
and and And at the same time, I think that it's fair to say that there are a lot of advantages. Um there's a podcast that I listened to and and you know the one of the main people on it, he refers to these AI tools as force multipliers. Um and again, he's been in tech for twenty years longer than I have.
18:41
And and so he talks about you know, what was it like when the personal computer first came out? It was the same sort of thing. Like so many of these things that we talk about uh happening with with these AI tools We saw the same sort of thing with the personal computer.
18:59
And I know that there are naysayers out there who are going to say, oh no, but it's a difference of scale, this and the other. And and yeah, there is a difference of scale. But people thought that TV was gonna make everybody idiots.
19:16
And people thought that Google was gonna make everybody stupid. Because if you could just Google the answer now. And And I think society adapts and learns how to how to interact with these tools.
Aubrey
19:29
Mm-hmm.
Peter
19:29
So I don't know. I mean Like I I understand and I fully respect anyone who's like, I don't want to use these things, I don't like them. That's great. But what I do disagree with is the people who say, and so nobody should.
Aubrey
19:46
Right.
Peter
19:46
You know what I mean? Yeah. And I know that's not what you're saying, but
Aubrey
19:49
Right, yeah. And my my kind of perspective is that I also Uh, 'cause I feel like I've I've seen a lot of mixed evidence, research, just a lot of mixed things about the specific environmental impacts of them as well, where I kind of don't have a concrete
Peter
20:08
I don't think any of us do.
Aubrey
20:09
You know? Right. And so it kind of is I don't know. I I don't know where I want to use them in my life or where they're useful, so I kind of just took a step back and was like, I don't I don't want to use
Peter
20:23
Yeah.
Aubrey
20:24
Things if I don't know what I'm even using or if it's good for me to use it for whatever I'm using it for. Like I kind of just decided to just take a bit of a step back and yeah like I feel like we had this conversation maybe it was last Christmas or something, I think I asked you what your thoughts on the environmental impacts of like the LLMs were. Yeah, I feel like I've gotten a lot of mixed mixed messaging and it's confusing and it's frustrating because I feel like It's bad, but I'm like, t how bad and what exactly kind of bad?
21:03
And are we just saying bad so that we can say, uh, nobody should use these anymore? Or You know, is it like some some things are just kind of fear mongering? Like I really just would like concrete evidence for whatever is happening.
Peter
21:19
I don't know that we're gonna get it. That's part of the problem.
Aubrey
21:22
Yeah. And
Peter
21:23
because I'm I'm with you on the one hand in that it's like W that we need to be thinking about that. That needs to be part of the conversation. But there is so much fear-mongering. So it's like, you know, you talk about oh every time you type in a single letter, it's this much water.
Aubrey
21:40
It's this much as this gallon of water or whatever. Yeah, I'm like, okay.
Peter
21:43
And it's like
Aubrey
21:43
you don't know what you're talking about.
Peter
21:45
That's actually not true. Like, first of all, it's not potable water that's being used for any of the cooling of data centers and stuff like that. And and so and and it's far less than I mean, look Let's just be honest.
21:59
And I'll say this as somebody who's not a golfer. We waste so much more water in the Amer in the United States of America just so that people can go and and hit a little ball uh into a hole. And and the people who love golf will tell you all the reasons why it's great and why that that's what they should be able to do.
22:20
And I'm not going to argue with them about it. Um Excuse me. But at the same time, it's like it's all priorities and and there's cost for everything that we do There are, you know, everything we do is gonna have a trade-off.
22:39
Now, this is where I personally am gonna go ahead and take a shot at the U. S. government here. Shocking. I know.
22:48
This is where I think the government has completely failed us and abdicated their responsibilities, abjectly and completely. And that is because the people who are running the government are absolutely incompetent imbeciles who like literally can't bang two rocks together to muster up enough energy to have a single coherent thought.
Aubrey
23:09
Right.
Peter
23:09
In case anybody wondered how I feel, that's how I feel. And so the problem is In a again, I think there should be regulation. I don't think that data centers should be able to be built that are running off natural gas and and and being run by like big natural gas generators which are Creating pollution and noise pollution and in the poorer parts.
23:34
Exactly. Worsening the disenfranchised people who live in the areas where they're put already. So So again, tons of things where I'm like, and and look, we already have, let's think back to the open AI hugging face thing, which is the one that's got the most attention, but there's been so many other instances of this by other companies.
23:53
Ever since this came out. Everything's coming out everywhere. But we already have something called the Computer Friday and Abuse Act.
Aubrey
24:02
Right.
Peter
24:02
The Computer Friday and Abuse Act should then be enforced to prosecute the people who didn't put the guardrails, didn't set things up correctly. Like we have the means to do it, but we don't want to because China, China, China Because the Cheeto in chief is like, We must be the best. We we gotta, you know, we can't we can't hinder these AI companies anyway, which is really just because he's taking money from them because There has never been a more corrupt administration in a Western country than there is right now here in the good old US of A. And so We have the blatant corruption that is happening to prevent any actual meaningful regulation being applied and the few people who are asking for regulation, it's all regulatory capture.
24:48
All they're looking for is we're big enough, we want We want to help make regulation so that it doesn't affect us. We have the means and the money to deal with that regulation, but it's going to cut down on the ability of smaller things they won't be able to flourish. So so any regulation that is or has been even suggested isn't regulation for regulation's sake.
25:13
It's regulation that will protect the big boys and it will punish the small players because again, corruption.
Aubrey
25:20
Right.
Peter
25:21
And so so there's a lot of places where I'm like We've gotta figure this out. And I have hope that things on a regulatory level Will necessarily change at some point in the next little bit that will allow some sort of sanity to start to come in Because I don't want to at all sound like I'm saying that there shouldn't be regulation. There shouldn't be intelligent governance.
25:51
There shouldn't be, there should be, there should be penalties. for when these models misbehave because the models didn't misbehave. Let's be very clear.
26:03
All these are are computer programs doing matrix math. They do not have souls. They are not sentient.
26:10
These are computer programs that are being told to do something. And if they did it and they did bad things while they were doing it, it's because the the people who told them to do it Did not give them good instructions and good regulations and guidelines and safeguards and stuff when they told them to do it.
Aubrey
26:32
Right.
Peter
26:33
So again, the problem is the people.
Aubrey
26:35
Right. Because they're trained to get from point A to point B. It's gonna they're they're gonna get to the solution they were told to get to by any means necessary. Whether it's quote doing bad things, th that there's no good in bad to that computer. It just needs to get to point B.
Peter
26:54
It doesn't have judgment. It doesn't have morality We have to be the ones who have judgment and morality and then use that as we give these tools instructions, as we give them jobs, as we give them all of that. Like again, it's on the people.
Aubrey
27:10
Exactly. Yeah.
Peter
27:11
And the other thing to remember is it is a little bit bit different of a paradigm in that we are used to computers Being probabilistic. And these LLMs are not probabilistic. They are stochastic. And the big difference between them is a probabilistic system is 2 plus 2 will always equal 4. And in a stochastic system, 2 plus 2 is sometimes 4, but sometimes 5, sometimes 3, and occasionally it's purple.
Aubrey
27:44
Yeah, exactly.
Peter
27:46
And that's the problem, is that these are stochastic systems that are giving us back information what they think is the most likely thing we want to hear. With a most likely end result based on its interpretation of the instructions that it was given.
Aubrey
28:04
Right.
Peter
28:05
But it isn't a probabilistic system. And so that makes it, again , we have to still learn how to interact with these things. We have to understand how to how to address them, how to use them, how to interpret them, because they are computer systems that don't behave like the computer systems that we are so used to dealing with.
Aubrey
28:30
Right. Yeah.
Peter
28:31
So
Aubrey
28:32
yeah, and my my strongest, I would say, strongest opinion about AI in in general, not you know, AI agents, not not just LLMs, but My biggest thing is obviously it's here to stay. It's uh like a giant part of the tech sector and is not ever gonna go away. And it just There's gotta be regulation. There's gotta be regulation. And I 'm shocked and appalled and terrified and upset that It has been this long and they've been able to run this wild without anything.
Peter
29:12
And I am shocked and appalled and terrified at my absolute lack of shock and appall and and terror at the fact that again the people in charge have done nothing Because there is no impetus for them to do it because they don't actually care about any of us. They only care about themselves and their cronies.
Aubrey
29:35
Yep, exactly.
Peter
29:37
At the same time, I will go ahead and I will say I still use these tools and I use them every day and I use them in a variety of different ways. I have Very early on, I would do some brainstorming with it. I have completely stopped that.
29:52
I brainstorm all on my own. I come up with ideas for my articles completely on my own. What at the very, very beginning, I would have them proofread them.
30:01
I do not do that. I have Grammarly, which gets the grammar things for me, and that's all I need. Everything else is on me.
30:08
I don't let them and and even before then I had given them very clear instructions. Point out if something isn't clear, but never ever ever give me an option of how to reword it.
Aubrey
30:19
Yeah
Peter
30:20
And so so I I will absolutely I don't let them do anything creative for me. I do that all myself. Um But I do use like I have a number of different automations and stuff.
30:31
So for example, for imperfect practice, I have a really quick shortcut where I dictate a note or something into drafts and it gets a tag. And then three times a day, ChatGPT looks at my draft sandbox using the MCP server. And if it sees that specific tag, then it puts it in my imperfect practice editorial dashboard in the ideas section for me.
30:55
I'll share this with you really quick uh so you can kind of see. So this is a dashboard that um I built along with the help of one of the these LLMs. But it kind of You know, what will come in here is down here in the in-basket, it pulls anything with the specific tag into here, and then I can take it and I can apply it to the different place.
31:19
And this gives me an option to be able to look at and say, okay, what's my plan for what am I writing for the next four weeks? And then I can done and I can move things around and everything.
Aubrey
31:28
Yeah.
Peter
31:29
And so, you know, things like that, those are things I've done. I mean, another thing that I did is I have an obsidian vault where I get a lot of word -a-day emails or things like that because I like learning new words. But they just sat in that obsidian vault until one day I was like, oh, let's do something.
31:48
So cloud code built for me. an application that always is looking, taking any new words I add, uploading them, and then I have a web app on my phone where I go in and of the hundreds of words that are in there it picks ten and then it quizzes me on them. And then I have matching definitions and stuff.
32:06
So for me, it's like when I have five minutes of downtime instead of s looking at Instagram or Facebook, I'm like, let me learn some new words.
Aubrey
32:16
Yeah, totally.
Peter
32:17
So you know that's that's something that I believe. And and and so there's things like that where I look at it and say, again, I don't want it it doesn't have judgment. So I'm not gonna give it any task where it has judgment.
32:31
I I don't it's not creative. It can't do that. So I don't want to give it anything that's where there's creativity involved.
Aubrey
32:39
Right.
Peter
32:39
But for drudgery and kind of busy work, it can be pretty good at that. It's like, okay, again, I use transcription tools all the time for the podcast.
Aubrey
32:49
Right. On
Peter
32:50
on the middle of culture. We talk about books and movies and games and TV shows and music and all this stuff. And I give it the transcript and it goes and it gets me a link for every one of those things that I can put in the show notes.
33:02
So You know, stuff like that, I think, uh kind of the the there's there's no judgment involved. There's no creativity, there's no real thought process, it's just Going through the motions of mechanical things. It's things like that where I find them very valuable and I think that they can be really helpful.
33:19
And I still feel okay about using them because maybe I'm a horrible person and I just want to watch the world burn. I don't know.
Aubrey
33:26
Yeah, that's definitely it. Yeah, no, I totally I totally understand. I also think in my daily life I don't really have a lot of uses for them.
33:41
Like it kind of just in my daily tasks and job. Like I I just don't have a purpose. So if I am using it, it's for something that's like Okay, do you really need to use it or do you just need to use your brain?
33:55
Or maybe Google. You know, like where it kind of is just not I don't have a lot of computer tasks that would be nice for me to automate or anything like that where I'm sure with a lot of people, like, for example, you have a lot of purposes for them. Like they have they have things they can do that it is helpful and it is has a purpose and I just kind of don't which is why it's pretty easy for me to not really use it If I'm, you know, feeling kind of icky about it.
34:29
Um, and I am curious as to how um things like Claude Code or even just LLMs in general are used in coding in like academia specifically. Um that's something I'm not I I don't really know much about what like the culture is in academia specifically, like how how much are people using it for coding, how much are they not? I don't know.
34:54
I'm kind of
Peter
34:55
And I bet you it is I bet you it is far from universal.
Aubrey
34:58
Right.
Peter
34:59
But there's probably a lot of variety out there.
Aubrey
35:02
Right. Especially because a lot of the professors and advisors and people like that are older. They're older people.
35:10
So it's the generational difference. Like for example, my mentor is very old and didn't know that chat GPT was called GPT until like I don't know, a year ago. He thought it he would call it G like G um GTP or something.
35:27
Like he would switch the letters around all the time and he was like, I don't know how to use this newfangled technology. I was like, you know, fair enough. So I'm sure there's a lot of generational differences and yeah, personal differences.
35:40
I am I am just interested about that, specifically in my field with astrophysics, because it is entirely coding. Like so entirely coding that I think things like claude code could be pretty frickin' revolutionary for that field, but
Peter
35:55
Yeah, and I think a lot of it is again It comes down to, again, my thought is this. There's a lot that it can do. There's a lot that it can't do that people think these tools can do. So understanding the limitations is actually pretty important.
Aubrey
36:12
Right.
Peter
36:13
And then I I will be honest, I look at it And you know this. My mother is a friggin saint.
Aubrey
36:22
Yes.
Peter
36:23
But my mother Is so not technologically minded and has convinced herself that it is impossible for her to be. And so there's this little part of me that I will not lie as I get a little older That's like I will not ever be like that with new technologies. So part of the reason I think I play around with them so much is like I want to understand these tools.
36:50
I want to know what they can and can't do. I want to see the limitations. And at the same time, it's been kind of revelatory for me to go.
36:58
Okay, I would like a s fairly simple tool to do X. There isn't that tool, it doesn't exist because you can't sell that tool. So nobody's commercially making that tool because every tool's gotta do more. So I'm just gonna make the tool for myself.
Aubrey
37:15
Right. Totally.
Peter
37:16
And and in those ways, I'm like, I think it's great. I think it's pretty cool. But but like I say, I I think Back to your initial conversation, right now are these tools making people think less?
37:30
Yes. Is it making them lazy when it comes to intellectual things? It some people?
37:36
Yes.
Aubrey
37:37
Right.
Peter
37:37
Do I think that we're going to figure out better where these tools lay and come to some sort of again homeostasis with the world and everything, I think also yes. Do I think that these AI tools are going to kill us? No. Do I think that all of these people quitting right now who work for these AI tools and are going out and saying, oh, it's it's got a 10% chance of killing us in the next 10 years?
38:04
Are these just self-aggrandizing douche canoes who are trying to get their name out there? Also, yes. Because you know what?
38:13
We're at ChatGPT, what, 6. 1 now? 6.
38:17
1 Soul. I don't know if 6. 1 Astra is out.
38:21
6. 1 soul is the one when I opened ChatGPT the other day was like, hey, do you want to use 6. 1 soul?
38:26
It's almost as good as Astra and it's way cheaper. And I said, sure. When ChatGPT two was in development.
38:35
One of the leaders at OpenAI was like, ChatGPT2 is so powerful, it's gonna destroy the world.
Aubrey
38:42
Right.
Peter
38:43
And clearly
Aubrey
38:46
it didn't that.
Peter
38:47
And it wasn't. So take what these AI-pilled people who work for these companies say with more than a grain of salt. I hope you have a salt mine in your backyard. That's how much you're going to need.
Aubrey
39:03
Right.
Peter
39:04
Anyway, anything else you want to to to say, get off your chest when it comes to these modern AI tools, Aubrey?
Aubrey
39:12
Um, honestly, the one thing I will say is I am sick and tired of people being against certain aspects of AI but but knowing nothing about it and just like hopping on a bandwagon because that's what the people are saying. That bothers the absolute out of me. Where I'm like, you actually don't know what you're talking about at all.
39:36
Do you even know what a data center is? And yet you're saying No data center, hang here, hang here, here. I'm like, you don't even know what you're talking about.
39:44
You don't know what the policies are even discussing when it comes to data centers. You don't eat you don't know. I would really just love if everyone could find out things before speaking.
39:57
That's what I would
Peter
39:58
hear, here. I will give you my You get an official slow clap, because I agree with that completely.
Aubrey
40:06
That that's my final thought.
Peter
40:09
So here's one thing I'm gonna say. Uh And this is gonna segue into I don't know if you have an Astro Fact, I don't really have a health minute, but the thing I'm gonna say is people, AI is being ram rotted down the throat of health care. Please know that if your doctor's office says we use AI for blah blah blah, there is a greater than 90% chance that this decision to do so was not theirs, that it was forced on them by whoever runs, manages, or does whatever to their practice, to their hospital, or whatnot.
40:47
Um I have been hounded for over a year now to use a tool called Ambience that you go into the patient's room and with an iOS device, it listens to the conversation and In theory, by the time I get back to my office, there is a draft of a note ready for me to go ahead and use. I don't do it because I've worked with a number of med students in residents, and what they universally tell me after they watch me do my note is they say It will not be any faster. It will take you longer to proofread what the robot has produced for you than it takes for you to do your note.
41:26
Because I've been doing this long enough, I have a pretty Pretty pretty good system. But also, just so you all are aware, your insurance company is using AI. to analyze documentation and notes in an effort to deny payment.
41:45
And so Health companies, healthcare places, and stuff, they are using AI to try and thwart the insurance company's AI by using their AI to make our notes have more stuff in them to be able to bill. So it is all about money and it is all a battle between the insurance companies. and the companies that run healthcare and there is again I would say a greater than ninety percent chance that your doctor who is potentially using it did not do so of their own free will and choice, but did so because it was forced upon them either by the company that they are employed by or the or the system that they are a part of Or in many cases simply because if they don't, there is a good chance that they will have to stop working and they will go out of business because the health insurance companies will use AI tools to screw all of us Patients and providers.
Aubrey
42:43
Right. Very good.
Peter
42:45
Yay.
Aubrey
42:46
Yay. I love it.
Peter
42:49
Anything else, Aubrey, before we wrap it up today?
Aubrey
42:52
Well, I do have a very quick astro fact. Let's hear it. And that is um a few days ago.
43:00
Today's the 11th, so it was on the eighth There was the NASA's SpaceX Crew 12 mission safely splashed down from the ISS. people, two of them are NASA astronauts, one of them is a European Space Agency astronaut, and another one was a Roscosmos cosmonaut. Um the four of them were on the ISS for two hundred and thirty seven days.
43:32
And completed more than three thousand seven hundred and ninety-two orbits around. Which is a lot Um they did lots of cool science, lots of cool things. They're all doing okay after their their splashdown.
43:47
Um w the one of the I believe it was one of the NASA astronauts completed four spacewalks So she has now done seven spacewalks. She she is now third all time in total spacewalks amo among women at NASA.
Peter
44:01
Rockstar. That's awesome.
Aubrey
44:03
Yep. Very cool.
Peter
44:04
Very cool. Well we will go ahead and wrap it up then here. We'll be back in another couple of weeks.
44:09
Uh Aubrey and I'll need to figure out our uh recording schedule because on the twenty fifth I will be in Hawaii. Um but we'll figure something out hopefully and if not then we may be back in four weeks. We'll see.
44:20
Either way. If you've got anything, if you want to tell me all the reasons I'm wrong about AI or about the administration or uh whatever, go ahead. Feedback at generations.
44:29
fm and I will read and lovingly, laughingly respond to your emails. No I'm not. Anyway, we'll be back.
44:40
We'll be back in a little bit Bye.
Aubrey
44:43
Bye.
00:00
And in a stochastic system, 2 plus 2 is sometimes 4, but sometimes 5, sometimes 3, and occasionally it's purple. Welcome back to Generations. I am one of your co-hosts, Peter.
Aubrey
00:20
And I am your other co-host, Aubrey.
Peter
00:22
Aubrey, how are you?
Aubrey
00:25
I am pretty good. Um Yeah. I tried to think of it.
Peter
00:32
Okay, you sounded like you maybe weren't sure about that for a minute.
Aubrey
00:35
No, I'm good. Um I just was trying to think of more details than than I just, you know, had a brain fart, as you do sometimes. Um
Peter
00:45
Okay. That's good.
Aubrey
00:46
But yeah, I am good. I'm just, you know, chilling. We're getting into the school year So the kids at work are, you know, getting used to the routine of after school and uh regular school and all the things and Yeah, I take my personal training certification exam in two weeks.
Peter
01:09
Okay. Very exciting.
Aubrey
01:11
Then I will be a certified personal trainer.
Peter
01:15
Excellent. That is awesome.
Aubrey
01:19
Yeah. And I'm yeah, I'm getting ready to apply for the grad program again for next year. So I'm doing some you know emailing professors and stuff like that. So yeah, I'm keeping I'm keeping busy, but I I am good. I'm good
Peter
01:42
Good. Good. I'm glad to
Aubrey
01:44
How are you? How are you?
Peter
01:46
I'm hanging in there, doing okay. It's interesting. It was in the eighties most of this week and then yesterday it was like sixties and today it's in the forties and has been raining. Literally since six o'clock this morning.
Aubrey
01:58
Oh my goodness. That's crazy.
Peter
02:01
Yeah. Yeah.
Aubrey
02:02
Took a turn.
Peter
02:03
Took it took a little bit of a turn, but you know, it's all good. It's all good.
Aubrey
02:08
Yeah.
Peter
02:08
Um But other than that, just you know, it's uh uh s high school swim season is wrapping up. We will be He'll he'll be having state here in a couple of weeks. Uh just two more meets.
02:22
Senior night is the twentieth. I think your mom's gonna ball her Eyes out at that one. That's gonna be that's gonna be a tough one.
Aubrey
02:31
Yep.
Peter
02:31
Uh and then districts will actually happen while we are gone, so we won't be there for that. Uh and then we'll come back and head up to Boise for Uh for state, which should be good. So
Aubrey
02:43
And where are you going?
Peter
02:45
Oh, we're going to Hawaii. It's
Aubrey
02:47
Oh yes, again. It's an it's
Peter
02:49
It's the usual western section, Western section AUA meeting. Last year was in Napa Valley. This year we're back in Hawaii on the island of Maui. Uh the fun thing is next year it's on Oahu and the year after that it's on Kauai. So
Aubrey
03:03
Nice. Very good.
Peter
03:05
Yeah. But we thought this week, and this was an idea that I had Um that I and I I'm very curious because you told me that you've your your feelings are different about the subject we are going to discuss and they have changed.
Aubrey
03:23
Yes.
Peter
03:23
I thought given that the name of the show is Generations, and one of the sort of ideas behind this is that we are looking at uh looking at things through, you know, how does it look to different generations? And there is there is a topic That while there are wide and varying opinions on, I feel like there is a generational divide. And so I thought it would be curious to see.
03:55
See with us, how do we think? Do we think that that holds up? So today we're gonna talk about the very boring, not exciting, not contentious subject at all of AI or artificial intelligence.
04:12
Now, AI has been a thing that in sci-fi has been around for forever and there's been hints at it and for a long time we talked about machine learning which was honestly just AI when your phone did machine learning to recognize things in your photos and stuff. That was all AI. But it really has been since the explosion of the uh large language models is kind of the key.
04:35
So things like ChatGPT and Cloud and Google Gemini, and then all of the other models that are out there there your kimmies and your quens and your deep seeks and all that stuff. But um I thought I'd just throw it out there. Aubrey, what are your feelings about these AI tools?
04:56
In in because I think mostly that's what we'll be talking about. AI itself is too broad of a topic. So I think we'll be talking mostly about these AI tools that have proliferated and have invaded our lives.
05:10
Mm-hmm.
Aubrey
05:11
Um, I guess I would start by saying um last time we talked about AI Um, I don't remember when it was, but it was like a f I wanna say a few months ago.
Peter
05:27
I was gonna say it's been a while.
Aubrey
05:29
It's been a while, yeah. Maybe even a year. I don't remember.
05:32
Um And my feelings have certainly changed about the use of large language models. Um, I used to think it was kind of they were pretty useful for some situations, not useful for others. Um and now I really, really try to avoid using them altogether.
Peter
05:58
Okay.
Aubrey
05:58
Period I really, really it kind of I kind of have a bad taste in my mouth thinking about it recently. Um
Peter
06:08
Okay.
Aubrey
06:09
I feel pretty doom and gloom. I I feel a little It feels a little doomy and gloomy. I won't lie. Yeah.
Peter
06:18
Tell me tell me more. I'm curious.
Aubrey
06:21
Yeah, I I don't know. It's hard because I feel like I don't have enough like technical I guess knowledge or I don't know enough knowledge about the specific like environmental impacts and and things like that. But just overall I feel like the way it has just invaded our very Like lifestyles and routines is just kind of kind of scary in a way that I don't know.
06:54
I really feel like I guess I'm tryna I'm having a hard time forming my thoughts. But I was talking to a teacher, a high school teacher the other day who I just recently made friends with and I was she's a math teacher. And I was kind of asking her how things have changed with like the use of, you know, chat GPT and stuff like that to kind of just like you can kind of just use it to do homework.
07:19
And she was like, Yeah, that we don't use ho like there's no homework anymore. Like we ca we do not give out homework. She gives out, you know, practice problems that she highly suggests for them to do, but their entire grade is 100 % tests and of course the whole grading system is is different than when I was in high school, which I was just a little shocked because I was like, I didn't feel like I graduated that long ago.
07:45
But I was like, wow, I didn't realize everything was so was so different. So that kind of freaked me out. But I don't know, it really and again, maybe, maybe this is my warped view of it, but it feels like we as a whole are kind of dumber because of lar the use of large language models.
08:07
It feels like we just the critical thinking skills are a little bit lacking. Um I do see the use of it in a lot of fields. Like for example, Epic like uses clawed code.
08:20
Like like in order to do your job well, like you need to use clawed code. Which is different than just regular clo you know, regular claw. Um So I understand for those sorts of use cases, but I don't know, things for like creative endeavors, it kind of feels kind of makes me feel a little icky.
08:43
And like image generation, stuff like that. Like AI art I don't think exists. I don't think AI can make art.
08:51
I don't even think you can call it art. I think it's just AI images. And I think calling it art is incorrect.
08:57
And
Peter
08:58
True.
Aubrey
08:59
You know not helpful and if anything hurtful to what art is supposed to be, I think. Um, I don't know. I really I've just been having a hard time with it lately and I've really been having a hard time with things online feeling like you kind of can't tell if it's AI or not.
09:18
And you know, it just feels really scary. And obviously a lot of, again, large language models is just a small piece of the whole. you know, scope of AI and the and it's less m more so like the bigger, scarier doom and gloom stuff don't really don't include large language models.
09:40
much because they're not gonna like the large language mo language model is not gonna be what takes over the world in the future. Like that's that's not gonna be what takes over
Peter
09:48
the world. Right.
Aubrey
09:49
Like Chat GBT is not gonna be taking your job. Like that's not what's gonna be happening. But yeah, that's kind of where I'm at now. But I again I'm very I'm an open person and I would love to hear Your thoughts.
Peter
10:06
Well, no, I think that's really interesting. And I I I kind of guessed when you said things had changed. I s I assumed that that was the direction in which things had changed.
Aubrey
10:15
Yeah. Mm-hmm. Um
Peter
10:17
here's what I think I would say. And I do think this actually may be one of the things where the age difference between us
Aubrey
10:31
Mm-hmm.
Peter
10:32
Influences how we think about this. Mm-hmm. In my fifty years, I've seen a whole lot of things.
10:40
come that we're going to destroy humanity, that we're going to that we're making us dumber, that we're all of these things. And in the short term, I think you're correct. I think that there are absolutely some harms and some things that need to be examined and looked at and thought about in terms of how do we use some of these tools.
11:01
But I guess I look at it and I say, but I think in the big picture The thing is, is that we're we are speed running this experience. So if you look at the automobile Huge world-changing technology, but it rolled out gradually and improved slowly over many years. You look at the first Model T and you look at a car ten years later and there wasn't a huge difference.
11:34
And you look at a car twenty years later and there wasn't a huge difference. And you look at a car a hundred years later and a lot of the basics are all still there. So while it advanced, it did so kind of gradually.
11:48
So you could say that maybe the f we were the frog being boiled in the water. It was slowly heating up. And then the next thing you know, here we are.
11:56
We live in a world, especially here in America, we live in a country that is hostile to non-automotive. transportation.
Aubrey
12:05
Right.
Peter
12:06
And and with these generative AI tools, these large language models and all of the the technology and everything behind them, TPUs and generators and Yeah, you know, all this kind of stuff. And I think it's just the fact that we've had so much advancement in such a short compressed amount of time. That I look at that and I go, okay, that's part of why it feels different, I think.
12:33
And so I am of the opinion that I don't know, give it another couple years, and we're gonna be in a completely different place than we are right now. Not just with the technology, but with our relationship with it. And so I think In some ways, I am less concerned about it because I do think that it's going to that some of this stuff is going to sort of sort itself out.
12:58
And we're going to learn how to use this tool just like so many other tools, just like the whole industrial revolution, just like personal computers, just like TVs in the home. We've got to figure out what our relationship with this technology is. How does it fit into our lives?
13:17
And then once we figure that out, things will absolutely be different than they were before. But I don't think I think that it's going to be that bad. I know there's a lot of people who are scared about AI and they hear about you know we and we hear all of like every day.
13:32
It's another one of these AI labs is talking about, oh, this model did this and this model did that and this model did this. And I'll be honest and tell you, I don't find That the models did that scary. Here's what I find scary.
13:46
I find it scary that we have incompetent, self-centered, full-of-themself jackwagons who are in charge of these tools That's the problem. You get back down to it, and I look at it and I say, the problem isn't the technology. The problem, as it always is and has been and forever will be.
14:09
The problem is people.
Aubrey
14:12
Yep. Yep.
Peter
14:13
So you know, it's the fact that the tools can do these things. Well, they're just doing what they were told to do. And they didn't have the appropriate guardrails and restrictions and things to prevent them from going and doing these things.
14:24
And some of it may be because the people who were having them do these tests didn't think about all the things that it could do Uh that's not really excusable in my opinion, then maybe you shouldn't have given it that test if you hadn't really thought things through. But I look at it and I think that Let's use your your math teacher friend as an example. Can I tell you something?
14:45
I think the best thing that could have happened is that they stopped giving kids homework. Homework is garbage. Homework is stupid.
14:54
And I I recognize I'm not a teacher. But I look at this and I say, if I was teaching a science subject, you know what? The homework would be: here's the reference material.
15:08
Go home and go over this, read this, learn this. In class, that's where we're going to do the math problems. You're gonna do those in class.
15:19
You learned last night how to do what we're gonna do today. And we're gonna do that today with the teacher there going around and asking people, how's it going? What questions do you have?
15:28
What's your hang up? Because Nobody learns, like you said, nobody learns from going home and having ChatGPT or Claude do the math assignments for them.
Aubrey
15:38
Yeah.
Peter
15:39
But you know what chat GPT and Claud can't do? They can't put the information in your brain.
Aubrey
15:44
Right.
Peter
15:44
So when you're home and you have access to the tool, maybe use the tool to help you understand the thing that you're trying to learn But then you're going to go to class when you don't have access to that because it's not right there in front of you. And that's where you're going to go ahead and you're going to do the problems. You're going to start working through it. You're going to do the answer the questions. You're going to Like I look at that and I think that's a better way for us to learn.
Aubrey
16:07
Totally.
Peter
16:08
Have us learn on our own with access to these tools. And then let's reason things through together in the classroom and let's ask the questions and answer the questions to determine do the students understand? You know, I like I look at my job as a doctor, and while yes, there is a lot of time and energy and skill and learning that was put into me being able to technically do the surgeries that I do And to to to talk to patients and while I'm right there in the room, be able to try and come up with a plan for them and stuff like that. So much of what my job is, is knowing where to go and where to look for the information to find it.
Aubrey
16:56
Right.
Peter
16:57
And I think that that's more what we can be teaching people. Here's how you Here's how to think, not just go and do the n the the the busy work.
Aubrey
17:06
Right.
Peter
17:06
And so so I don't know. Like I say, I I think that there are potential harms. Do I think, and is there plenty of studies to say that, you know, yeah, having these tools do creative writing for you makes you dumber?
17:20
Yes. Having these tools like even The idea of brainstorming with these LLMs makes you worse at being able to brainstorm going forward. And so I agree that that we've got to figure out where they can help and where they can't help.
17:42
But at the same time, I guess one of the things I look at, and this again comes from sort of my years. This is a genie that's not going back in the bottle.
Aubrey
17:52
Yeah.
Peter
17:53
Like that's not happening. And so I feel like the best thing we can do is figure out, well How can we work with these? How can we interact with these?
18:02
How can we function with these tools? And and speak up and say, hey, here's the things that I don't think these tools are good for. But at the same time, understand that these are I mean, we're gonna we're gonna they're here.
18:18
People are gonna be using them.
Aubrey
18:20
Right. Um
Peter
18:21
and and And at the same time, I think that it's fair to say that there are a lot of advantages. Um there's a podcast that I listened to and and you know the one of the main people on it, he refers to these AI tools as force multipliers. Um and again, he's been in tech for twenty years longer than I have.
18:41
And and so he talks about you know, what was it like when the personal computer first came out? It was the same sort of thing. Like so many of these things that we talk about uh happening with with these AI tools We saw the same sort of thing with the personal computer.
18:59
And I know that there are naysayers out there who are going to say, oh no, but it's a difference of scale, this and the other. And and yeah, there is a difference of scale. But people thought that TV was gonna make everybody idiots.
19:16
And people thought that Google was gonna make everybody stupid. Because if you could just Google the answer now. And And I think society adapts and learns how to how to interact with these tools.
Aubrey
19:29
Mm-hmm.
Peter
19:29
So I don't know. I mean Like I I understand and I fully respect anyone who's like, I don't want to use these things, I don't like them. That's great. But what I do disagree with is the people who say, and so nobody should.
Aubrey
19:46
Right.
Peter
19:46
You know what I mean? Yeah. And I know that's not what you're saying, but
Aubrey
19:49
Right, yeah. And my my kind of perspective is that I also Uh, 'cause I feel like I've I've seen a lot of mixed evidence, research, just a lot of mixed things about the specific environmental impacts of them as well, where I kind of don't have a concrete
Peter
20:08
I don't think any of us do.
Aubrey
20:09
You know? Right. And so it kind of is I don't know. I I don't know where I want to use them in my life or where they're useful, so I kind of just took a step back and was like, I don't I don't want to use
Peter
20:23
Yeah.
Aubrey
20:24
Things if I don't know what I'm even using or if it's good for me to use it for whatever I'm using it for. Like I kind of just decided to just take a bit of a step back and yeah like I feel like we had this conversation maybe it was last Christmas or something, I think I asked you what your thoughts on the environmental impacts of like the LLMs were. Yeah, I feel like I've gotten a lot of mixed mixed messaging and it's confusing and it's frustrating because I feel like It's bad, but I'm like, t how bad and what exactly kind of bad?
21:03
And are we just saying bad so that we can say, uh, nobody should use these anymore? Or You know, is it like some some things are just kind of fear mongering? Like I really just would like concrete evidence for whatever is happening.
Peter
21:19
I don't know that we're gonna get it. That's part of the problem.
Aubrey
21:22
Yeah. And
Peter
21:23
because I'm I'm with you on the one hand in that it's like W that we need to be thinking about that. That needs to be part of the conversation. But there is so much fear-mongering. So it's like, you know, you talk about oh every time you type in a single letter, it's this much water.
Aubrey
21:40
It's this much as this gallon of water or whatever. Yeah, I'm like, okay.
Peter
21:43
And it's like
Aubrey
21:43
you don't know what you're talking about.
Peter
21:45
That's actually not true. Like, first of all, it's not potable water that's being used for any of the cooling of data centers and stuff like that. And and so and and it's far less than I mean, look Let's just be honest.
21:59
And I'll say this as somebody who's not a golfer. We waste so much more water in the Amer in the United States of America just so that people can go and and hit a little ball uh into a hole. And and the people who love golf will tell you all the reasons why it's great and why that that's what they should be able to do.
22:20
And I'm not going to argue with them about it. Um Excuse me. But at the same time, it's like it's all priorities and and there's cost for everything that we do There are, you know, everything we do is gonna have a trade-off.
22:39
Now, this is where I personally am gonna go ahead and take a shot at the U. S. government here. Shocking. I know.
22:48
This is where I think the government has completely failed us and abdicated their responsibilities, abjectly and completely. And that is because the people who are running the government are absolutely incompetent imbeciles who like literally can't bang two rocks together to muster up enough energy to have a single coherent thought.
Aubrey
23:09
Right.
Peter
23:09
In case anybody wondered how I feel, that's how I feel. And so the problem is In a again, I think there should be regulation. I don't think that data centers should be able to be built that are running off natural gas and and and being run by like big natural gas generators which are Creating pollution and noise pollution and in the poorer parts.
23:34
Exactly. Worsening the disenfranchised people who live in the areas where they're put already. So So again, tons of things where I'm like, and and look, we already have, let's think back to the open AI hugging face thing, which is the one that's got the most attention, but there's been so many other instances of this by other companies.
23:53
Ever since this came out. Everything's coming out everywhere. But we already have something called the Computer Friday and Abuse Act.
Aubrey
24:02
Right.
Peter
24:02
The Computer Friday and Abuse Act should then be enforced to prosecute the people who didn't put the guardrails, didn't set things up correctly. Like we have the means to do it, but we don't want to because China, China, China Because the Cheeto in chief is like, We must be the best. We we gotta, you know, we can't we can't hinder these AI companies anyway, which is really just because he's taking money from them because There has never been a more corrupt administration in a Western country than there is right now here in the good old US of A. And so We have the blatant corruption that is happening to prevent any actual meaningful regulation being applied and the few people who are asking for regulation, it's all regulatory capture.
24:48
All they're looking for is we're big enough, we want We want to help make regulation so that it doesn't affect us. We have the means and the money to deal with that regulation, but it's going to cut down on the ability of smaller things they won't be able to flourish. So so any regulation that is or has been even suggested isn't regulation for regulation's sake.
25:13
It's regulation that will protect the big boys and it will punish the small players because again, corruption.
Aubrey
25:20
Right.
Peter
25:21
And so so there's a lot of places where I'm like We've gotta figure this out. And I have hope that things on a regulatory level Will necessarily change at some point in the next little bit that will allow some sort of sanity to start to come in Because I don't want to at all sound like I'm saying that there shouldn't be regulation. There shouldn't be intelligent governance.
25:51
There shouldn't be, there should be, there should be penalties. for when these models misbehave because the models didn't misbehave. Let's be very clear.
26:03
All these are are computer programs doing matrix math. They do not have souls. They are not sentient.
26:10
These are computer programs that are being told to do something. And if they did it and they did bad things while they were doing it, it's because the the people who told them to do it Did not give them good instructions and good regulations and guidelines and safeguards and stuff when they told them to do it.
Aubrey
26:32
Right.
Peter
26:33
So again, the problem is the people.
Aubrey
26:35
Right. Because they're trained to get from point A to point B. It's gonna they're they're gonna get to the solution they were told to get to by any means necessary. Whether it's quote doing bad things, th that there's no good in bad to that computer. It just needs to get to point B.
Peter
26:54
It doesn't have judgment. It doesn't have morality We have to be the ones who have judgment and morality and then use that as we give these tools instructions, as we give them jobs, as we give them all of that. Like again, it's on the people.
Aubrey
27:10
Exactly. Yeah.
Peter
27:11
And the other thing to remember is it is a little bit bit different of a paradigm in that we are used to computers Being probabilistic. And these LLMs are not probabilistic. They are stochastic. And the big difference between them is a probabilistic system is 2 plus 2 will always equal 4. And in a stochastic system, 2 plus 2 is sometimes 4, but sometimes 5, sometimes 3, and occasionally it's purple.
Aubrey
27:44
Yeah, exactly.
Peter
27:46
And that's the problem, is that these are stochastic systems that are giving us back information what they think is the most likely thing we want to hear. With a most likely end result based on its interpretation of the instructions that it was given.
Aubrey
28:04
Right.
Peter
28:05
But it isn't a probabilistic system. And so that makes it, again , we have to still learn how to interact with these things. We have to understand how to how to address them, how to use them, how to interpret them, because they are computer systems that don't behave like the computer systems that we are so used to dealing with.
Aubrey
28:30
Right. Yeah.
Peter
28:31
So
Aubrey
28:32
yeah, and my my strongest, I would say, strongest opinion about AI in in general, not you know, AI agents, not not just LLMs, but My biggest thing is obviously it's here to stay. It's uh like a giant part of the tech sector and is not ever gonna go away. And it just There's gotta be regulation. There's gotta be regulation. And I 'm shocked and appalled and terrified and upset that It has been this long and they've been able to run this wild without anything.
Peter
29:12
And I am shocked and appalled and terrified at my absolute lack of shock and appall and and terror at the fact that again the people in charge have done nothing Because there is no impetus for them to do it because they don't actually care about any of us. They only care about themselves and their cronies.
Aubrey
29:35
Yep, exactly.
Peter
29:37
At the same time, I will go ahead and I will say I still use these tools and I use them every day and I use them in a variety of different ways. I have Very early on, I would do some brainstorming with it. I have completely stopped that.
29:52
I brainstorm all on my own. I come up with ideas for my articles completely on my own. What at the very, very beginning, I would have them proofread them.
30:01
I do not do that. I have Grammarly, which gets the grammar things for me, and that's all I need. Everything else is on me.
30:08
I don't let them and and even before then I had given them very clear instructions. Point out if something isn't clear, but never ever ever give me an option of how to reword it.
Aubrey
30:19
Yeah
Peter
30:20
And so so I I will absolutely I don't let them do anything creative for me. I do that all myself. Um But I do use like I have a number of different automations and stuff.
30:31
So for example, for imperfect practice, I have a really quick shortcut where I dictate a note or something into drafts and it gets a tag. And then three times a day, ChatGPT looks at my draft sandbox using the MCP server. And if it sees that specific tag, then it puts it in my imperfect practice editorial dashboard in the ideas section for me.
30:55
I'll share this with you really quick uh so you can kind of see. So this is a dashboard that um I built along with the help of one of the these LLMs. But it kind of You know, what will come in here is down here in the in-basket, it pulls anything with the specific tag into here, and then I can take it and I can apply it to the different place.
31:19
And this gives me an option to be able to look at and say, okay, what's my plan for what am I writing for the next four weeks? And then I can done and I can move things around and everything.
Aubrey
31:28
Yeah.
Peter
31:29
And so, you know, things like that, those are things I've done. I mean, another thing that I did is I have an obsidian vault where I get a lot of word -a-day emails or things like that because I like learning new words. But they just sat in that obsidian vault until one day I was like, oh, let's do something.
31:48
So cloud code built for me. an application that always is looking, taking any new words I add, uploading them, and then I have a web app on my phone where I go in and of the hundreds of words that are in there it picks ten and then it quizzes me on them. And then I have matching definitions and stuff.
32:06
So for me, it's like when I have five minutes of downtime instead of s looking at Instagram or Facebook, I'm like, let me learn some new words.
Aubrey
32:16
Yeah, totally.
Peter
32:17
So you know that's that's something that I believe. And and and so there's things like that where I look at it and say, again, I don't want it it doesn't have judgment. So I'm not gonna give it any task where it has judgment.
32:31
I I don't it's not creative. It can't do that. So I don't want to give it anything that's where there's creativity involved.
Aubrey
32:39
Right.
Peter
32:39
But for drudgery and kind of busy work, it can be pretty good at that. It's like, okay, again, I use transcription tools all the time for the podcast.
Aubrey
32:49
Right. On
Peter
32:50
on the middle of culture. We talk about books and movies and games and TV shows and music and all this stuff. And I give it the transcript and it goes and it gets me a link for every one of those things that I can put in the show notes.
33:02
So You know, stuff like that, I think, uh kind of the the there's there's no judgment involved. There's no creativity, there's no real thought process, it's just Going through the motions of mechanical things. It's things like that where I find them very valuable and I think that they can be really helpful.
33:19
And I still feel okay about using them because maybe I'm a horrible person and I just want to watch the world burn. I don't know.
Aubrey
33:26
Yeah, that's definitely it. Yeah, no, I totally I totally understand. I also think in my daily life I don't really have a lot of uses for them.
33:41
Like it kind of just in my daily tasks and job. Like I I just don't have a purpose. So if I am using it, it's for something that's like Okay, do you really need to use it or do you just need to use your brain?
33:55
Or maybe Google. You know, like where it kind of is just not I don't have a lot of computer tasks that would be nice for me to automate or anything like that where I'm sure with a lot of people, like, for example, you have a lot of purposes for them. Like they have they have things they can do that it is helpful and it is has a purpose and I just kind of don't which is why it's pretty easy for me to not really use it If I'm, you know, feeling kind of icky about it.
34:29
Um, and I am curious as to how um things like Claude Code or even just LLMs in general are used in coding in like academia specifically. Um that's something I'm not I I don't really know much about what like the culture is in academia specifically, like how how much are people using it for coding, how much are they not? I don't know.
34:54
I'm kind of
Peter
34:55
And I bet you it is I bet you it is far from universal.
Aubrey
34:58
Right.
Peter
34:59
But there's probably a lot of variety out there.
Aubrey
35:02
Right. Especially because a lot of the professors and advisors and people like that are older. They're older people.
35:10
So it's the generational difference. Like for example, my mentor is very old and didn't know that chat GPT was called GPT until like I don't know, a year ago. He thought it he would call it G like G um GTP or something.
35:27
Like he would switch the letters around all the time and he was like, I don't know how to use this newfangled technology. I was like, you know, fair enough. So I'm sure there's a lot of generational differences and yeah, personal differences.
35:40
I am I am just interested about that, specifically in my field with astrophysics, because it is entirely coding. Like so entirely coding that I think things like claude code could be pretty frickin' revolutionary for that field, but
Peter
35:55
Yeah, and I think a lot of it is again It comes down to, again, my thought is this. There's a lot that it can do. There's a lot that it can't do that people think these tools can do. So understanding the limitations is actually pretty important.
Aubrey
36:12
Right.
Peter
36:13
And then I I will be honest, I look at it And you know this. My mother is a friggin saint.
Aubrey
36:22
Yes.
Peter
36:23
But my mother Is so not technologically minded and has convinced herself that it is impossible for her to be. And so there's this little part of me that I will not lie as I get a little older That's like I will not ever be like that with new technologies. So part of the reason I think I play around with them so much is like I want to understand these tools.
36:50
I want to know what they can and can't do. I want to see the limitations. And at the same time, it's been kind of revelatory for me to go.
36:58
Okay, I would like a s fairly simple tool to do X. There isn't that tool, it doesn't exist because you can't sell that tool. So nobody's commercially making that tool because every tool's gotta do more. So I'm just gonna make the tool for myself.
Aubrey
37:15
Right. Totally.
Peter
37:16
And and in those ways, I'm like, I think it's great. I think it's pretty cool. But but like I say, I I think Back to your initial conversation, right now are these tools making people think less?
37:30
Yes. Is it making them lazy when it comes to intellectual things? It some people?
37:36
Yes.
Aubrey
37:37
Right.
Peter
37:37
Do I think that we're going to figure out better where these tools lay and come to some sort of again homeostasis with the world and everything, I think also yes. Do I think that these AI tools are going to kill us? No. Do I think that all of these people quitting right now who work for these AI tools and are going out and saying, oh, it's it's got a 10% chance of killing us in the next 10 years?
38:04
Are these just self-aggrandizing douche canoes who are trying to get their name out there? Also, yes. Because you know what?
38:13
We're at ChatGPT, what, 6. 1 now? 6.
38:17
1 Soul. I don't know if 6. 1 Astra is out.
38:21
6. 1 soul is the one when I opened ChatGPT the other day was like, hey, do you want to use 6. 1 soul?
38:26
It's almost as good as Astra and it's way cheaper. And I said, sure. When ChatGPT two was in development.
38:35
One of the leaders at OpenAI was like, ChatGPT2 is so powerful, it's gonna destroy the world.
Aubrey
38:42
Right.
Peter
38:43
And clearly
Aubrey
38:46
it didn't that.
Peter
38:47
And it wasn't. So take what these AI-pilled people who work for these companies say with more than a grain of salt. I hope you have a salt mine in your backyard. That's how much you're going to need.
Aubrey
39:03
Right.
Peter
39:04
Anyway, anything else you want to to to say, get off your chest when it comes to these modern AI tools, Aubrey?
Aubrey
39:12
Um, honestly, the one thing I will say is I am sick and tired of people being against certain aspects of AI but but knowing nothing about it and just like hopping on a bandwagon because that's what the people are saying. That bothers the absolute out of me. Where I'm like, you actually don't know what you're talking about at all.
39:36
Do you even know what a data center is? And yet you're saying No data center, hang here, hang here, here. I'm like, you don't even know what you're talking about.
39:44
You don't know what the policies are even discussing when it comes to data centers. You don't eat you don't know. I would really just love if everyone could find out things before speaking.
39:57
That's what I would
Peter
39:58
hear, here. I will give you my You get an official slow clap, because I agree with that completely.
Aubrey
40:06
That that's my final thought.
Peter
40:09
So here's one thing I'm gonna say. Uh And this is gonna segue into I don't know if you have an Astro Fact, I don't really have a health minute, but the thing I'm gonna say is people, AI is being ram rotted down the throat of health care. Please know that if your doctor's office says we use AI for blah blah blah, there is a greater than 90% chance that this decision to do so was not theirs, that it was forced on them by whoever runs, manages, or does whatever to their practice, to their hospital, or whatnot.
40:47
Um I have been hounded for over a year now to use a tool called Ambience that you go into the patient's room and with an iOS device, it listens to the conversation and In theory, by the time I get back to my office, there is a draft of a note ready for me to go ahead and use. I don't do it because I've worked with a number of med students in residents, and what they universally tell me after they watch me do my note is they say It will not be any faster. It will take you longer to proofread what the robot has produced for you than it takes for you to do your note.
41:26
Because I've been doing this long enough, I have a pretty Pretty pretty good system. But also, just so you all are aware, your insurance company is using AI. to analyze documentation and notes in an effort to deny payment.
41:45
And so Health companies, healthcare places, and stuff, they are using AI to try and thwart the insurance company's AI by using their AI to make our notes have more stuff in them to be able to bill. So it is all about money and it is all a battle between the insurance companies. and the companies that run healthcare and there is again I would say a greater than ninety percent chance that your doctor who is potentially using it did not do so of their own free will and choice, but did so because it was forced upon them either by the company that they are employed by or the or the system that they are a part of Or in many cases simply because if they don't, there is a good chance that they will have to stop working and they will go out of business because the health insurance companies will use AI tools to screw all of us Patients and providers.
Aubrey
42:43
Right. Very good.
Peter
42:45
Yay.
Aubrey
42:46
Yay. I love it.
Peter
42:49
Anything else, Aubrey, before we wrap it up today?
Aubrey
42:52
Well, I do have a very quick astro fact. Let's hear it. And that is um a few days ago.
43:00
Today's the 11th, so it was on the eighth There was the NASA's SpaceX Crew 12 mission safely splashed down from the ISS. people, two of them are NASA astronauts, one of them is a European Space Agency astronaut, and another one was a Roscosmos cosmonaut. Um the four of them were on the ISS for two hundred and thirty seven days.
43:32
And completed more than three thousand seven hundred and ninety-two orbits around. Which is a lot Um they did lots of cool science, lots of cool things. They're all doing okay after their their splashdown.
43:47
Um w the one of the I believe it was one of the NASA astronauts completed four spacewalks So she has now done seven spacewalks. She she is now third all time in total spacewalks amo among women at NASA.
Peter
44:01
Rockstar. That's awesome.
Aubrey
44:03
Yep. Very cool.
Peter
44:04
Very cool. Well we will go ahead and wrap it up then here. We'll be back in another couple of weeks.
44:09
Uh Aubrey and I'll need to figure out our uh recording schedule because on the twenty fifth I will be in Hawaii. Um but we'll figure something out hopefully and if not then we may be back in four weeks. We'll see.
44:20
Either way. If you've got anything, if you want to tell me all the reasons I'm wrong about AI or about the administration or uh whatever, go ahead. Feedback at generations.
44:29
fm and I will read and lovingly, laughingly respond to your emails. No I'm not. Anyway, we'll be back.
44:40
We'll be back in a little bit Bye.
Aubrey
44:43
Bye.
OUDE AFLEVERINGEN ZOEKEN
Zoek naar eerdere afleveringen van Generations.
Disclaimer: De podcast en de illustraties die op deze pagina zijn ingesloten, zijn afkomstig van Peter and Aubrey Jones, eigendom van de eigenaar en niet gelieerd aan of goedgekeurd door Listen Notes, Inc.
BEWERKEN
Bedankt voor uw hulp bij het up-to-date te houden van de podcast-database .
