We've heard from GenX-ers and Millennials talking about their views on AI-assisted software development. We've heard separately from GenZ-ers. But we haven't had all three perspectives in the same room together...until now. What advice do GenX-ers and Millennials have for GenZ-ers in the age of AI-assisted coding, and vice-versa? Find out, as we bring back some of our former panelists, Tim Banks (GenX), and Cassidy Williams (Millennial), and Divyasha Pahuja (GenZ) to talk about how all 3 generations can help each other in our new AI reality.
We've heard from GenX-ers and Millennials talking about their views on AI-assisted software development. We've heard separately from GenZ-ers. But we haven't had all three perspectives in the same room together...until now. What advice do GenX-ers and Millennials have for GenZ-ers in the age of AI-assisted coding, and vice-versa? Find out, as we bring back some of our former panelists, Tim Banks (GenX), and Cassidy Williams (Millennial), and Divyasha Pahuja (GenZ) to talk about how all 3 generations can help each other in our new AI reality.
Cassidy Williams
Cassidy is the Senior Director of Developer Advocacy at GitHub! Outside of that fancy title, Cassidy is a startup advisor and investor, open source-er, and meme-maker on the internet. She enjoys building mechanical keyboards, playing music, hanging out with family and friends, and teaching in her free time.
Find Cassidy on:
Tim Banks (they/them)
Tim’s tech career spans over 25 years through various sectors. Tim’s initial journey into tech started in avionics in the US Marine Corps and then into various government contracting roles. After moving to the private sector, Tim worked both in large corporate environments and in small startups, honing his skills in systems administration, automation, architecture, and operations for large cloud-based datastores.
Today, Tim leverages their years in operations, DevOps, and Site Reliability Engineering to advise and consult with the open source and cloud computing communities in his current role. Tim is also a competitive Brazilian Jiu-Jitsu practitioner. They are the 2-time American National and is the 5-time Pan American Brazilian Jiu-Jitsu champion in their division.
Find Tim on:
Divyasha Pahuja
Divyasha Pahuja is a GenZ AI engineer, multimodal ML researcher, and four-time peer-reviewed author at ACM/IEEE and Springer — someone whose career started before ChatGPT shipped and accelerated alongside it. She's had quite a journey: software engineer (back when LLMs weren't a household phrase), then back to school, then ML researcher, and now sitting at the cusp of both as an AI engineer.
Off the keyboard, she's spent the last few years with Feeding India leading a digital campaign team, distributing meals, and building recurring partnerships with local NGOs and old age homes.
Outside of all that, she loves playing badminton, is a self-admitted sore loser, has a soft spot for cats, coffee shops and Netflix because who doesn't, and is just starting an AI community at @un__prompted — say hi if you want to be there early.
Find Divyasha on:
00:08 Intro
01:11 Guest intros
03:30 The demo curse
06:18 Live demos at really large tech conferences
09:30 Tape backups & zip drives
14:27 What scares you most about AI?
16:59 People don't know how to have arguments anymore
23:26 Dev work has turned into AI babysitter
25:30 Do you mourn the loss of hand coding?
31:20 Use AI to augment
34:17 GenZ doesn't want to lose the problem solving muscle memory
45:03 What do you need from other generations to help you navigate the AI world?
50:29 Flattening organizations to restructure for the AI world
52:29 Delusions validated by AI ("brolulu")
55:52 What excites you about AI?
57:27 Do we have a reason to fear Skynet?
59:58 Wrap-up
SPEAKER 1:
Hello everyone, and welcome to another live edition of Geeking Out, where I have actually was a suggestion to bring back our AI generations panel, where the first time we had Tim and Cassidy were part of a millennial and Gen X panel and Division two others. Those are two others or three others I can't remember now. My my memory is terrible, so I do apologize.
SPEAKER 1:
Who represented Gen Z? And it was lots of fun. And now this is the all the generation representatives have come back for a battle royale.
SPEAKER 1:
So welcome everyone. Hello. I didn't.
SPEAKER 2:
Realize we were a battle royale.
SPEAKER 3:
No. Yeah, I was like.
SPEAKER 2:
We don't need to fight them. It's okay. No, we.
SPEAKER 1:
Don't need to fight anyone. No, no, it's not a fight. It's not.
SPEAKER 3:
About put it in. Right. You know.
SPEAKER 1:
So really quickly, why don't we. Why don't we do just brief intros? Your name, what you do, whether it's at work or for funsies and what generation you're present starting with. Tim.
SPEAKER 3:
Yeah. Yeah, I believe I am the oldest one here I am, Tim Banks.
SPEAKER 1:
That is true. Finally, I'm not the oldest one in the room. Yes.
SPEAKER 3:
I am a I am a black belt in Brazilian jiu jitsu. Multiple international title winner in Brazilian jiu jitsu. I pay the bills by being a solution to architect innovative solutions where I'm dealing mostly like 90% with AI workloads of various types and yes, very solidly middle like maybe like back third, I guess of Gen X is born in 75, right?
SPEAKER 3:
So like but yeah, Gen X for sure.
SPEAKER 1:
Also Cassidy.
SPEAKER 2:
I am the solid millennial in the group. I have peak millennial, as the New York Times call it. Born 1991 so got to go to college and stuff, high school and college in the recession of 2008 and so on, and came into a very fun tech industry. But, you know, I think things change and move a whole lot.
SPEAKER 2:
I work at GitHub. I'm on the team at GitHub, where we specifically focus on improving experiences for people using the GitHub platform itself, as well as all of the GitHub Copilot things as well. But mainly my team focuses on the GitHub platform, and outside of work I have two toddlers, and honestly, that's been consuming a whole lot of my time.
SPEAKER 2:
There's there's occasional hobbies thrown in there, but let's be real, I wrangle babies a little bit for a living right now, and that's very fun.
SPEAKER 1:
Awesome. Enjoy that while it lasts. And finally.
SPEAKER 1:
Oh, do we have a freeze?
SPEAKER 3:
Did we. Oh, oh.
SPEAKER 1:
Go back.
SPEAKER 2:
A little bit.
SPEAKER 1:
Oh damn it internet, why are you doing this to us?
SPEAKER 2:
It's a curse. Computers are so hard.
SPEAKER 1:
I know computers.
SPEAKER 3:
Are hard. It is an extension of the of the live demo curse.
SPEAKER 1:
Oh.
SPEAKER 2:
It's true.
SPEAKER 1:
That's so real. True. Blue. Okay.
SPEAKER 2:
I often do, like, send a blessing to the demo gods whenever I start in a conference talk. And it usually.
SPEAKER 1:
I mean, yeah, that's, you know, my, I don't usually do live demos for that reason.
SPEAKER 2:
Yeah.
SPEAKER 1:
Unless they're like a, you know, they don't have an online like a connect to the internet component and they're fully containerized. And then I'm like, all right, I have a better chance.
SPEAKER 2:
I've had some nightmare ones where I'm like, dang, the talk was going so well until that point when you know, everything gets worked.
SPEAKER 3:
The problem is I will, I will get I can get very thrown off by it and it's hard to get it was I just now will record it and then be like, if you want to, if you want to do it to go to this link.
SPEAKER 1:
Yes, yes, actually I'll do like I'll do a recording. I say like this is a recorded demo with a live narration.
SPEAKER 2:
There you go.
SPEAKER 1:
So there you go. Are we back? Hey, dude, I think.
SPEAKER 1:
Well, just in time. Intros.
SPEAKER 4:
Yes. So since we're going in order, I am obviously the youngest one here. I am the bachelor.
SPEAKER 5:
Oh, no, I think I lost her again.
SPEAKER 4:
I am the.
SPEAKER 4:
Oh, God.
SPEAKER 2:
Once again, computers.
SPEAKER 1:
Computers.
SPEAKER 3:
They were a bad idea.
SPEAKER 2:
They really were. But I appreciate that. They pay the bills.
SPEAKER 3:
Yes, but they also generate the bills.
SPEAKER 4:
Yeah, that's.
SPEAKER 2:
That's true.
SPEAKER 1:
Yes. They keep us employed. All right, here we go again.
SPEAKER 4:
Okay. As I was saying, going by order, I'm the youngest one here. I had to present the Gen Z voice on this panel. Though. I am the oldest flavor of Gen Z that you find. And my engineer at IFS, which is a little ironic because I'm literally building the thing we're questioning here today.
SPEAKER 1:
Perfect. Amazing.
SPEAKER 4:
But I actually have engineering. Did I get cut off again?
SPEAKER 1:
Yeah.
SPEAKER 3:
You're back now?
SPEAKER 1:
Yeah. You're back okay.
SPEAKER 4:
Yeah. Would you just give me a moment?
SPEAKER 3:
Yeah.
SPEAKER 2:
No excuse.
SPEAKER 4:
Yeah, yeah.
SPEAKER 1:
No worries. Well.
SPEAKER 2:
We we could talk about so many different things, so.
SPEAKER 2:
We were talking about live demos. One of the most fascinating things has been witnessing the live demos at the really, really large tech conferences, like, for example, Microsoft builds and GitHub universes and stuff that I've been a part of a lot more lately, just in this job and to make everything real, everything is so perfectly set up where there's a pre-recording just in case something goes wrong.
SPEAKER 2:
But then also while you're doing the live demo on stage, it has to be live like they want it to be as live as possible. They actually have what's called a shadow, where someone backstage is following your mouse and typing at the exact same time you are, so that if something goes wrong, they switch to that other live computer first.
SPEAKER 2:
If that works, great. If it doesn't, then they switch to the recording. And so there's all of these different backups, and there's actually a point in the Microsoft Build keynote this past June where that actually happened to someone. And it was the most spectacular thing where like their demo wasn't working in the speakers, just like now I'm going to hit this and like, we could see backstage that wasn't working.
SPEAKER 2:
They flipped it. The shadow was just doing it. And then once things started working, they flipped it back. No one could tell. And it was spectacular. It was just the amazing logistical feat of keeping the demo live was.
SPEAKER 3:
It was. So we're talking about high availability essentially.
SPEAKER 1:
That's right. High availability demos. That's amazing.
SPEAKER 2:
Yeah for sure.
SPEAKER 1:
Okay. Take three.
SPEAKER 4:
Apologize for that.
SPEAKER 1:
No. No worries.
SPEAKER 1:
Okay. So so you were you were saying that the the work you do is AI based, which is.
SPEAKER 4:
Yes. Which is I said I consider how dangerous it seems and that's what we're here to question. But I actually have done software engineering before since before the wave it. But now I'm in deep you know, agent AI and LLM pipelines and stuff. So yeah.
SPEAKER 1:
Cool. And you said you're you're late late Gen Z are you born.
SPEAKER 4:
97.
SPEAKER 1:
97.
SPEAKER 2:
Oh yeah. That's like right at the cusp.
SPEAKER 1:
Oh, that's the year I started university. I'm crying a little bit I.
SPEAKER 2:
One of my cousins is 1997 as well. And she always like she's she feels so weird being on the cusp because she's like I feel like I both understand the millennial memes really well and like saw a lot of like the internet come to be and stuff. But at the same time she relates a lot more to Gen Z stuff or other things.
SPEAKER 2:
It's it's yeah, it's an interesting.
SPEAKER 4:
Yeah. I think I might like to go on myself because I'm on both sides. And since I've, I've seen that side as well.
SPEAKER 2:
So yeah, the, the siblings thing is a major thing for sure.
SPEAKER 3:
Yeah. Now by 97 I was already married with kids and.
SPEAKER 1:
All right, well thanks for the intros, everyone.
SPEAKER 1:
I dropped him, all right. You went.
SPEAKER 3:
Okay. It was I think it was interesting. And I think maybe to kick us off like 97 was when I was first starting to really start interviewing for jobs in tech. Yeah. And it was like a lot of. As for hundreds and can you deal with punch cards and tape drives and like giant tapes because like, I, I remember right when you had to know all of the, like, tape backup related switches for the Tar command, right.
SPEAKER 3:
Oh my god. Because because Tar still has all of the tape backups in there for like where to position something in the silo and how far forward to move the tape and everything like that. So it's like that was stuff you had to do. And it was it wasn't just like to make to put to put things up in a single archive file.
SPEAKER 3:
It was like, now we got to back them up.
SPEAKER 1:
I remember actually at one point my dad had bought like a commercial, not a commercial, a consumer version of the tape backup. Yeah. And then I don't know if. Well, Tim will remember for sure. I don't know about Cassidy and but zip drives.
SPEAKER 3:
Oh, yeah. Yeah.
SPEAKER 1:
They were around for a for a brief period.
SPEAKER 3:
100 megabyte floppy drive. Are you kidding?
SPEAKER 2:
Oh, I know a floppy disk. Is that from a zip drive?
SPEAKER 1:
That's different from a zip drive.
SPEAKER 3:
So a zip drive is still a floppy disk, right?
SPEAKER 1:
Yeah, I guess. Yeah.
SPEAKER 3:
Okay. 100 megabyte magnetic floppy disk inside that zip drive. Right. But before you were constrained to 720 K or 1.44 Meg and so like and you, you know, you would, you would go places with like almost like a backpack full of floppies, right. Like, like your, like your papers and stuff like that. Any projects was like or what's worse or worse, a briefcase full of punch cards.
SPEAKER 2:
Wow.
SPEAKER 1:
Yeah. I never experienced that.
SPEAKER 2:
Yeah, I feel like I used, I used floppies probably until about middle school. And then we started moving to CDs.
SPEAKER 4:
Yeah, yeah, I.
SPEAKER 1:
Was going to.
SPEAKER 4:
I think I've done the same thing, but with CDs in our backpacks. So yeah.
SPEAKER 1:
Yeah, I still handed in assignments in university on floppies, I remember.
SPEAKER 2:
Wow.
SPEAKER 1:
Yeah. Was just becoming like the little USB flash drives were just starting when I when I finished.
SPEAKER 3:
But if you talk about like the amount of time it took to go from where floppies are used in a regular basis to now like no one even has magnetic drives, right? I think it was like five years at the most.
SPEAKER 2:
Yeah, yeah.
SPEAKER 3:
Right.
SPEAKER 4:
In that area I would say.
SPEAKER 1:
You know, like I'm still scared of, like having a magnet around a computer because of the trauma from growing up.
SPEAKER 3:
Yeah. But but I think about now, like, let's talk about what we're talking about, like, you know, like even five years ago, we were still basically mostly still doing like algorithm standard kind of machine learning models and stuff like that, where it's like prediction, like, you know, a recommendation for what social media stuff you see or shopping, you know, like, hey, is what is my bill going to be for this utility or whatever?
SPEAKER 3:
Right. And there were the there were the early jokes about, oh, you know, we fed a hundred rom coms into the, into this computer and it and it wrote something out. Do you remember those memes on, on on Twitter.
SPEAKER 2:
Yeah, I do remember those.
SPEAKER 3:
Those are like five years ago. And it was that's.
SPEAKER 2:
Actually mind blowing. Yeah. They were.
SPEAKER 1:
That's why.
SPEAKER 3:
This is so funny. This is he does never happen. And now he.
SPEAKER 2:
Yeah I remember making fun of like GPT 2.5 coming out and I was just like, ha ha, I can make it say Cassidy is great.
SPEAKER 1:
I remember like, just having so much fun with the AI generated images. I'm like, this is hilarious. I can get it to draw me. Kathy Burrows with a bunch of funny stuff on them.
SPEAKER 2:
Yeah, like the images were so bad back then, but I was like, it's magical though, because. And now I'm just like, now I loathe them. I don't know why it feels so different now than it did when it was first coming out.
SPEAKER 1:
Because everyone's doing it now.
SPEAKER 3:
Well, when it first came out, it was a novelty, right? It was like, this is cute and no one's taken seriously. And now we're being told, like, you have to use this all the time for everything, etc., etc. we are being unwilling participants in either contributing to or consumers of AI like it has become. I don't know if ubiquitous, maybe, but like.
SPEAKER 1:
Yeah.
SPEAKER 3:
In the worst possible way.
SPEAKER 1:
Yeah, yeah I agree. So on that.
SPEAKER 3:
Note, unavoidable are a better way to say it.
SPEAKER 2:
Yeah.
SPEAKER 1:
So so on that note, like I have two questions for for all of you, which is when what what most excites you about AI but also what scares you most about AI. Let's go in reverse order now with let's talk about what's what scares you first about AI.
SPEAKER 4:
I feel like, as we've talked about how ubiquitous it is and like what scares me is that the dependency we have established with AI. And here I actually speak for myself as well, because it's become such an integral part of how I work. And I think that if you actually took it away from me tomorrow, I'd probably live, but just a lot slower.
SPEAKER 4:
And I feel like that's the low stakes version of the problem, because what's here is like, I see people using, you know, AI for medical, legal, financial and actually decisions as well. But AI is not fit to answer all of that. So like even if you train it on every spoken or written whatever, it's it's not fit. It's not professional to actually answer it all that.
SPEAKER 4:
But people are on a daily basis using it, especially for medical and emotional, you know, advice. But that's not right. And that's what scares me quite a lot. And, you know, because it's like if you're if you're using it in a code base, if it is, it's fine. You can catch it later. But if it hallucinates in a medical decision then it's I can have amplifications later.
SPEAKER 4:
So that's what scares me quite a lot.
SPEAKER 1:
Yeah, yeah. And it's it's freaky because like, I think a lot of people like we already had Doctor Google as, as the.
SPEAKER 3:
MD came out.
SPEAKER 2:
Yeah.
SPEAKER 1:
Doctor Claude I think is a little bit scarier now because.
SPEAKER 3:
It's the same hypochondriacs up at 230 in the morning that are still asking.
SPEAKER 1:
It's true. But but but Claude, Doctor Claude says it with such certainty that you're almost like believing what it says. And that can be scary.
SPEAKER 2:
Yeah.
SPEAKER 1:
Cassidy. Oh, sorry.
SPEAKER 4:
No, no. Just saying. Yeah, it says it with a lot of confidence that you just believe it. And that's great.
SPEAKER 2:
And I think similarly, I'm worried about people losing their ability to think for themselves. And, and there's there's a whole lot in there. But I've been reading so many different studies about how, like, there's one that I saw recently where people were unable to handle rejection and arguments as well because they were so used to, like I just saying, like, you're absolutely right and like hearing them out rather than pushing back necessarily.
SPEAKER 2:
And so in real life situations, people like their actual I'm not a scientist. They did stuff with the brain and they were showing how people got way more frustrated, way faster, having argumentative interactions with humans because they weren't used to people being they weren't used to people pushing back on them. And so like that, that human interaction and then just the ability to think deeply in these human interactions in general.
SPEAKER 2:
I think we're seeing a lot of people getting impatient, both with AI, but also with like the quick hit social media Stipe type of entertainment where people are so impatient that they aren't thinking deeply about problems, solutions, how they interact with others and how they go about their lives.
SPEAKER 1:
Yeah, I, I agree with that. And I don't know if that happens to any of you, but I find myself like when I'm troubleshooting something and I start using AI and then the AI is not working and I'm like, oh my God, what am I going to do? And then I have to like, remind myself, hey, you have a brain, you have experience, you can actually troubleshoot.
SPEAKER 2:
Yeah. It's like if you're on an escalator and it gets stuck, you're like, dang it. Then you remember, oh, I can just walk the stairs. Yeah, yeah.
SPEAKER 1:
Exactly. Exactly. How about you, Tim?
SPEAKER 3:
So I think the thing that scares me most is that AI, it feels like the primary use for AI is for large scale manipulation of public opinion. Either via bots on social media, generation of videos and propaganda, you know, like, you can see like someone can type any prompt into a thing and get get a propaganda video of literally anything.
SPEAKER 3:
Right. But also the extent to which that the companies that that profit from this, you know, have completely unchecked. Like, again, like companies like meta, meta, Facebook, Twitter because it will always be Twitter and things like that. Where they are, they, they, they have their own LMS or they take stake in other LMS and now they're having content produced by this and, and largely interacted with with, by, by also other bots right to to give the image of large scale adoption of an idea.
SPEAKER 3:
Right. Couple that with you know like how you know what what is what was the summer like this summer? What's next summer going to be like? Like the the going over the going over the the the crest of the of the climate change roller coaster, which it appears we have done that and there's no sign of any of this slowing.
SPEAKER 3:
So like AI, for all its potential, is mostly being used for absolute garbage reasons, much the detriment of not only the people consuming it, but also the people who have nothing to do with it and just happen to want to like, you know, not not pay astronomical electric bills or deal with floods and things like that.
SPEAKER 1:
Yeah, that's so true. And it feels it. It's almost like, you know, whenever there's like a new technology there's like right. And it's just like everyone wants to get in on it. I don't care, like if this is useful to me, I want in and that's what it feels like. But extra like everyone's like, I guess scared of the FOMO.
SPEAKER 3:
But I think on top of that, and with AI more than anything else, is that people who are saying or are skeptical, who are cautious, are warning, are shouted down as Luddites. Oh yeah, and and raising very, very valid concerns as they do it. And the disappointing part is there are people that I have or used to respect, you know, in this industry who have totally, abandoned.
SPEAKER 3:
Right, the type of diligence that I was familiar with them having and the type of kind of like ethical competence and having for the guise of promoting AI.
SPEAKER 1:
Yeah, yeah, that's so true. And, you know, I think this leads I know we were we were going to talk about what excites you most about AI, but I'm going to like leave that for a little bit later because I feel like it leads into another good question, which is, you know what? What's something that's changed about your job as a result of AI?
SPEAKER 1:
Because, I mean, you know, there's a couple of interesting angles because I think there's like, I think everyone's like kind of being forced to do it. And then, you know, like as a survival mechanism for our type of work. Like we kind of have to embrace it to a certain extent. But then there are people who are like so into it in a way that's a little bit terrifying.
SPEAKER 1:
So thoughts.
SPEAKER 2:
I like to bring a balance to a room where I can or like I yeah, because I could talk about things that I think are cool about AI. Like it is. It is cool to be able to build a little bit faster and get through certain bugs and things faster. I do like that, but at the same time, I loathe listening to an AI script.
SPEAKER 2:
If someone is reading from a script in a meeting that that is AI written, or there is a interview that I delivered relatively recently where the candidate literally was, I think they were feeding my questions into AI and reading it back. And like I could tell, there was a point where they said, yeah, and that's the real unlock.
SPEAKER 2:
I was.
SPEAKER 1:
Like, oh.
SPEAKER 2:
Lord, come on, I know, I know that I like.
SPEAKER 2:
That kind of stuff. Blows my mind that that people are relying on it in, in that way. And so I do think that like there is a balance. But people the industry doesn't necessarily make money preaching balance. And so that, that is that is where a lot of that comes from.
SPEAKER 1:
Yeah, definitely.
SPEAKER 1:
What do you what about you. What do you think.
SPEAKER 4:
So I if you ask about work I feel like at work they're asking for a lot more now. Like it's like the output is much higher. We're trying to ship features and modules faster than ever. And, you know, a big chunk of my time actually goes into not actually writing the code, but monitoring and observing what we are producing, like on top of actually building.
SPEAKER 1:
So like you're a babysitter of sorts.
SPEAKER 4:
Yeah, I'm.
SPEAKER 1:
Going to be productive, but it feels sometimes. Like I definitely feel like managing.
SPEAKER 4:
I like to call myself the manager.
SPEAKER 1:
Of AI.
SPEAKER 2:
A curator.
SPEAKER 1:
So, so like, do you do you find that fulfilling?
SPEAKER 4:
I feel like in some ways, because now, like now, my domain knowledge is much more appreciated and value than it was before. Because, you know, before it was all about writing code or if you actually know this tech stack. But right now, if you actually know the context and the domain that you're working in, that's actually valued more and I feel more appreciated in that way.
SPEAKER 3:
And so I literally, literally just gave a talk on that last week about how the specificity of under understanding all the specificities and specification specifications and syntaxes of like your favorite JavaScript framework are becoming largely irrelevant. Right. And it's now that you need to know systems and how they operate and how they interoperate. It's not just because the writing of code and the the novelty of that knowledge is starting to go away.
SPEAKER 1:
Do you do any of you kind of mourn that, you know, the loss of like hand coding, like, or do you still get to. Okay.
SPEAKER 2:
I still do I, I.
SPEAKER 3:
I I'd never liked coding. Coding was something you had to do right in order to solve a problem, but it was not something I ever enjoyed. Like I say, I said it over and over again. I write code like like a pop singer sings. You just good enough, right? But not great. But it's a means to an end.
SPEAKER 3:
Like I, I am into systems. I'm into building systems that do things like that. And you know, writing code is usually the last thing that you do for it. But before that, you've got to do all these, you know, planning and architecture and all these other kinds of things, which which is what I enjoy. Right. And those things are now becoming much more important and much more relevant because like I said, as I said, the writing of the actual code becomes a lot less a lot less of a of the important part.
SPEAKER 3:
Right? You know, in my in my job now, I, you know, I'm dealing primarily with AI based workloads, right. So like the what my job entails, right. When for customer facing thing is a lot a lot of interpreting what they want either to their first forays into AI, which is usually less and less of the case now. Now it's sort of like, well, we did this thing and it was expensive or didn't work like we wanted to, but we want to see, you know, how we'd reduce the cost, how we make it more effective, how we make it more efficient.
SPEAKER 3:
What do we do with our data? Right. You know, why is our token cost so high and things like that. But internally though, like we use AI constantly, right. Either to generate statements, work to generate our our architecture diagrams, our delivery teams are using it to, to write code. You know, like primarily they're rarely doing hand code. And we're talking like the teams we're delivering, you know, dozens of dozens of different engagements a month, you know, hundreds of thousands and millions of lines of code.
SPEAKER 3:
And they're not writing anymore.
SPEAKER 2:
It's wild. I on my personal blog, I like on principle, handwrite all of that and like, that's that's something I care about. Or like, if I want to change a component, if I want to do something, I'm going to handwrite that just because it feels like it's my space and I don't need a robot touching that at the same time.
SPEAKER 2:
It's exciting to see, like my side projects are getting done much faster because I do have some of these bots where I'm just like, if I don't want to deal with this side of the code base, I don't have to necessarily. But at the same time, I like knowing what's going on and depending on the side project, if I care more about the output, I handwrite more of it.
SPEAKER 2:
But that's probably a trust problem on my end, too.
SPEAKER 3:
Well, I think it's also like it's it's lower stakes usually. Right. Like this is not something like it's a side project, you know, like if I'm building a garden box for outside, like I'm not going to build that the same way. I'm going to frame a house, you know what I mean? So it's like, yeah, I do think I do think that's a great place for it because you're not it's not really load bearing code you're writing.
SPEAKER 3:
Right. It's it's a thing to, you know, not not single use. But again it's something that has a lower impact. But I think that, you know, we talk about what it frees you up and allows you to do like okay I can I can realize these kinds of like it's easy. And people said the same thing about using synthesizers to make music or whatever else like that, or like, I don't I don't have to splice tape together anymore.
SPEAKER 3:
I can use Fruity Loops, or I can use Final Cut Pro or whatever. Like there's, there's there's always tools to automation that help you know, to you to realize your vision. Right. But that's a long way from oh, I typed in a prompt to AI and now I made this song for me.
SPEAKER 2:
I also don't think I've used Fruity Loops since like 2012. That was a blast of the past.
SPEAKER 3:
I am, I am dyed in the wool like I swear by fruit I love, I love really?
SPEAKER 2:
Wow. I got to look back at it.
SPEAKER 1:
I don't even know what that is.
SPEAKER 3:
Fruity loops was a software that you would literally use to make like music loops, drum loops, you know, basically. No way. Very extensible, very, very visual in how it was. It was fantastic. Yeah, yeah. But it's.
SPEAKER 2:
It's called FL studio now.
SPEAKER 1:
Wow.
SPEAKER 2:
Yeah. I'm I'm pulling it up, but I'm just like, dang, it's so pro now.
SPEAKER 3:
Yeah. Become a real thing.
SPEAKER 1:
By the way, I just wanted to mention, you know, you said you had code like stuff on your website and for, for everyone else here. I, my website is is a fork of Cassidy's website which is publicly available, which thank you for that because otherwise I would have never gotten this thing off the ground.
SPEAKER 2:
Yeah.
SPEAKER 1:
This was one of my AI side projects because I've had my domain parked for over 20 years at this point now, and AI was finally enabling me to like, get things up and running. Like initially I it was an exact clone of of the, the, the repo that you had available for forking. And then I wanted to get up a little bit and I'm like, I know nothing about JavaScript, but I want to make it look pretty and I don't have time to learn JavaScript.
SPEAKER 1:
Also, I don't care for JavaScript. I'm sorry, JavaScript people. So then I was like, hey Claud, I want to add this thing that looks like this on this website. Do it for me. And I am, you know.
SPEAKER 2:
You go.
SPEAKER 1:
Code on there looks like shit. But I'm like, but my website looks so damn pretty now.
SPEAKER 2:
But there you go. Let's see, that makes it all worth it. I'm so happy. Yes. Yeah, that's a different end of fulfillment. How about that?
SPEAKER 1:
Yes. Yeah. So there's there's my endorsement also for for Cassidy's. Was it called blog repo?
SPEAKER 2:
Yeah. My blog.
SPEAKER 3:
Yeah I think.
SPEAKER 3:
I think the important part was, is that you were using AI to put finishing touches on something, essentially. Right? Not to do the heavy lifting of the code writing like.
SPEAKER 1:
No, not not the heavy lifting. Yeah.
SPEAKER 3:
So like I am, I am one of the things that I guess I don't really talk about, which maybe I should, is I am a gigantic astrophysics nerd. Like. What a huge.
SPEAKER 2:
I got to send you more like.
SPEAKER 1:
Astrophysics on Instagram team.
SPEAKER 3:
If I watch, if I watch two hours of YouTube a day, like, say, an hour, 45 minutes, that's going to be astrophysics. If I listen to two hours of podcast a day, an hour and 45 minutes, that's going to be like astrophysics. Most of the books I own nonphysical nonfiction, actual physics. Right. Hugely like and by extension, you know, particle physics, field theory, quantum mechanics and things like that.
SPEAKER 3:
Right. For someone who does not have a college degree, I can sit down and have conversations with astrophysicists and particle physicists pretty casually. Right? And so yesterday I had an idea which I don't know if I should reveal on the podcast yet, because it's actually a pretty cool idea, a notion that I wanted to pursue. Right? And I was like, okay, I know this.
SPEAKER 3:
I've read this, I understand these principles, right, blah, blah, blah. So I'm going to cause like, I think this thing. Right. And I've looked at this thing here, here, here, here, here, help me bridge the gap between what this equation said here and this research and what this equation is like that. And so like, you know, and I'm using like I'm using like, you know, Claude, like opus for this kind of thing because it's a big context.
SPEAKER 3:
Right. But it's like also I understand what I'm doing, understand what I'm talking about, and I just need you to help me put the finishing touches on this idea to see if it has any, see if it stands like, you know, just a review of the of the mathematics of mechanics. Right. But I'm not somebody who doesn't know anything about astrophysics and just said like, hey, you know, how does how does a neutron star work, you know?
SPEAKER 3:
And so I think, I think the important part, this is all to say, the important part of it is as professionals in this industry, people who have experience, right, is that we are using AI tools usually to augment our, our, our efforts, right, not outright replace them. And that's where the danger, I think is and how we proceed is because like, whoo hoo hoo!
SPEAKER 3:
You know, I would say individuals like you being like as a Xenia, right? If you look at the people who were born five, ten years after you. Right, who maybe starting to enter the the workforce, what are they being taught? Right. Are they getting the same body of experience to support what is probably a very useful and practical usage of AI, or are they just being told to rely on AI for answering questions for them?
SPEAKER 3:
And I feel like it's much more the latter than the former, and I think that's how we screw ourselves.
SPEAKER 1:
It's interesting you mentioned that, because one of the things that came out in our in our panel, the Gen Z panel was and the thing and you can speak more to that as well. But a lot of the I would say all the ladies on the panel were kind of like felt robbed by having by having AI problem solve for them if you want to be.
SPEAKER 1:
Actually if you want to talk a little bit more about that.
SPEAKER 4:
We did discuss about how now that we are doing everything with, they don't really get the satisfaction of doing it by themselves. And because they were never brought up in the environment where you could, you were actually, you know, motivated to do it yourself. Now you just AI reliant and you just really don't feel like you have accomplished anything at this point because and like, that's, that's that's where we come from when we say we are a native and we were just never really taught how to do it by ourselves.
SPEAKER 4:
And that's the kind of where we are pushed into.
SPEAKER 2:
I do think a big aspect of learning is failure, where like, like like you, you learn so much. Like you could probably think of a time where you spell the word wrong and now you've never spelled it wrong again. Like, like a big part of the rewiring of your brain and understanding something is falling in your face a little bit.
SPEAKER 2:
And a lot of these tools give you enough answers and, and again, help you figure things out enough that you don't fall on your face as much as you might have anymore just because it'll catch you.
SPEAKER 3:
I think what's I think the other thing with that too, and I like how you said, because I definitely agree that, like, failure is one of the most important parts of learning. Like how do you how do you get good at some things? You got to suck at it first, right? Is that even what failure looks like now is not even the same, right?
SPEAKER 3:
You know, like, I didn't get this answer. Well, let me go prompt it again and see. You know, I'm going to ask the question in a different way and see if I get something that works. Right. You never understand why it works.
SPEAKER 1:
Yeah.
SPEAKER 3:
The and the why something works is important for your mastery of the thing. Like, I talk about this, I'm going to surprise you very much by bringing jujitsu this. But I talk about this in jujitsu as, as, like people that first learn how to do a thing, right? They just go through the rope technique and like, yeah, you move it this way, you do this way, and the thing works, right?
SPEAKER 3:
But if they never learn to understand why the thing works in the first why? Why does the work when you do this? Well, because you have to drop this shoulder change like there's physics, there's body mechanics in it. And if you never understand that, you're never going to progress beyond just being able to do this in a drill.
SPEAKER 3:
You're never going to actually do it in a real fight. You never actually do it when it matters, because you don't understand why it works. And there's a lot of that easy, easy answer getting in a lot of just just making the things work that, as you said, I feel like robs you of a true understanding of what it is you're doing.
SPEAKER 4:
What I would like to sweep away from here for one point that we actually, I feel like have a lot more experimentation going on, like we have the benefit of being able to fail because we have a fallback. We have AI to help us so we can experiment. So like, I have a different take on this, I feel.
SPEAKER 1:
Oh, that's so interesting.
SPEAKER 4:
Like you have to, you know, thinker around that. Even if you fail, we have a fallback, a safe place to go to. Like, I feel like the previous generation didn't have that because, you know, when you when they were, you know, working on code, they had longer compiles, no StackOverflow. Now you their a deployment couldn't be taken back.
SPEAKER 4:
But right now it's mostly just, you know, experiment around experimenting around with AI. You know, even if it is wrong, I can just, you know, it would cost me like eight seconds and I can just do control Z and go back. That's that's what I figured.
SPEAKER 3:
Yeah, I do think that.
SPEAKER 1:
Oh I was going to ask, do you, do you think though, like there was stuff that as part of your education, as part of your experience that prepared you for that? Like do you think? Do you think other people in your generation would be a would see it the same way, or would have the skills to do it the same way?
SPEAKER 4:
I think so, I think I can say that for the whole generation when I say it, because we are had, we have, we have been brought up in environment where we have the space to fail. That was not the scenario before.
SPEAKER 1:
That's interesting.
SPEAKER 3:
Yeah. I do think there's there's a lot of merit in like supercharging the ability to experiment. Right. You know, because because that's part of the process of learning. Let me try this and try this. But but I think it's contingent on curiosity. Right, right. You you have to be invested in more than just does it work, right? You have to be invested in the process.
SPEAKER 3:
And so like I said, using AI to to help speed up the process, to help augment what your skills and talents, I think is a great thing. It's just I feel like, you know, I think we are being pushed as an industry towards the reliance on the AI as, as both as the solution versus the tool.
SPEAKER 4:
Right? We have to have to find that balance when it comes to using AI, of course, use it as a tool and not, you know, it's not your end goal. It's not everything. It's just it's just a catalyst. I feel.
SPEAKER 1:
Like, you know, on, on on that note, I do find that the best way to effectively interact with AI is to have like a certain amount of base knowledge, and so that it's almost like, you know, before AI was a big deal and you were troubleshooting part of the part of the trick to troubleshooting was even knowing what question like what to Google.
SPEAKER 1:
And I feel like that's the case with AI, like almost amplified, because you have to know what question to ask and you have to be like very precise, and you almost need to have like some background information as well to be able to catch it. If it's doing something wonky, like I'll have like these whole conversations with AI, like I'm researching whatever for whatever talk, and it'll go down the path and I'm like, yeah, but I thought my understanding was this.
SPEAKER 1:
And then I'll have like this back and forth and it's like, oh, I think I was asking the wrong question because my knowledge was incorrect on that. Like, what do you all think of what's your take on that?
SPEAKER 2:
I was very proud of my Google fu.
SPEAKER 2:
I loved being able to be just like, oh, I just have to ask this, and I have to do that. And like, yeah, knowing knowing how to ask the right question to get to the right place, even if you did end up copying and pasting from Stack Overflow or something. It was it was a satisfying thing to be able to know.
SPEAKER 2:
And that is an interesting thing now where like, I still keep a code editor open, but there's also like the various agent windows and stuff open. Again, I work at GitHub, so there's a whole lot of GitHub Copilot open at times. And it's interesting where sometimes if I'm like tossing something at the wall, I see it make a bug and I know exactly how to fix that bug.
SPEAKER 2:
And I go into like a little mental debate with myself sometimes being just like, should I just go fix it? Or should I just say, oh, this is exactly the thing that you have to do. Check if this works, then do this. Like it? It's there's I like the idea of asking a question and testing things, but I also sometimes just want to fix it myself.
SPEAKER 2:
I don't know, I go back and forth on how to properly agent engineer something very regularly these days.
SPEAKER 3:
I think there's I think and to that point, like there is you have to have an understanding of context of just beyond like, what is this question I ask? What is the impact of this question? What is how does this relate to the whole system? And sometimes you have to give that to the put that in the prompt for it to be able to sit back at something else.
SPEAKER 3:
Right? You know, if you say like just, you know, if you say like, hey, I want to know, you know what? The velocity of swallow is, right? You know, it will give you an answer. But if you say, I want to know what the link swallow is, so that way I can, you know, figure out if I can get over the wall and blah, blah, blah.
SPEAKER 3:
And you give the whole story, you're probably going to get a better answer, right? But that's for anyone.
SPEAKER 2:
Yeah, that's for asking a human a question.
SPEAKER 3:
That's for asking a human. Right. And so and, and I think that goes along with it. It's like, how do you, how do you how well do you communicate in the first place? Right. Because AI for for all its faults, it is built to communicate or receive communication much the same way that a human does. Right. And so if you, if you, you know, if you it's like it's like, you know, for those of us who are parents, your kid will ask you a question, right?
SPEAKER 3:
And then you answer the question, but then comes like a zillion more questions and you're like, what do you really need right now? You want to know if you can have cereal? Is that it?
Speaker 6
You know, and then and then and then so like.
SPEAKER 3:
It does waste a lot of time, effort and energy. And it's frustrating for everybody involved. Right. So having having the ability to understand a bigger picture, how many ability to communicate the first place does help to figure out what how does help for you to get the result that you wanted, right?
SPEAKER 1:
Yeah, that's such a good point. How about you.
SPEAKER 4:
Think you're right? Obviously specificity is where you know, details is there. Everything is at. And that is why prompt engineering is an actual job right now. Because it's really important. More than writing actually need to be able to talk to. I better in this day of age. So yes I and but again somewhere fundamentals are still important. You still have to go through, you know, all the basic software engineering skills before you can actually before you even know what you need to build, you need to architect it first.
SPEAKER 4:
And that's where the real skill is. So yes, fundamentals basics are really important.
SPEAKER 1:
Still awesome. Now switching gears a little bit since we do have like three different generations represented, what do you feel like you need from each other generally speaking, to help you better navigate this AI world?
SPEAKER 5:
That's a.
SPEAKER 2:
Good question. I kind of just like hearing how people think about their work and like how how people approach it. Like, for example, it's so loud on social media of people saying like either they're anti-air or saying like they have a thousand windows going at any given time and their agents are always running like there's always the extremes.
SPEAKER 2:
And I think it's really valuable to hear how different groups of people are using these tools these days, because I feel like we're all kind of learning them now, but we're also all learning them differently because of our different experiences to like, I'm literally asking deviation. Tim, right now, do you have agents running right now? What are you doing?
SPEAKER 2:
Like what's going on?
SPEAKER 3:
I almost never have an agent running, right. My, my my use of AI is very transactional. I, by and large, I do not rely on it. Like I said, if I, if I'm working on, like a client engagement or something like that, like, sounds like that I will have it. Hey, here's here is like the thing that I wrote or actually recorded, like discussing like what the what the architecture should be, what is or turn this into a sound.
SPEAKER 3:
Right. And then I close the window. I'm done. Right I don't I by and large, right. Very minimal use of LMS, AI in general. Right. Because I don't need it, I think. And not not in like in Nevada. But it's just like not it is not as relevant to my job. I'm spending more time researching what AI does, what it is, how these things work, and understanding how these systems function together.
SPEAKER 3:
And that's not something I can do. For every now and then I'll be, I will I will pop into Claude and I'll be like, okay, help me round this out. Like, this is my understanding of this and this and this and this. This is what a diffusion model runs on this. Like that. Am I miss is there is there any piece that I'm missing.
SPEAKER 3:
Right.
Speaker 6
Yeah.
SPEAKER 3:
And something like that. But it's not agent. It is very transactional. By and large, everything that I need to do in an automated fashion. I've already written the automation for and I prefer prefer something deterministic if I need something done.
Speaker 6
Yeah, yeah.
SPEAKER 4:
I think I agree with him here. I used LMS heavily, but it's always at my back end call. It's never like I'm going to just let it run on my own even if I. Yeah, but even if I do, it's mostly it's always sandbox. I will never let it just go rogue or like it will always have to ask me the permissions.
SPEAKER 4:
It's just.
SPEAKER 2:
Not know like open claw on your personal machine.
SPEAKER 4:
Yeah. No, not at all.
Speaker 6
First ever.
SPEAKER 3:
Heard about that? I was.
Speaker 6
I was blown away.
SPEAKER 3:
I'm like, you were.
Speaker 6
Doing what?
SPEAKER 2:
Yeah.
SPEAKER 1:
Yeah, that's just too much like.
SPEAKER 2:
You got to do the work. Anything like. I'm.
SPEAKER 1:
I'm kind of with Tim and Divyasha, actually, I, my agents are on my beck and call like I, I am so like anal about like what I allow AI to touch on my machine that every single dev project I have going, it's in a dev container. Dev containers are I have a love hate relationship with them. But so my, you know, like cloud only has access to whatever's in that in that dev container.
SPEAKER 1:
And like it's best that way I yeah, I don't know, I'm a little scared. Like I've had some instances where it's done some stupid things and I'm like, okay. But within this, you know, I limit the blast radius. We did.
SPEAKER 2:
Yeah, I'm very similar.
SPEAKER 5:
Go ahead.
SPEAKER 4:
No, that was just saying that I was going on the same thing along the same lines, that it always has to be sandbox. Otherwise I'm just not letting get the power over my data or machine.
SPEAKER 2:
Yeah. I'm kind of trying to figure out, like, what are the use cases where I'll be like, wow, this has changed my life. And there's there's some things I'm like, yeah, that's nice. Yeah. Like like like one of the things that I haven't do now is like, it'll email like my meeting blocks to my husband every morning. And so when I.
SPEAKER 2:
And so now like he knows like, oh, okay, I probably shouldn't call or message during this time because she's streaming right now or something like that. That's nice. But that's also something I probably could have done. So like, otherwise I don't have anything. Yeah, running at all times. Always. But there's the FOMO that you get again, on having to scroll on the social medias and seeing everything.
SPEAKER 2:
You're like, should I be doing more?
Speaker 6
Yeah.
SPEAKER 2:
Missing something?
Speaker 6
Oh my God, so much.
SPEAKER 3:
For me. Because like, when I, when I see those things, I'm like. But there's already something that does that, you know the it's the big like AI agents and microservices debate like, oh we're deploying all these agents. I was like, you've already been able to do that deterministically for like a fraction of the cost, right? Yeah.
SPEAKER 4:
That's the way that has been going on. I feel like when to use a I went to not like I think we have reached to a point where even at my workplace, that we were just using AI for everything that could have just been done with the piece.
SPEAKER 2:
Of that could be a script. Yeah.
SPEAKER 4:
It is.
SPEAKER 3:
It's an argument we had with why. Why are you running a WordPress blog on Kubernetes? Right. It is adequate.
SPEAKER 1:
Yeah, yeah, yeah. I will say one thing that I think has been an interesting side effect of this whole AI thing is that a lot of organizations have flattened a lot more, and I don't see that necessarily as a bad thing, because I do think there was probably a lot of management overhead that didn't need to be there.
SPEAKER 1:
And so it's putting like different emphasis on on like it's almost putting emphasis on, on technical skills, except what those technical skills are have changed because to Tim's point, it's more of a architecture sort of skill set. But it does put an emphasis on on the technical role, less on the management role, or at least the layers of management, which I think is interesting.
SPEAKER 1:
But I think and I'd love to hear what you all think on this, which is because of AI, you know, like technically you can ship out like so many features that much faster and all this. But I have found like my experience and maybe I haven't tweaked things well enough, but I don't trust my agents to like, do the things properly.
SPEAKER 1:
So it requires babysitting. And so the human becomes the bottleneck. The human that's supposed to like check the work becomes the bottleneck. So then no matter how fast your agents get stuff done, you're still having to sit there and make sure that they haven't fucked it up.
SPEAKER 3:
So I'm definitely seeing a flattening of organizations. But again, I've seen it in the opposite way, where people are relying less on tech. And I will say it's the same way that like when DevOps came out like, oh, we don't need an ops team, we'll just make the dev team do it right. And they did it wrong and fired the ops team.
SPEAKER 3:
And now it doesn't work right. But what we're seeing now and I talk and this is talking to lots of customers who want the AI projects. It is like a salesperson. It's a CTO, you know, who vibe coded something together and it works. Now we want to run this in production, right. And you have to have the conversation.
SPEAKER 3:
And it's like, you know, you built a model car, right? And it's a great model car. It is not a real car. In order to have a real car, you're going to need a real mechanic. You're going to need a real engineer, you know, and have these conversations. And I feel like AI has validated a lot of delusions, both in.
SPEAKER 2:
My team calls that “brolulu”, by the
SPEAKER 2:
way, continuously.
SPEAKER 3:
But but because everyone thinks, oh, you can make this work. I just designed this thing and and it does this and does that, does that. I'm like, I was like, it actually does something that looks like that, but it's not the thing. If you wanted to do this, this is what it's going to take, right? And you used to have to have someone in an org that could validate that information before it could, even a prototype could even be built.
SPEAKER 3:
And now you don't have to. And because of that, they don't want anyone suffering on the dreams, so they don't hire the person that can validate any of that until it's like time to go get funding.
SPEAKER 2:
Yeah. There's that. I have I have a whole rant on that and we simply do not have time. I there's, there's so many ideas in the world and CEOs and CTOs, people who are very high up, who probably shouldn't be touching the code anymore, just like you're saying, saying like, I vibed this thing and now I can do anything and they are fully just drinking their own Kool-Aid and kind of like what I was saying before, they can't handle the pushback because the AI said that it was a good idea.
SPEAKER 2:
And I've seen it in so many different instances. There is a VC that I used to work with where he legitimately thinks that he's going to be a world class musician, because he used an AI tool to generate a song, and it was not a good one. But he he used AI to generate the lyrics. He used AI to generate the music and like it.
SPEAKER 2:
It's one of those things where as I try to question and push a little bit, they it's they don't even hear it because their AI has validated them. And that's something that I've seen. Again, this VC example, I've seen CEOs where they say you can't handle the pushback, where where people are saying, like, I made this entire training platform from scratch, and now I'm making all of my employees take it.
SPEAKER 2:
The employees don't want to take it. It's not good. But then they they're just like, okay, now productize it. And now the team has to figure out how to do it because the CEO said so. That's a real example. And I've seen that in so many cases across this industry. I'll cut myself off soon, I promise. Where the Lulu is deep and I don't know how we can fight against it, because it's always the people with a whole lot of money and time and power able to do that, and I need to reel myself back in.
SPEAKER 2:
That being said, I love keeping my job and I will work with these tools to find the best and balanced way to be able to use them effectively for the industry. Because developers matter.
SPEAKER 3:
You mean the tools, the people, or the actual tools?
SPEAKER 3:
Yes, because I work with tools to.
SPEAKER 5:
Yeah, what.
SPEAKER 2:
A world we live in. I like the.
SPEAKER 5:
Kind of like.
SPEAKER 2:
You did have the question, Adriana, at the beginning that we didn't answer. Like, I do think that there's some really cool use cases. I think that there's there's like a literacy and comprehension problem that we need to solve in the industry and how we do that. I'm not entirely sure yet.
SPEAKER 3:
I think if I can, one of the things about which I'm most excited about AI is that it allows you to customize learning content for people that work for them. So many neurodivergent kids out there do not learn the same way, use the same methods like that, and it is very difficult to create curriculum specific to them if you're teaching a class.
SPEAKER 3:
So now you can generate these things in a way that will work for them based on whatever the standards are. Right. And I think the potential for that is great.
SPEAKER 1:
Absolutely. How about you to since we're talking about the pluses, the exciting things of of AI, what's your thought?
SPEAKER 4:
I mean, if you talk to me about exciting things, I think I'm going to be the tool that they're talking about because I feel like now everything is possible for me. It's so exhilarating. Like the feeling is just it's so exciting that I can do anything now. I can create anything without even knowing what it is about. So yeah, I would.
SPEAKER 3:
We are very much if you like it, if you like it. We love it. Right?
SPEAKER 1:
That's awesome.
SPEAKER 4:
That's awesome.
SPEAKER 1:
Now we are coming up on time, so I'll leave. I'll leave you with a final question. And this will probably age me.
SPEAKER 1:
Do we have reason to fear Skynet?
SPEAKER 5:
I.
SPEAKER 2:
Have a really? This is also going to age me. I don't actually know what Skynet is.
Speaker 7
Oh, okay. All right, I. Okay.
SPEAKER 1:
Did you actually. Do you know what Skynet is?
Speaker 7
Him and me. Oh. Oh.
SPEAKER 2:
It's Transformers.
SPEAKER 1:
Terminator.
SPEAKER 2:
Terminator.
Speaker 7
Okay, cool.
SPEAKER 3:
So I have a very good answer for this. And the answer is yes. We do have a reason to fear Skynet. Because Skynet was the merger of techno fascism and actual fascism. Right? And we are seeing that very much play out in front of our eyes now between flock cameras and Palantir and the the push of the AI from from a very central authoritarian government and other central authoritarian governments.
SPEAKER 3:
There is no reason that we should not be alarmed at this. Right. And it isn't. You have to be very diligent about these things. What's happening like in small towns and, and, and counties and stuff like that, where they're putting up cameras without your knowledge. And in mind you, this is already happening with ring, but now it is stepped on, it is shifted into sixth year now with Flock and Palantir and so like, yes, we have very much reason to fear Skynet and we should be doing everything we can to prevent Skynet from happening.
SPEAKER 3:
That was, in fact not people.
SPEAKER 2:
Now that I now that I understand.
Speaker 7
That.
SPEAKER 2:
Yeah, I don't like the normalization of mass surveillance, but I'll leave it at that.
SPEAKER 1:
Fair, fair. How about you?
SPEAKER 4:
I think it's funny because I was actually having a conversation with a friend last week. It was jokingly that how AI is getting self-aware and Skynet waiting to happen, but actually do have a different take than Tim here because I build a genetic systems for a living, so I think I might be qualified to answer this. I've seen my agent forget a file.
SPEAKER 4:
It has just edited twice. We've seen models not being able to tell how many hours there are in strawberry. So I think we are safe for a while before that.
SPEAKER 1:
So, so we're, we're, we're hoping that the this this side effect works to our advantage.
SPEAKER 4:
Yes. That's also amazing.
SPEAKER 1:
Well we are coming up on time. We've come up on time. I know folks have to get to meetings and whatnot. So I appreciate all three of you coming back for a broader conversation on the impact of AI in our lives and for bringing your your generational perspectives. So thank you so much for this. Appreciate that.
SPEAKER 1:
until next time,
SPEAKER 1:
Geeking Out is hosted and produced by me, Adriana Villela. I also compose and perform the theme music on my trusty clarinet. Geeking Out is also produced by my daughter, Hannah Maxwell, who, incidentally, designed all the cool graphics. Be sure to follow us on all the socials by going to bio.site/geekingout.
SPEAKER 1:
And.