945: AI is a Joke, with Joel Beasley

24:37Jon Krohn:You named many of them there, named many of them there. I didn't know Sir Tim Berners-Lee had been on the show. That's super cool. I'll have to check that one out. But just kind of tell us, tell us about some of the most incredible guests you've had on recently, what they talked about, you know, kind of what you learned, maybe the nugget that you got out of them that they didn't want to tell you. Yeah. I'll give you my top three. How's that? Yeah. Sounds good. Okay. Art Hugh, he's a CTL at Novo. He's been on the show many times over the past couple of years, but this time, I think it was like two weeks ago, he dropped some really good advice when he was talking about that you should be getting speeding tickets, not parking tickets.

25:16And the context of that is a lot of companies will move too slow to try for a variety of reasons, and then they'll have to pay the cost of moving too slow long-term. His advice is move fast and pay the cost of the speeding ticket, of moving too fast. And this was specifically in the context of AI implementation. so not like large-scale manufacturing or anything like that just implementing ai you should be getting speeding tickets not parking tickets do you agree with that what do you think i like that

25:49Jon Krohn:that makes a lot of that makes a lot of sense it kind of seems similar to me at the famous move move fast and break things slogan that facebook had in the early days and uh yeah i mean it's it's definitely i think that there's a i i don't necessarily ascribe to the school of breaking things necessarily deliberately. Uber was famous for that kind of attitude, for example, really pushing things legally, for example. I don't necessarily think that that's a great idea, but certainly trying things out, getting into the arena to use, I think it's a Roosevelt kind of saying to say that, to get into the arena and try things out and see what works is way more useful than reading more books about it, doing more planning about it.

26:39Jon Krohn:Just try things out and you will discover, you know, the path will unfold as you try things out. Yeah, and the context is super important because we're not talking about a artificial heart that we're making, or we're not talking about a medical device that fills you up with the IV. Like we're not talking about moving fast and breaking things with those people. We're talking about it often in the context of a situation where the worst case is a little bit of downtime, not death. So move fast, break things. Sounds great. I love Facebook for it, but I hate it when I'm a customer of their ads manager trying to use their ads manager.

27:19And I'm like, this is not good. Number two, Guy Kawasaki. Pretty popular, legendary Apple evangelist. Really legendary. I think he might be Canadian as well. Oh, that's kind of a, I'm going to look that up. Oh yeah. You guys have your club, don't you?

27:34Jon Krohn:Yeah, exactly. I always, anytime a Canadian comes up, I've got to mention them. Yeah, he was great. He was a lot of fun and he's out there surfing with his daughter. He's like, I'm going surfing after this show. And I'm like, you're the coolest 60 year old I know. So he was a good time, but my, he was funny too. He was, he was a funny guy. Yeah. Yeah. And he shared a lot of stories. He was super blunt. And he just said what he thought, which was always great. But one of the things that he really looks for when hiring or working with smart people is humor. He considers humor a sign of intelligence.

28:14And for me, I was like, maybe you should become a comedian, guy.

28:18Jon Krohn:Hey, Gal, why don't you become a comedian? I mean, so Canada is famous for comedians. Norm Macdonald, Mike Myers, John Candy, Jim Carrey, Catherine O 'Hara, Martin Short, Eugene Levy, Seth Rogen more recently, Dan Aykroyd, tons of great Canadian comedians, but one of them is not Guy Kawasaki. He was born in Hawaii. He was born in Hawaii? Yeah. I don't know why I thought he was Canadian. It's like Southern Canada. The Canada of the South. Hawaii. Just completely wrong. Not all the jokes work. That's why you throw away 90 % of what you write. uh zach cass he was the head of go to market at open ai and this is your number three this is my number three third and final recent incredible guest yeah re this this is contextualized to very recent so these are all in the past couple weeks zach cass by the way shout out to zach he's a papa for the first time this week so i got some awesome pictures on my my phone of his new addition to his family, but he was the head of go-to-market for OpenAI.

29:22So he joined them and helped launch ChatGPT. It turns out, do you know ChatGPT was like an experiment to show people what the models could do? It wasn't meant to be this like massive consumer product? Because

29:37Jon Krohn:I didn't know that. I didn't know that. I did not know that. Yeah, they were trying. So they were selling it. They were at$2 million in revenue. They were selling this technology to different companies and engineering firms and things like that. And it was just such a lift to, and this is my interpretation of it. You can listen to the full episode to get the actual story, but it was this big lift to explain it. And so they ended up making this chat GPT thing so that people could just feel it and touch it and understand it. And they tuned it for conversation specifically because they were selling these models that models could be tuned for anything, calculations for quantum stuff or for architecture or whatever it may be.

30:14So they tuned it a little bit for conversation, put it out there, and it became the fastest growing products of our lifetime. And he was there when it went from$2 million to$2 billion. And now he's out there on the speaking circuit. So I'm sure you're going to run into him over there, John. And then he's also doing a consulting type company too. but uh he he had some great talking points uh he the thing i remember the most was the the double standard for technology so your listeners might appreciate this so if we're killing each other at the rate of like 1.2 million people per year in car accidents it's like a crazy high number that's a large number right because our population is what 350 i think last known population in the

31:01Jon Krohn:us yeah 350 million so yeah so that's that's really insane so that's like a third of one percent like one in 300 people every year that's way worse than like 10 of amazon getting cut and that's a callback thank you john uh yeah so that that's exactly what's happening and but but the so we don't get up in arms about that right we're not super riled about that obviously it sounds bad, but I will tell you this. If a Waymo got in a fender bender or even a mildly difficult or one fatality, there would be pitchforks in the streets. And it's because of this concept that we're really okay with human error.

31:41Humans are okay with human error. We are not okay with machine error. Machine error is unacceptable. Human error is acceptable. And that's fascinating because we're watching our world start to blend with things like the autonomous driving and all of that.

31:56Jon Krohn:I'm actually currently writing a query to a chatbot to get, I think there's actually a name of that effect. So I'm asking, what is the name of the effect that people are intolerant of machine error but okay with humans doing the same? We'll see what comes back in real time. I'm not going to, I don't know, automation bias. I thought it would have a good, it needs something better. Tell it, tell it. and be like, you're wrong. Your confidence score is too high. Make it better. So we're going to do it right here. We're going to name it right now. So this comes from, there's like a key paper from 2015 called Algorithm Aversion.

32:36Jon Krohn:And I'll put this in the show notes. The first author on that paper is named Berkeley Dietforst, which first of all, amazing name. Berkeley is the first name, apparently, unless this is a Gen AI hallucination. He has a beard. I feel like Berkeley Dietforst has a beard, right? Well, I mean, I can find that out in real time. Search Google for Berkeley deed forced. He's got like a five o 'clock shadow kind of thing. All right. That's not nothing. Redhead. A lot younger than I was expecting. But so, yeah, we're going to call this thing the deed forced effect. Let's do it. Yeah. From now on. Deed forced effect.

33:13I like it. I'll pass it along to people. Nice. We also talked a lot about bias and LLMs. I hate that phrase. I don't like it at all. And the fact that it's not a bug and it's a feature, it's a core principle of all thinking things. Like bias is this absolutely necessary function for intelligence. Like you have to bias putting your hand on the stove versus not. Like you have to understand that. What we argue about as humans is the popularity of different biases as trending topics like this group's preferenced over this group or something like that. That's what people are really saying when they're saying there's bias in technology.

34:00They're really saying we want our preferences for our groups enacted. And you can see that play out in the real world too, because the GPTs will treat you differently based off of different cultural norms in different countries from where you're logging into. I don't know if you know that or not, but they have them trained on the cultural norms of the places that you're accessing them from.

34:22Jon Krohn:They do. And I heard, I don't know if you heard this as well, but as a Grok fan, maybe you can fill me in on whether this is accurate. But my understanding is that, so one of the ideas behind Grok was to provide a different kind of bias to the bias, which is relatively homogeneous that we get from the other frontier lab providers. And so the idea with Grok was that we could get a different perspective, maybe a more centrist or more right-wing approach than what the Frontier Labs are generally reinforcement learning, training their chatbots to be like. but some of the initial results weren't great and so they it appears that they changed the prompt for particular circumstances to align with elon musk's views on certain topics oh they did yeah like it wasn't even they didn't even try to reinforce the learning training it was just like added to the prompt added to the context like if this comes up this is your opinion that's interesting and you know i like maximally truth-seeking ai like i was sold on that concept from him uh i don't know how i feel about about that i'd i'd have to think about that and research it more about those specific instances i'm not saying i believe it or don't believe it i just it's too early for me to tell like how i would feel about it haven't thought it through but uh yeah, it's a, that's a lot.

35:51I like the maximally true seeking part though.

35:54Jon Krohn:Yeah. It's a nice idea. It's just, it's hard to know what truth is and. Well, that's why I like transparency. So that was a part of the conversation with Zach, the open AI guy. I am pro, I was using different words. I learned from him. It's the word to use is transparency. So I am pro legislating model transparency. I think that's a good legislation point because it should be like a nutrition label where I get to see the ingredients. I get to see what you've told this thing that was trained on raw data. And then you started to shape its future. I want to see what you've told it so that I at least have an option in the marketplace and I know what I'm consuming.

36:30I'm okay. Give me a woke LLM. Give me a redneck LLM. Give me a gangster. Give me all these different types of LLM. I would love them, but I need to choose them and use them for the different reasons that I need them. I can't have this like ambiguous, well, we've got this culture here and we're making decisions and we're doing it behind the scenes. You just use the product. It's like, I need transparency.

36:53Jon Krohn:I like that. I like that a lot. All right. So now we're going to have to kind of, we're going to do a lightning round to kind of round things out here. Cause I know you have a lot to talk about on, on this. This was actually when we first started talking about you doing an episode on the show, the first topic idea is now what we're going to go over, which is the best leadership advice that you've received from guests on the modern CTO podcast that you host over all time. So you kind of recently just told us about some recent incredible guests and some cool lessons that they had for us, but now we're going to stretch back over all the guests you've ever had.

37:25Jon Krohn:And yeah, I guess spend about a minute on each one. Yeah, yeah, yeah. We don't have to hit all of them, But I've ordered them almost in the most impactful applying to my life. So these are lessons that I've learned over the thousand episodes. I've applied them to my life and they've still impacted me today. So it's used experience validated. The first one was Tom Marshburn. He is the first astronaut to command Elon's manned flight. So he was the first astronaut on Elon's first manned flight. Super cool guy. The guy is just next level brilliant. John, he's like a pilot and a surgeon and all of these things.

38:08He's smart and an astronaut. Yeah. That's great. I know. I know. And so he came on the show after his mission with Musk. And we were talking. And I said, look, you do all these crazy things. He shared this story about how they'll fly into an emergency bay flight situation where there's just chaos happening all around, people screaming, dismembered, all these bad sort of things. And they're trying to piece together this situation. And he also shared the space flight and the connection between these two, like if you're doing a spacewalk. And he says, what happens is when you get into this chaos, this chaos of this stress or all these things happening around you, the answer is to keep your world really, really, really small.

38:55He's like, so when he's on that crash site, he's not thinking about the helicopter 30 yards away. He's not thinking about the police all around in the fire truck. He is focused on what is my next move? What's the next stitch? what's the next thing that I have to do to save this person's life? Or what's when he's in the, in space, what's the next thing I'm grabbing with my hand and how am I moving this latch? And what's the next knob that I'm twisting? He, he just makes his world really small. So I heard that advice maybe three or four years ago. And I've put that through practice and stressful situations with my family and my kids work all different ways.

39:34And every time it rings true, It's my hands down, my number one best piece of leadership advice I've ever received on the show.

39:41Jon Krohn:Nice. I like that a lot. And it kind of reminds me of how when you see professional athletes, the best in the world of what they do, mess things up, it's because you can kind of see them in the replays thinking about the next thing instead of just focusing on what's immediately in front of them. So, you know, a baseball, an outfielder in baseball going to catch a pop-up, routine pop-up, but they're thinking, they're already thinking about throwing the ball to first to pick off the base runner instead of just staying focused on what's immediately in front of them catching the ball. It's very hard to be present, to live and act in the present.

40:16And especially when you're a planner or an engineer, which a lot of the people that listen to your show are. Next best piece of advice is more in the business side of things, time management side. And it came from Kyle. When I first met Kyle, he was becoming, he was in the transition process of becoming the CTO of Verizon. And so I had called the CTO of Verizon. I sent him a cold email. He called me back in five minutes. Hans. Hans was like, hey, I thought it was the coolest thing in the world that he called me back from a cold email. He's like, hey, you don't want to interview me because it's not public knowledge.

40:51I'll tell you now. It's private. He goes, but next week, I'm becoming the CEO of Verizon. And this new guy, Kyle, is getting promoted to CTO. He goes, so you want to talk to Kyle? Because by the time this thing airs, I won't even be the CTO. I said, thanks, Hans. So we got Kyle on. This was maybe five, six, seven years ago or something. And he was just brilliant. But he shared this one very specific piece of advice. And it's in relation to time management. And so what he does is every quarter, he'll have like a reoccurring event in his calendar and he'll sit down and he'll say, make a list of everything he is doing and everything that needs to get done.

41:30Like, here's everything I'm actually doing on a daily basis. Here's everything that matters. And then he syncs those lists up because what will inevitably happen is you end up doing, working and spending a lot of time on things that don't actually matter to the goals right now. And so that pie chart of looking at how you're spending your time and deciding if that's the way you want to be spending your time or should be spending your time is something I have been doing now for several years. He's been on the show many times over the past couple of years, but he shared that advice when I first met him and that's stuck with me.

42:02Jon Krohn:I like that a lot. That's important. I also have a quarterly reminder to kind of to review key habit. So every day I have a habit tracking spreadsheet and I track these key things, uh, inputs as well as outputs. So, you know, the kinds of things that you can change, like how many hours did I spend in bed is like a leading metric that I can track every day. And then, um, you know, I could, I could have a lagging metric of, did I wake up feeling refreshed? Something like that. And, uh, there can be all kinds of things like the number of drinks I had has a big negative impact on whether i had a refreshing sleep zero bro you don't drink no list of things you could do before you have kids drink you want to try to be hung over with children they wake up at like 6 a.m and they climb on you there's no there's no drinking with kids that's cool i didn't know that i did i drank a lot in my 20s before but i was like i can't do this anymore even like special occasions or something i just i mean i just i i just don't really no no i like gummies i'll eat a gummy for a special occasion yeah those don't whatever my body when i turned uh so i'm 38 when i turned like 27 or whatever just like stopped wanting to process alcohol at all uh it's just like nope we don't want alcohol anymore and i was like are you sure because it's a lot of fun and it's like yeah we don't want that anymore.

43:29Jon Krohn:Yeah. That happens a lot. I've had, you know, uh, one of my roommates in college, he was, who is still one of my best friends. Uh, we used to drink together a lot. And the last time I saw him, you know, I hadn't caught up with him in a couple of years and yeah, he, he drinks, uh, you know, cannabis beverages because he's like, you know, I just got to a point where I realized if I have one beer, I feel terrible. Um, so I just don't do it. Um, And so just to, so anyway, to get us back on track, this guy, Kyle, oh, I was going to say the thing that got me, got us talking about this was with the habit tracking.

44:07Jon Krohn:Once a quarter, I go back and review. The idea of doing it for time is a really good idea as well. I'd like to incorporate that into kind of. Yeah. Like where are you spending time? Sales delivery. Like where are you spending time? Also another one that I throw in there too, with that recurring event reminder is, are you doing anything you don't like doing? Because we're founders. So inevitably we're the garbage collector. We get to do all the stuff that nobody wants to do or that is new that we don't know if it's going to work yet. But our job is to do things, figure out how they should be done, and then hand them off.

44:35And our ability to do that directly impacts our ability to scale. Because if we never hand anything off, we're a team of one. 100%.

44:42Jon Krohn:Nicely said. And yeah, so you have mentioned how Kyle is now CEO of Verizon. I think you might not have said his full name, which is Kyle Malady, if I'm pronouncing that correctly. I think so. Um, yeah. And, um, yeah. So now this is potentially based on how much time we have left for you to record. This might be your third and final piece of leadership advice, which is kind of nice doing things in threes. So we have three topic areas, news, recent incredible guests, and best leadership advice you've ever gotten on the show. And in each of those three categories, we have three topics. And while I would love to give you guys the best advice from Seth Godin or Ed McLaughlin, who is the president of MasterCard and CTO.

45:24I'm going to go with this one here at the bottom, but you can go listen to their episodes on Modern CTO. You can go listen to their episodes and the information's in there. But the third biggest impact to my personal and professional career leadership advice that I ever got, Mark Porter, CTO, DBT Labs, and drafting drafting your emails like in Evernote when you're angry before sending them. So you don't want to do it in the draft because you might accidentally hit, you know, you know, this might be against you. You might tap that, that send button halfway through, which is the worst. And inevitably, anytime that you're emotionally charged, it's not, you need to wait for yourself to get back to neutral review the writing often you'll be like that sounds like a crazy person and you won't send it like i have a a not send rate of basically a hundred percent of every angry email i've ever

46:23Jon Krohn:done this that's what i was gonna say is that i it's hard for me to imagine based on and granted all basically a hundred percent of our interactions they've been digital most of them most of the time that we have had has literally been while recording podcast episodes either your show or now mine but it is impossible for me to imagine you angry. I don't get angry. You can talk to my wife. Uh, I, I, I don't, I'm not a very angry person. Um, you might be angry after you find out that I already had been talking to your wife.

46:56Hey, look at that. No. Um, but yeah, there's a, yeah, I don't know why, man. I've always been people like earlier on in my life. They, somebody told me they're like you remind me of the guy from like the lego movie like everything is awesome and i was like yeah i guess because like we're alive and we could not exist john but we do exist and you mostly don't exist there's like mostly billions of years of emptiness in a vast empty universe and for some reason you're standing on a comedy stage making jokes and people are laughing to me it's beautiful and i think whatever you consider god is fine but like I, when I use the label God and I'm referencing the thing that created everything.

47:40And so whatever, whatever that ends up being is what it is. But like, I think the fact that we are here by whatever means necessary or however, whatever words you want to use to describe it. I just think the act of being here is the coolest thing in the world.

47:54Jon Krohn:That's a really good, uh, really nice point to end the episode on and yeah, something to not lose sight of for sure. So yes, that brings That brings us to the end of the topics for the show. But as you know, at the end of every episode, I ask my guests for a book recommendation. Joel, what have you got for us? Well, since we brought up God at the end, I'm going to go with the Bible. No, no, no. He's from Nashville, people. Or he's from Nashville. He's from Tennessee, people. No, no. I would say that one of the books that helped me the most spiritually was the book Untethered Soul. and that book the first three chapters really walks you through almost this exercise of understanding what you are and it's like a consciousness exercise if anything like if you cut off your finger like are you still you it's like well yeah and then it just goes through this whole situation where it kind of helps you understand internally like what you are and what i've come to and the conclusion is like i am the ability to make a choice like i get to choose and there's a bunch of different options that appear to me every day in every single way.

49:04And I am just this choosing mechanism and my choices yield are directly related to the quality of my overall experience here. And so what I try to do is I try to make better choices. And I think the Untethered Soul book, I don't know if it'll help you make better choices, but it will definitely help you understand how to think about who you are and yourself. and I think it's applicable to any religion. So for example, I brought this up at like a small group or our church and they're like, well, is it a Christian author? I was like, no, but it doesn't matter because God gave us consciousness and this author is talking about better understanding consciousness.

49:38So it doesn't need to be explicitly a Christian author. And for that regard, I believe it would be applicable to any religion.

49:47Jon Krohn:Nice. I like that, Joel. Really nice. For a funny guy, you've had his pretty serious ending to this episode. Oh, well. And now we're out of time. And yeah, you know, most comedians, they like to end the podcast strong. I like to end it on an awkward note. So thank you for listening. JoelComedy.com, Modern CTO. Yeah, so I'll have in the show notes your upcoming tour dates and people can keep that, you know, keep an eye on that. Even if you're listening to this in the distant future, there will probably just be more tour dates at even bigger venues in years to come. So joelcomedy.com. Go on Instagram.

50:29I'm trying to do that. There's a QR code. Oh, yeah.

50:33Jon Krohn:So if you're watching the video version, Joel is holding up a printout of a QR code. And it leads to Mr. Joel Beasley on Instagram, which we will also have in the show notes. Any other way that people should be following you other than Monarch TV Podcast, Joel Comedy? if you if you thought if you had fun if you're like oh that's a funny guy then you like my style of humor instagram's the best i do a 60 second update every week on my comedy journey i'm in week 44 and uh that's if you like the technology leadership stuff it would be modern cto available everywhere you can stream podcast favorite episode in the history is john crone's episode so go listen to john on our show had a great time and uh thank you guys for listening my first episode or my second.

51:20I like all your episodes. They're equal.

51:24Jon Krohn:Uh, equally your two favorites ever. Yeah. Wow. What are the odds? Yeah. Um, that's why they made the list too for today. Uh, if you're like, well, the lists were all about best advice and you're like any advice that he had in those episodes, don't take them. But we had a good laugh. No, this is great. Nice. Yeah. I had a really great time, Joel. Hope to have you on the show again soon. Thank you so much for taking the time and best of luck. Awesome. Now that was a fun one. In today's episode, Joel covered his systematic approach to comedy using AI, including recording every performance, analyzing laughs per minute through Gemini, and maintaining a detailed spreadsheet tracking what works.

52:07Jon Krohn:He provided the top advice he's ever received on his Modern CTO podcast, including astronaut Tom Marshburn, suggesting you keep your world small to come out of stressful moments on top. And he talked about the case for AI model transparency is a form of a nutrition label, allowing users to understand what biases and instructions have been built into different models so they can make informed choices about which tools to use. As always, you can get all the show notes, including the transcript for this episode, the video recording, any materials mentioned on the show, the URLs for Joel's tour dates, and social media profiles, as well as my own, at superdatascience.com slash slash 945.

52:45Jon Krohn:Thanks to everyone on the Super Data Science podcast team, our podcast manager Sonja Brejevich, media editor Mario Pombo, our partnerships manager Natalie Zajewski, researcher Serge Macis, writer Dr. Zara Karche, and our founder Kirill Eremenko. Thanks to all of them for producing another excellent episode for us today, for enabling that super team to create this free podcast for you. We are deeply grateful to our sponsors. You can support this show by checking out our sponsors' links, which you can find in the show notes. And if you want to sponsor an episode yourself, you can go to johnkrone.com slash podcast to learn more.

53:18Jon Krohn:Otherwise, share this episode with folks who would like to have it shared with them, who would like a good laugh about AI. Review the show on your favorite podcasting app or whatever, wherever you view episodes, listen to episodes, subscribe if you're not already a subscriber, but most importantly, just keep on tuning in. I'm so grateful to have you listening, and I hope I can continue to make episodes you love for years and years to come. Till next time, keep on rocking it out there, and I'm looking forward to enjoying another round of the Super Data Science Podcast with you very soon.