1 00:00:01,000 --> 00:00:04,759 How'd you like to listen to dot net Rocks with no ads? Easy? 2 00:00:05,320 --> 00:00:09,880 Become a patron For just five dollars a month you get access to a private 3 00:00:10,000 --> 00:00:14,359 RSS feed where all the shows have no ads. Twenty dollars a month will 4 00:00:14,400 --> 00:00:18,800 get you that and a special dot net Rocks patron mug. Sign up now 5 00:00:18,839 --> 00:00:24,199 at Patreon dot dot net rocks dot com. Hey Carlin Richard here. As 6 00:00:24,239 --> 00:00:29,519 you may have heard, NDC is back offering their incredible in person conferences around 7 00:00:29,559 --> 00:00:33,960 the world, and we'd like to tell you about them. NDC Copenhagen is 8 00:00:34,000 --> 00:00:39,600 happening August twenty seventh through the thirty first. Go to NDC Copenhagen dot com 9 00:00:39,799 --> 00:00:45,039 for more information. NDC Porto is happening October sixteenth through the twentieth. Go 10 00:00:45,159 --> 00:00:50,119 to Dcporto dot com to register and check out the full lineup of conferences at 11 00:00:50,200 --> 00:00:54,840 NDC Conferences dot com. Hey there, this is Jeff Fritz, the purple 12 00:00:54,880 --> 00:00:59,280 blazer guy from Microsoft, letting you in on a little secret about my friend 13 00:00:59,439 --> 00:01:03,560 Carl Franklin. You know, the guy who started dot net Rocks, the 14 00:01:03,640 --> 00:01:07,439 first podcast about dot net in two thousand and two. The guy who's been 15 00:01:07,480 --> 00:01:14,599 teaching Blazer on YouTube since twenty twenty. Yeah that Carl Franklin. Well, 16 00:01:14,959 --> 00:01:18,640 Carl's joined up with the folks from Code in a Castle to teach a week 17 00:01:18,760 --> 00:01:23,120 long hands on Blazer class at Are you ready to get this? At a 18 00:01:23,239 --> 00:01:30,200 castle slash villa in Tuscany. It's sort of a luxury vacation with Blazer learning 19 00:01:30,280 --> 00:01:37,359 built in. Carl's calling it the Blazer master Class. You'll learn Blazer from 20 00:01:37,359 --> 00:01:41,680 the ground up, finishing the week with the ability to build and deploy Blazer 21 00:01:41,719 --> 00:01:47,120 applications. Since the training happens for only four hours in the morning over six 22 00:01:47,200 --> 00:01:51,680 days, you can bring your significant other, your partner with you and you 23 00:01:51,719 --> 00:01:57,319 should right This part of Italy is absolutely beautiful. There's so much to see 24 00:01:57,400 --> 00:02:02,159 and do and in Larry and Marco from Code into Castle are organizing daily activities 25 00:02:02,200 --> 00:02:07,760 both at the castle and in the area. The castle is in the Marema, 26 00:02:07,840 --> 00:02:13,319 a less touristed region of Tuscany, offering both classic Tuscan hill country as 27 00:02:13,319 --> 00:02:17,759 well as easy access to the Etruscan Riviera, with sublime local food, wine 28 00:02:19,039 --> 00:02:23,120 and olive oil around every corner. Breakfast is included every day. There will 29 00:02:23,120 --> 00:02:28,919 be two communal dinners at the Castle book ending the experience and most other meals 30 00:02:29,080 --> 00:02:34,439 and all activities are included. And did I mention you'll learn Blazer in person 31 00:02:34,520 --> 00:02:38,159 from Carl Franklin. Listen, space is limited and for very good reason. 32 00:02:38,439 --> 00:02:45,000 This is quality training in a beautiful setting. Go to code in Acastle dot 33 00:02:45,000 --> 00:02:53,039 com slash Blazer twenty twenty three that's bla z o R two zero two three 34 00:02:53,080 --> 00:02:58,599 to take advantage of this amazing opportunity to join Carl in Tuscany for an unforgettable 35 00:02:58,639 --> 00:03:13,680 week of La dulce vita while advancing your programming skills in this important new technology. 36 00:03:17,120 --> 00:03:21,199 Welcome back to dot net rocks. I'm Carl Franklin, and this is 37 00:03:21,280 --> 00:03:25,639 Richard cattle in what's going on over on your side of the continent? My 38 00:03:25,639 --> 00:03:30,080 friend what packing boxes and hauling him up to the coast, Like after after 39 00:03:30,080 --> 00:03:34,280 we finished the show today, I'm literally loading the truck and taking a run 40 00:03:34,400 --> 00:03:38,719 up. We're trying not to end up with a storage locker, seeing how 41 00:03:38,719 --> 00:03:40,840 we're moving. We have a very big house that we're moving out over a 42 00:03:40,840 --> 00:03:46,120 lot of stuff into a smaller house that's already furnished. Yeah, so you 43 00:03:46,159 --> 00:03:50,680 know, it's just a juggling of a lot of stuff. But you're going 44 00:03:50,719 --> 00:03:53,560 to have a party. Oh, there's a few parties. There's been a 45 00:03:53,560 --> 00:03:55,319 few parties. Well, let me know, I might just make the trek. 46 00:03:55,439 --> 00:03:59,360 It's a long way out, but yeah, we'll figure something out. 47 00:03:59,400 --> 00:04:00,879 Certainly. When we're settled in up there, we'll we'll do something fun. 48 00:04:00,919 --> 00:04:05,560 But the it's good to be married to a woman who who tends to argue 49 00:04:05,599 --> 00:04:10,360 with the spreadsheet anyway, because didn't think it's very long to figure out how 50 00:04:10,439 --> 00:04:12,719 much stuff the garage holds said, Okay, that's ho much stuff we're allowed 51 00:04:12,719 --> 00:04:15,639 to keep, Yeah, and then condensed accordingly. Love it. Yeah, 52 00:04:15,920 --> 00:04:19,680 it's organized. Well, that's good to hear, Richard. Let's get started 53 00:04:19,720 --> 00:04:30,279 with better know a framework. Awesome, all man, what do you got? 54 00:04:30,519 --> 00:04:33,319 Well? This being show eighteen fifty seven. You can go to eighteen 55 00:04:33,439 --> 00:04:40,920 fifty seven dot twamp dot m E and that's a link to an article at 56 00:04:40,959 --> 00:04:45,279 Digital Trends, which I think is kind of important. Maybe it's important. 57 00:04:45,319 --> 00:04:48,399 We'll find out, but anyway, it's called here's why people think GPT four 58 00:04:48,600 --> 00:04:55,160 might be getting dumber over time. As impressive as GPT four was at launched, 59 00:04:55,240 --> 00:04:59,800 some onlookers have observed that it has lost some of its accuracy and power. 60 00:05:00,560 --> 00:05:02,279 First of all, was it ever accurate? I don't know, but 61 00:05:03,759 --> 00:05:11,439 sometimes I mean it mean the underlying Harvard study is pretty great. Yeah, 62 00:05:11,639 --> 00:05:14,959 and they are modeling it well. I thought the bigger message from this thing 63 00:05:15,120 --> 00:05:18,319 was they're doing updates and not telling anyone. Yes, right, like this 64 00:05:18,399 --> 00:05:23,920 is all guesswork from everybody else. Is that they're constantly refreshing this data set, 65 00:05:23,959 --> 00:05:27,319 it seems, and they're not really talking about where their sources are or 66 00:05:27,360 --> 00:05:31,680 anything like that. No, we don't, We don't really know. Yeah, 67 00:05:31,720 --> 00:05:35,319 Well, anyway, and also I've noticed that it don't ask it to 68 00:05:35,319 --> 00:05:42,360 do math. That's not his thing. It doesn't. It will lie and 69 00:05:42,439 --> 00:05:46,160 and be very adamant that it is correct. Yeah, it's uh yeah in 70 00:05:46,199 --> 00:05:50,439 the end, when you realize it's basically just text driven and you're hoping that 71 00:05:50,560 --> 00:05:54,920 the text on the Internet was accurate, which is a hilarious statement all by 72 00:05:54,959 --> 00:05:58,560 itself. It is absolutely hilarious. Well, yeah, I'm sure Amber's got 73 00:05:58,600 --> 00:06:01,519 some opinions about this we're talking about, but we'll bring her on in a 74 00:06:01,519 --> 00:06:05,000 minute. First of all, who is talking to us today, Richard Campbell? 75 00:06:05,240 --> 00:06:08,720 I grabbed a command of the show sixteen oh five, So that's going 76 00:06:08,759 --> 00:06:12,720 to back a bit in December of twenty eighteen, and not really about large 77 00:06:12,759 --> 00:06:15,639 language models, because who was talking about large languge of models in twenty eighteen. 78 00:06:15,680 --> 00:06:17,360 But this was a show we did with Jared Rhodes who's at MVP, 79 00:06:18,600 --> 00:06:24,040 and we were talking about IoT and edge computing at the time, and mostly 80 00:06:24,040 --> 00:06:28,480 what we were talking about was this continuous data collection mechanism, like just our 81 00:06:28,519 --> 00:06:31,120 ability to collect a lot of data about people, especially with so much compute 82 00:06:31,120 --> 00:06:34,959 in so many places, with so much ability to gather and stream that information 83 00:06:35,040 --> 00:06:40,399 back. And we ended up in an ethical conversation around data collection. And 84 00:06:40,439 --> 00:06:43,800 so Mike had this comment on the show shortly after it was publisher. He 85 00:06:43,839 --> 00:06:46,519 says, this is a great ethical discussion about collecting data. I always revert 86 00:06:46,600 --> 00:06:49,800 back to Kransberg's first low tech, which is it tech is neither good nor 87 00:06:49,879 --> 00:06:55,199 bad, nor is it neutral, which you know, I actually went and 88 00:06:55,240 --> 00:06:59,839 looked it up. It doesn't have to be right. It's like, again, 89 00:07:00,040 --> 00:07:04,879 we are personifying or anthromomorphizing technology. It's the people using the tool that 90 00:07:05,079 --> 00:07:09,600 is the issue. So it just is yeah, well, yeah, don't 91 00:07:09,680 --> 00:07:12,040 the thing is when we attribute it to the technology rather than the people who 92 00:07:12,120 --> 00:07:15,600 operated. Right, Right, we're giving agency where agency doesn't exist. The 93 00:07:15,720 --> 00:07:19,160 mentioned stores, data gathering is troubling. So this is one of the devices 94 00:07:19,199 --> 00:07:24,439 we were talking about. You obviously opt in by seeing the monitor and continue 95 00:07:24,480 --> 00:07:27,319 to enter, But is that enough. Shouldn't the customer know to what extent 96 00:07:27,360 --> 00:07:30,079 their behavior will be analyzed and for what purpose? What are the patrons who, 97 00:07:30,120 --> 00:07:32,959 beyond their own control, only have access to that store? Are you 98 00:07:33,000 --> 00:07:36,399 now alienating their business and taking advantage of the situation to gather data? So 99 00:07:36,439 --> 00:07:42,920 this is the store monitoring system where they're figuring out what people's buying behavior is 100 00:07:43,160 --> 00:07:46,959 just by having these sensors, right Yeah. Programming ethics have always fascinated me. 101 00:07:47,160 --> 00:07:49,920 Recent protests, of course, is the twenty eighteen if you can call 102 00:07:49,959 --> 00:07:55,920 back to then, recent protests by Google employees over developing censoring products for the 103 00:07:56,000 --> 00:08:01,040 Chinese government and military machine learning for the US government highlights the backlash she's company's 104 00:08:01,079 --> 00:08:05,079 face when adopting technologies for different applications. But it always boils down to this. 105 00:08:05,439 --> 00:08:07,319 You will always be able to find a developer to do what you want 106 00:08:07,360 --> 00:08:11,560 to have done. Right there, will always be someone with the ethics that 107 00:08:11,639 --> 00:08:13,720 align with yours or who's desperate enough to do that work for you. This 108 00:08:13,879 --> 00:08:18,279 is unique to programming. However, programming has a unique advantage of being cheap 109 00:08:18,360 --> 00:08:22,959 to scale, So one developer going to have a profound impact on humanity versus 110 00:08:22,959 --> 00:08:26,360 one unethical police officer or one unethical to investor. Had Bernie Madoff know how 111 00:08:26,360 --> 00:08:31,800 to make an asp dot net site, yeah, he could have offered his 112 00:08:31,879 --> 00:08:35,679 services anonymously to many more people. So it's always a great discussion on what 113 00:08:35,720 --> 00:08:39,600 to do responsibly. How do companies talk about this? My experience is most 114 00:08:39,799 --> 00:08:43,639 don't, but we will eventually need to. So just seemed to be on 115 00:08:43,759 --> 00:08:48,679 point with what we were talking about today, and certainly what I love is 116 00:08:48,759 --> 00:08:52,720 dragging out a conversation from five years ago, a gay pretty relevant, kind 117 00:08:52,720 --> 00:08:56,440 of on the point kind of there. Hey, Mike, thanks so much 118 00:08:56,440 --> 00:08:58,000 for your comment and a copy of us to Cobi. It's on its way 119 00:08:58,000 --> 00:09:01,000 to you, and if you'd like a copy used to go buy write a 120 00:09:01,039 --> 00:09:03,279 comment on the website at dot net rocks dot com are on the facebooks. 121 00:09:03,320 --> 00:09:07,080 We publish every show there and if you comment there and reading the show we'll 122 00:09:07,080 --> 00:09:09,960 send your copy mused to go by and you can definitely follow us on Twitter 123 00:09:09,000 --> 00:09:13,480 if you want, But we're on Mastodon. I'm on Mastodon a lot more 124 00:09:13,480 --> 00:09:15,720 of these days. I don't know about you, Richard, but I'm here 125 00:09:15,720 --> 00:09:16,759 and there, you know. Yeah, if you want to follow me on 126 00:09:16,840 --> 00:09:22,320 Mastodon, it's Carl Franklin at tech Hub dot social and I'm rich Campbell at 127 00:09:22,360 --> 00:09:26,759 mastodon dot social. Send us a two root two twot two two boot scoot 128 00:09:28,840 --> 00:09:33,440 you're back there, are you? Okay? Toot scoot boogie. Okay, 129 00:09:33,559 --> 00:09:39,639 let's bring in Amber. Amber McKenzie is head of data science and analytics at 130 00:09:39,799 --> 00:09:46,000 hr tech startup Fama Fama. Her background is in computer science and linguistics, 131 00:09:46,120 --> 00:09:48,919 and she's been doing data science, machine learning, and natural language processing for 132 00:09:48,960 --> 00:09:52,519 almost fifteen years now. Welcome back, Amber, Yeah, thanks for having 133 00:09:52,559 --> 00:09:56,039 me, holcome, thanks for being happy. And we had the ethics and 134 00:09:56,279 --> 00:10:01,639 AI conversation with you like four years ago. But I think when I wrote 135 00:10:01,639 --> 00:10:03,679 the email too, and like, I don't know if you've noticed, but 136 00:10:03,759 --> 00:10:07,480 some stuff has happened and you pretty much picked up that threat. You're like, 137 00:10:07,600 --> 00:10:11,600 yeah, a few things have happened, little stuff here and there Yeah, 138 00:10:11,600 --> 00:10:15,720 it's an interesting time. Are we now in a place where we're really 139 00:10:15,799 --> 00:10:20,240 challenging that debate like that four years ago and we were talking about sort of 140 00:10:20,559 --> 00:10:24,279 the bias in the machine learning models and that kind of thing. Is that 141 00:10:24,759 --> 00:10:28,720 is chat GPT? Is that writ large? You know, it's interesting to 142 00:10:28,759 --> 00:10:33,639 me and I've had this conversation a lot with people lately. There is sort 143 00:10:33,679 --> 00:10:41,480 of this idea that chat GBT is like some new way, different, next 144 00:10:41,559 --> 00:10:46,519 level, and to some degrees it is. But in my mind, it's 145 00:10:46,559 --> 00:10:50,720 not any different than some of the innovations we've had over the years. I 146 00:10:50,799 --> 00:10:56,279 mean, you know, we went from always sending letters to sending email. 147 00:10:56,360 --> 00:11:00,240 That was a heatge, big different thing, right we came out, you 148 00:11:00,240 --> 00:11:01,840 know, when Watson came out, everybody was like, Oh, that's going 149 00:11:01,919 --> 00:11:05,600 to change the game. That's going to be the new thing, right, 150 00:11:05,679 --> 00:11:11,279 and then it was Google and now it's chat GPT. And what essentially is 151 00:11:11,279 --> 00:11:16,200 going to happen is that the the opposite side, the people who are trying 152 00:11:16,200 --> 00:11:22,159 to regulate it, trying to create things to combat you know, bias or 153 00:11:22,320 --> 00:11:26,200 combat uh you know, deep fakes. That type of thing is just going 154 00:11:26,279 --> 00:11:30,399 to catch up, right, It's always kind of a game of came ross. 155 00:11:30,799 --> 00:11:35,120 Yeah, yeah, it always is. Yeah, but at least we 156 00:11:35,200 --> 00:11:39,559 know what we're racing against, like I would hope, Yeah, accuracy is 157 00:11:39,559 --> 00:11:46,000 actually important, it is. I mean I don't know. I look at 158 00:11:46,519 --> 00:11:50,600 Yeah, I liken it a lot to Google. We look to Google, 159 00:11:50,639 --> 00:11:54,840 and Google doesn't have all of the right things. And when we first came 160 00:11:54,840 --> 00:11:58,600 out, we were like, you know, oh, it's all magic information. 161 00:11:58,639 --> 00:12:01,399 And then everybody's like, well, you gotta check the sources. I 162 00:12:01,399 --> 00:12:05,120 would argue that when we used to all look up everything on encyclopedias, that 163 00:12:05,240 --> 00:12:11,480 wasn't all exactly right. So like it's just clearly though, the way people 164 00:12:11,480 --> 00:12:15,120 are perceiving its change is different than all of these other things. I mean, 165 00:12:15,200 --> 00:12:22,480 I don't remember that much kerfuffle between governments and agencies getting together and saying, 166 00:12:22,559 --> 00:12:24,879 what are we going to do about this? You know, we need 167 00:12:26,000 --> 00:12:28,720 regulation around this. I mean I guess they did with Google and Microsoft to 168 00:12:28,759 --> 00:12:35,120 some extent, but you know, other AI things haven't been Yes as hitting 169 00:12:35,159 --> 00:12:41,120 you over the head what this is. I'm glad to see it, though, 170 00:12:41,679 --> 00:12:45,480 I mean it took I remember, so I worked at Sandia in an 171 00:12:45,480 --> 00:12:50,679 internship and one of the problems they had at the time is all the national 172 00:12:50,759 --> 00:12:56,919 labs and all the agencies weren't even working together for you know, cybersecurity purposes. 173 00:12:58,320 --> 00:13:01,759 So you know, they would come up with a one person would come 174 00:13:01,840 --> 00:13:05,039 up with a solution for a zero day attack, and then they wouldn't share 175 00:13:05,080 --> 00:13:09,639 it. And that's what they were working on at the time, was just 176 00:13:09,759 --> 00:13:13,840 getting to a point where they would even share information. So the fact that 177 00:13:13,879 --> 00:13:16,960 we're already like, hey, we've got this new thing out there, we 178 00:13:18,000 --> 00:13:22,240 need to come together and put the guard wheels in place is for me kind 179 00:13:22,279 --> 00:13:26,159 of a refreshing sure take on things. And it's like, oh yeah, 180 00:13:26,200 --> 00:13:31,720 people recognize now that it's not just to put it out there, let it 181 00:13:31,799 --> 00:13:35,759 run wild thing. Let's figure out the best way to you know, especially 182 00:13:37,159 --> 00:13:43,799 for the sort of non technically inclined people who don't you know, have some 183 00:13:43,840 --> 00:13:46,919 of the understanding that is, oh, I've got to take this with a 184 00:13:46,919 --> 00:13:50,240 green as salt. I need to you know, check this or check that. 185 00:13:50,519 --> 00:13:54,039 So saying hey, let's let's try and put you know, the bowling 186 00:13:54,120 --> 00:13:58,919 bumpers on and make this thing, you know, go down the lane a 187 00:13:58,000 --> 00:14:01,919 little bit easier. Yeah, the thing that worries me about government regulation. 188 00:14:03,120 --> 00:14:07,279 Isn't the regulation itself. I think that's great, but it's that government takes 189 00:14:07,360 --> 00:14:13,519 so long to do things and to regulate, and this thing is evolving faster 190 00:14:13,720 --> 00:14:18,559 than anything we've seen before in tech, at at a much more rapid pace. 191 00:14:18,080 --> 00:14:22,759 For example, just read an article in The Verge that chat GPT can 192 00:14:22,799 --> 00:14:28,919 now remember who you are and what you want. With new custom instructions. 193 00:14:28,960 --> 00:14:31,279 You might not have to tell the chat about your life story every time you 194 00:14:31,320 --> 00:14:35,159 have a question, but do you know what you want the bot to know? 195 00:14:35,360 --> 00:14:39,399 I'll put a link to that. That doesn't surprise me. And I 196 00:14:39,799 --> 00:14:43,200 think in those instances, the private sector is going to be the one. 197 00:14:43,320 --> 00:14:50,639 I mean, there's going to be essentially the market opens up for yeah, 198 00:14:50,679 --> 00:14:54,120 for industry to say, hey, let's let's find a tech that combats that, 199 00:14:54,240 --> 00:15:00,879 and they'll be a market for that. People looking to to fill that 200 00:15:00,919 --> 00:15:05,879 gap. But I agree with you, the big cogs are slow. Yeah. 201 00:15:05,919 --> 00:15:09,559 I hope the government does big picture regulation rather than you know, picking 202 00:15:09,559 --> 00:15:13,639 on these features and those features which can easily change from under the regulations and 203 00:15:13,720 --> 00:15:18,320 screw everything up. That's where I like the data source one you brought up 204 00:15:18,320 --> 00:15:20,480 before it's like, just at least publish where we got your data from. 205 00:15:20,639 --> 00:15:24,480 Publish where that came from. Yet, I don't know if you see that 206 00:15:24,519 --> 00:15:28,159 as reasonable, Amber like, oh for sure, I mean that in terms 207 00:15:28,159 --> 00:15:33,480 of that's what's interesting about what's happened recently. For me, it is not 208 00:15:33,679 --> 00:15:37,759 the development of these algorithms. It's I mean, that's been happening. They're 209 00:15:37,799 --> 00:15:45,600 just taking it to a new level. It was the the mass market of 210 00:15:45,679 --> 00:15:50,120 it. They put marketing resources around it, they packaged it that they made 211 00:15:50,120 --> 00:15:56,120 it suddenly visible, and that changes the stakes of the game from something that 212 00:15:56,279 --> 00:16:00,600 is research based. You put papers around, people have to review the papers, 213 00:16:00,639 --> 00:16:06,360 they need to have reproducibility to you know. Now it is a product, 214 00:16:06,399 --> 00:16:10,039 a commodity, which puts it in a different space. Which, by 215 00:16:10,080 --> 00:16:12,799 the way, Microsoft just got a nice, big old fat stock price bump 216 00:16:12,919 --> 00:16:18,279 on like the marketing spoke and you're killing it in the stock market. We 217 00:16:18,399 --> 00:16:22,360 talked about this on Windows Weekly. It's like they put out an Inspire that 218 00:16:22,399 --> 00:16:26,320 it's thirty dollars a month per person and you're like, holy men, that's 219 00:16:26,320 --> 00:16:30,480 a lot of any three subscriptions only thirty six dollars. But if you want 220 00:16:30,480 --> 00:16:34,360 the copilot Now it's an extra thirty bucks. Stock price goes up five percent, 221 00:16:34,639 --> 00:16:40,080 Like if I'm sat an Adela, Like I'm laughing all the way to 222 00:16:40,120 --> 00:16:42,840 the bank, Like, that's a lot of money. What's your opinion amber 223 00:16:44,000 --> 00:16:48,840 on the general public's acceptance of this stuff. I mean, it's been my 224 00:16:48,480 --> 00:16:52,639 experience that people are kind of like, oh, that's cool, you know, 225 00:16:52,720 --> 00:16:56,320 and they use it when they can and they don't think too much about 226 00:16:56,320 --> 00:16:59,440 it. But I guess these new regulations, some of the stuff that they're 227 00:16:59,480 --> 00:17:04,880 floating around sound is we want to have some sort of watermark on anything that 228 00:17:04,920 --> 00:17:11,640 has been generated by AI, whether it's text or video or or you know, 229 00:17:11,759 --> 00:17:15,039 or whatever. But but that of course puts it in the hands of 230 00:17:15,079 --> 00:17:21,279 the content creators, and so you have the you'll have the problem where everybody 231 00:17:21,319 --> 00:17:25,240 who's following the law and everything will use the watermark, and then those who 232 00:17:25,319 --> 00:17:29,200 want to publish stuff for nefarious purposes will use it. Honey, that's the 233 00:17:29,279 --> 00:17:34,200 case of everything, right Yeah, you know, for me, it's it's 234 00:17:34,240 --> 00:17:40,200 one more thing that we have to teach about. I mean, right now, 235 00:17:40,240 --> 00:17:45,400 even in my kids English classes, they are taught about primary sources. 236 00:17:45,480 --> 00:17:48,319 They are taught about not just finding something on the web and taking it at 237 00:17:48,359 --> 00:17:52,400 face value, and that they can't just take that information roll it into an 238 00:17:52,519 --> 00:17:56,440 essay and their teachers are going to take it. That's going to be just 239 00:17:56,559 --> 00:18:03,720 one more thing in a realm that has to be taught. But I don't 240 00:18:03,759 --> 00:18:07,880 see I mean, you see degrees. It's almost like the Kevin Bacon thing. 241 00:18:07,920 --> 00:18:11,759 You see to the degrees of separation. It's like, you know, 242 00:18:11,000 --> 00:18:15,720 we're in tech, we are uniquely aware of it. And then you've got 243 00:18:15,799 --> 00:18:18,279 people who are pretty tech savvy who are like, yeah, this is cool, 244 00:18:18,640 --> 00:18:21,880 maybe you know, you know, like I have a friend who had 245 00:18:21,920 --> 00:18:25,400 it right as resignation letter, you know, stuff like that. And then 246 00:18:25,440 --> 00:18:29,200 you get, you know, to my parents who probably don't even know that 247 00:18:29,240 --> 00:18:33,839 it's happening right now. And it's like that degree of separation requires just more 248 00:18:33,839 --> 00:18:41,279 and more education about it. And I think that's what I hope the government 249 00:18:41,519 --> 00:18:42,759 sort of, I mean, even if it takes a little while, the 250 00:18:44,119 --> 00:18:48,519 sort of the guardrails, if you will, will essentially help those people who 251 00:18:48,440 --> 00:18:53,480 you know, aren't aren't going to go through the diligence of understanding it and 252 00:18:53,599 --> 00:18:57,680 understanding how to use it. And what to do with it and try and 253 00:18:59,079 --> 00:19:03,799 mitigate, um, you know, the damage that can be done there, 254 00:19:03,240 --> 00:19:07,680 because again, it's the same thing. I liken it to Google. I 255 00:19:07,720 --> 00:19:11,240 mean, I have grandparents who go and Google something and they're like, oh, 256 00:19:11,319 --> 00:19:12,240 but this, this, this, And I'm like, right, no, 257 00:19:12,920 --> 00:19:17,680 that's just not accurate. That's just not the case, right, And 258 00:19:17,920 --> 00:19:18,920 it's the same, you know, it's the same type of thing. And 259 00:19:18,920 --> 00:19:22,640 then the watermark stuff. I mean, I think we're going to come out 260 00:19:22,720 --> 00:19:30,079 that the software is going to come out to try and spot plagiarism, fakes, 261 00:19:30,160 --> 00:19:33,359 that type of thing. It's just going to have to catch up right 262 00:19:33,400 --> 00:19:38,759 at this twenty ugly it's but well I don't I don't know there's going to 263 00:19:38,839 --> 00:19:44,160 be any uglier than it already is. It's just becoming more obvious than it 264 00:19:44,240 --> 00:19:48,440 is. Ugly, Like I think your language they're amber about they Google this, 265 00:19:48,519 --> 00:19:52,000 so it must be true, right, Just there's no thought to the 266 00:19:52,079 --> 00:19:56,039 idea of where Google where where that data came from that Google surfaced for you 267 00:19:56,440 --> 00:20:00,559 is not in consideration for most people, right yep, or at least for 268 00:20:00,640 --> 00:20:03,960 some people anyway. Certainly, then we get into the confirmation biased thing where 269 00:20:04,000 --> 00:20:07,519 it's like I look for the link that sounds most like what I want must 270 00:20:07,519 --> 00:20:11,599 sound like, and that's the one I use absolutely, And in that respect, 271 00:20:11,640 --> 00:20:15,119 nothing has changed again. It's as you talked about earlier. It's the 272 00:20:15,160 --> 00:20:19,359 scalability issue. It just makes it more accessible. In the olden days, 273 00:20:19,400 --> 00:20:22,839 we just had you know, snake oil people, right, and they were 274 00:20:22,920 --> 00:20:26,000 still doing that right, Like they're going around in their wagons and they're telling 275 00:20:26,039 --> 00:20:30,200 people stuff and selling stuff, and it was harder for them to get around 276 00:20:30,279 --> 00:20:34,640 and spread that message, but it was still happening. It's just now it's 277 00:20:34,680 --> 00:20:38,319 easy, it's in your face, it's accessible. It just gets further and 278 00:20:38,440 --> 00:20:44,799 further that way. What if water marks become required and then there are penalties 279 00:20:44,880 --> 00:20:51,640 for those who don't use them. I can just see that turning into a 280 00:20:51,759 --> 00:20:56,440 shite show. I mean how I mean, when any we're going to put 281 00:20:56,440 --> 00:21:00,200 it? You know, I find it something to tweet and chat. GPT 282 00:21:00,359 --> 00:21:03,839 helped me write it. Do I have to put that on my tweet? 283 00:21:03,160 --> 00:21:07,400 And now does that mean that somebody's going to look at that and say, 284 00:21:07,440 --> 00:21:11,200 oh, that didn't come from my brain, you know what I mean? 285 00:21:11,359 --> 00:21:14,480 And then if I don't do it, when am I going to pay a 286 00:21:14,559 --> 00:21:15,960 fine? And how much is the fine going to be? Is it going 287 00:21:17,000 --> 00:21:18,559 to be more than Elon must charges me a month? To you sweeter, 288 00:21:18,559 --> 00:21:22,200 I can't imagine it'd be a fun I think more likely my viewer that said 289 00:21:22,319 --> 00:21:26,519 I will say, I'll have the option to say I only want stuff that 290 00:21:26,559 --> 00:21:30,079 has a watermark, and so it's your stuff just won't be visible if it 291 00:21:30,079 --> 00:21:33,640 doesn't pass that quality gate. Yeah, the market's going to dictate that. 292 00:21:33,799 --> 00:21:37,400 I mean to some degree. You're already seeing some of it where there have 293 00:21:37,519 --> 00:21:42,279 been ramifications for people that have used chat GPT and their jobs right or on 294 00:21:42,319 --> 00:21:47,799 their resumes or something like that, and it's already you've seen instances where it's 295 00:21:47,799 --> 00:21:52,839 come to kind of bite them. Those those things are going to happen. 296 00:21:52,680 --> 00:21:56,640 The people are still going to do some of the bad stuff and get away 297 00:21:56,640 --> 00:21:59,759 with it as they always do. And they're in my mind, we're gonna 298 00:21:59,839 --> 00:22:03,640 end in some middle ground. I don't think it's ever going to be as 299 00:22:03,680 --> 00:22:08,200 far as like super regulated. That just doesn't ever really happen. They're going 300 00:22:08,279 --> 00:22:11,880 to try, and there's going to be some backlash and then there's you know, 301 00:22:11,880 --> 00:22:15,279 we're going to cut out a bulk of the easy stuff that people are 302 00:22:15,279 --> 00:22:18,799 trying to do, and people will find that it's not worth it to lose 303 00:22:18,839 --> 00:22:22,599 their job or whatever. And right, we'll end up in a space where 304 00:22:22,640 --> 00:22:26,680 some people use it for good and some people use it for bad. And 305 00:22:26,359 --> 00:22:30,079 well, this is that Gartner hype cycle right where I think we've come off 306 00:22:30,119 --> 00:22:36,279 the peak of inflated expectations and we're headed down into the trough of disillusionment right 307 00:22:36,319 --> 00:22:40,720 now, and then you sort of climb your way back out into sort of 308 00:22:40,839 --> 00:22:45,200 reasonable expectations what it can and can't do. I mean, I've certainly been 309 00:22:45,240 --> 00:22:49,119 encouraged people don't ask, don't use this software to try and discover facts, 310 00:22:49,640 --> 00:22:55,160 right, use it to discover ideas. It's a good creative tool. But 311 00:22:55,400 --> 00:22:59,200 any fact you want that it spits out, you need to validate, and 312 00:22:59,240 --> 00:23:03,640 you'll find a stunning number of them are incorrect. Yes, running is right. 313 00:23:04,160 --> 00:23:08,119 Absolutely, Yeah, it's the classic one. The one that I first 314 00:23:08,119 --> 00:23:11,720 said, okay, well here we go down the trough was the lawyers that 315 00:23:11,920 --> 00:23:17,200 used it to write summaries to their cases that cited cases that simply did not 316 00:23:17,319 --> 00:23:21,240 exist. Didn't exist, yep, and they didn't check yep. And but 317 00:23:21,359 --> 00:23:26,559 you know who did check the judge and then they were in big trouble because 318 00:23:26,599 --> 00:23:30,519 it was their name on it. Yeah, you can't use the software wrote 319 00:23:30,559 --> 00:23:33,960 it. Excuse like, it's not a thing. You submitted this as you 320 00:23:34,599 --> 00:23:40,839 you're on the hook for submitting inaccurate information. Yeah. And and then you 321 00:23:40,880 --> 00:23:44,240 know, the industry, they'll you know, at that point we'll see and 322 00:23:44,319 --> 00:23:45,519 be like, oh, you know what, that's not lucrative for me to 323 00:23:45,559 --> 00:23:49,279 do. It doesn't It might save me time, but it doesn't that the 324 00:23:49,440 --> 00:23:52,519 risk is not worth the reward. And I gotta think that's just like a 325 00:23:52,559 --> 00:23:56,799 cautionary tales for lawyers full stop. Like how many lawyers are like, Okay, 326 00:23:56,839 --> 00:23:59,680 well we're not using this now because it's at a risk. Although it 327 00:23:59,759 --> 00:24:02,680 same time, I think that's an overreaction too. It's like check your facts, 328 00:24:03,160 --> 00:24:07,119 right, that's always a good idea by the way you set down guidelines. 329 00:24:07,279 --> 00:24:11,319 I mean, like my company sent out so they encouraged us to explore 330 00:24:11,559 --> 00:24:15,839 chat GBT find ways that it could kind of help us do things, but 331 00:24:15,880 --> 00:24:21,720 they also sent out a you know, essentially a regulatory doc that we need 332 00:24:21,720 --> 00:24:23,759 to sign that's like, hey, these are the boundaries, right, like, 333 00:24:23,839 --> 00:24:26,759 these are the ways to use it and the ways not to use it. 334 00:24:26,799 --> 00:24:30,559 And I imagine a lot of companies will do that, right, and 335 00:24:30,599 --> 00:24:34,559 it's like, hey, this is how we're okay with you guys using this 336 00:24:34,799 --> 00:24:38,039 tool. This is going to become a part of the employee handbook. Yep. 337 00:24:38,319 --> 00:24:44,599 Yeah, especially when you see Microsoft and presumably Google productizing it. Yeah. 338 00:24:44,640 --> 00:24:48,160 Absolutely, Now if the company is paying for it for you, they're 339 00:24:48,200 --> 00:24:52,839 better be guidelines. Yeah, you would, hope, I think so. 340 00:24:53,079 --> 00:24:56,079 I mean we, of course we're still premature here. This is coming out 341 00:24:56,079 --> 00:25:00,880 in early August. We haven't actually later hands on M three sixty five copilot, 342 00:25:00,920 --> 00:25:04,200 even though they've announced a price, like you can't just buy it yet, 343 00:25:04,880 --> 00:25:10,400 so we still don't actually know what it's going to do for us as 344 00:25:10,440 --> 00:25:12,880 that whether or not, as we're thirty dollars a month per per person. 345 00:25:14,200 --> 00:25:18,839 Wow, And that's where the you know, the market dictates, right, 346 00:25:18,960 --> 00:25:21,839 Yeah, they act. I mean, just with any other software, you've 347 00:25:21,880 --> 00:25:26,119 got to come out with the value to warrant the price. And if they 348 00:25:26,160 --> 00:25:29,279 don't, yeah, then people won't pick it up. Right. Well, 349 00:25:29,359 --> 00:25:30,519 I think this is the other side of this, which is that we're still 350 00:25:30,519 --> 00:25:33,480 trying to figure out if this is a product people are actually willing to pay 351 00:25:33,480 --> 00:25:37,559 for. Yeah, and because the other thing we don't really know is the 352 00:25:37,559 --> 00:25:41,359 cost of operations, Like we know what they're willing to charge for, but 353 00:25:41,480 --> 00:25:45,319 I fascinated to know how much it's costing Microsoft to run those models on our 354 00:25:45,400 --> 00:25:49,440 behalf essentially for free. You know, I don't know how many folks have 355 00:25:49,559 --> 00:25:55,920 signed up for Chat GPT plus and much less what it's costing chats GPT on 356 00:25:55,960 --> 00:25:59,799 the back end. I did. I have Chat GPT plus and then I 357 00:26:00,000 --> 00:26:03,839 and to use the API, and I found that, oh, that's a 358 00:26:03,880 --> 00:26:07,759 separate charge, the GPT four API. And not only that, but you 359 00:26:07,799 --> 00:26:11,000 have to put your name on a waiting list, right, So yeah, 360 00:26:11,039 --> 00:26:14,119 if you want to use the API, even to use stuff like the playground, 361 00:26:14,319 --> 00:26:18,079 which I've been using on the AI Bought Show with Brian McKay, you 362 00:26:18,119 --> 00:26:22,839 have to have you have to be you have to be validated, and pay 363 00:26:23,720 --> 00:26:27,079 extra for that because it does more than Chat GPT. I mean, I 364 00:26:27,119 --> 00:26:32,440 do feel a little bit good about this gatekeeping, Like it's not like this 365 00:26:32,480 --> 00:26:34,680 thing sitting out an open source and anybody can set one of these lms up 366 00:26:34,720 --> 00:26:38,599 and running. However, they want like the fact that you have to be 367 00:26:38,640 --> 00:26:44,799 an authenticated developer so they kind of know who you are, and then you're 368 00:26:44,799 --> 00:26:48,880 working against their instance like that at least gives me some sense of governance, 369 00:26:49,200 --> 00:26:52,279 right, Yeah, and I just don't I mean at this time, you 370 00:26:52,319 --> 00:26:56,920 know, it takes an enormous amount of resources to train pre train one of 371 00:26:56,960 --> 00:27:00,519 these, right, and then a lot of what they're doing over over the 372 00:27:00,559 --> 00:27:08,640 time is fine tuning, right, tweaking it and so and I'm seeing, 373 00:27:08,759 --> 00:27:14,319 you know, the job listings are out there now right to work on these 374 00:27:14,359 --> 00:27:17,599 things, to do some of the fine tuning and that sort of stuff. 375 00:27:17,640 --> 00:27:19,839 And that that too, kind of what you were talking about in the beginning 376 00:27:21,240 --> 00:27:26,759 about whether whether it's degrading or not. And that's I think it does so 377 00:27:26,799 --> 00:27:32,359 many things that trying to test a handful of things that it that it does 378 00:27:32,519 --> 00:27:37,880 does not really test the entirety of it and it what it's capable of that 379 00:27:37,960 --> 00:27:41,640 sort of thing. But that's going to be a it's going to take I 380 00:27:41,680 --> 00:27:48,720 think a good bit to try and manage that process. Right, they're fine 381 00:27:48,759 --> 00:27:52,000 tuning it and having to basically because they've put it out to the public. 382 00:27:52,119 --> 00:27:57,400 Look at how that's received just like a software product, right, you know, 383 00:27:57,720 --> 00:28:00,599 do you put out an update and people get all mad about it, 384 00:28:00,720 --> 00:28:03,839 right, and then you have to kind of go back and tweak it. 385 00:28:03,839 --> 00:28:08,240 It's going to take them a while to sort out that that management. I 386 00:28:08,279 --> 00:28:11,880 think I got it. I got a sense that when they put chat Gypt 387 00:28:11,000 --> 00:28:15,559 out in November of twenty two, it's because they had runned all the tests 388 00:28:15,599 --> 00:28:18,160 they could think of, and so now it was well, instead of continuing 389 00:28:18,160 --> 00:28:25,000 to pay for people to write prompts to validate that this behavior, we can 390 00:28:25,000 --> 00:28:27,759 get it for free by putting it out in public. Yeah right, you 391 00:28:27,799 --> 00:28:32,839 know, except for that whole hundred million users signing up in sixty days thing, 392 00:28:32,920 --> 00:28:36,880 which it was weird. You know, I don't think anybody planned on 393 00:28:36,920 --> 00:28:41,720 that. I mean, their reaction to that was clearly intentional, But I 394 00:28:41,759 --> 00:28:45,160 don't think anybody started in November of twenty two saying, you know what, 395 00:28:45,200 --> 00:28:47,359 one hundred million people are going to sign up for this in two months. 396 00:28:47,480 --> 00:28:49,599 Well, that's a new you know, in the last few years, that's 397 00:28:49,599 --> 00:28:53,119 a new space for us. It didn't used to be that things like that 398 00:28:53,200 --> 00:28:57,599 traveled that quickly, right, Even though, right we're all digital and that 399 00:28:57,640 --> 00:29:02,640 sort of thing. But now, I mean, my kids know about new 400 00:29:02,680 --> 00:29:07,599 things before I do, because it gets blasted out on TikTok and goes all 401 00:29:07,640 --> 00:29:11,799 over the place, right, And that's that's new, especially for us older 402 00:29:11,799 --> 00:29:15,720 people. Right, It's like that, how does everybody in my neighborhood now 403 00:29:15,759 --> 00:29:19,000 know about this thing? No? About this thing because TikTok, you know, 404 00:29:21,079 --> 00:29:25,039 craziness. Yeah, no, it's an interesting aspect to that. But 405 00:29:25,880 --> 00:29:27,240 you know, I'm not a big fan of the folks saying, hey, 406 00:29:27,240 --> 00:29:30,960 we should put a pause on this, you know, shouldn't be productized. 407 00:29:32,920 --> 00:29:36,920 My instinct is against that. And yet when we talk this way, it's 408 00:29:36,960 --> 00:29:40,960 like, it's pretty clear you probably productized too soon, like you have not 409 00:29:41,319 --> 00:29:45,160 tested the edges of this product. At the same time, I also see 410 00:29:45,200 --> 00:29:51,960 Microsoft looking at it going they're probably spending millions a month in Azure resources operating 411 00:29:52,000 --> 00:29:56,440 this thing. If you a customers signed up for it soon, you might 412 00:29:56,480 --> 00:29:59,680 have to turn it off. Like that's just a lot of money. So 413 00:29:59,720 --> 00:30:04,039 we're i'd stay stitty in I mean, and it's you know, it's a 414 00:30:04,039 --> 00:30:10,039 product essentially that has gone viral that you really didn't you know, I imagine 415 00:30:10,079 --> 00:30:14,119 they expected some but I'm with you. Imagine they didn't expect no sort of 416 00:30:14,119 --> 00:30:17,559 the degree to it, and it's like, how do you how do you 417 00:30:17,599 --> 00:30:19,519 manage that with any sort of product? Right, all of a sudden, 418 00:30:21,119 --> 00:30:25,359 everybody has it and they all have opinions about it, and you know, 419 00:30:26,079 --> 00:30:29,440 how do you scale that up? And do you pull it back or do 420 00:30:29,519 --> 00:30:33,720 you try and tweak it? And how you manage expectations and backlash and all 421 00:30:33,759 --> 00:30:38,119 of that stuff. I mean, I don't there's a reason I work at 422 00:30:38,160 --> 00:30:42,400 smaller company because there's just a lot of stuff there that I'm just not interested 423 00:30:42,440 --> 00:30:47,519 in managing. Because at the same time, you've got Google with bards saying, 424 00:30:47,559 --> 00:30:48,480 hey, this isn't ready yet. We're going to hold us back. 425 00:30:48,519 --> 00:30:51,960 And I think there's a sense of folks who're saying, oh, do you 426 00:30:52,000 --> 00:30:53,240 just say your thing's not working yet? It's like, I'm pretty sure the 427 00:30:53,279 --> 00:30:57,599 other thing isn't working yet either. But you know, there's all these degrees 428 00:30:57,599 --> 00:31:02,079 of working. They're playing game, right, and there's there's always a thing 429 00:31:02,119 --> 00:31:07,559 like you always have to take that that gamble when do you do something wins 430 00:31:07,599 --> 00:31:10,960 the right time? Am I you know, is it going to pay off 431 00:31:11,000 --> 00:31:14,400 for me to wait and put out something that's later, you know, maybe 432 00:31:14,440 --> 00:31:18,759 more mature and received better, or maybe they're missing out because they're missing out 433 00:31:18,759 --> 00:31:23,039 on this initial fervor. Yeah. Again, that's not stuff that I am 434 00:31:23,119 --> 00:31:26,759 very good at. So also a reason I will never run a company. 435 00:31:26,880 --> 00:31:30,039 It's a great start up story. It's like, did you want to be 436 00:31:30,079 --> 00:31:33,000 the first mover because first mover has advantages? Or do you recog or do 437 00:31:33,039 --> 00:31:37,319 you look at it as the pioneers the one with the arrows in his back? 438 00:31:37,119 --> 00:31:42,279 Right? Absolutely? Yeah, both can be true. It makes you 439 00:31:42,279 --> 00:31:45,720 think of that. Have you seen that F one show on Netflix. I 440 00:31:45,759 --> 00:31:48,720 wasn't really into F one until I saw it. It's really good. I 441 00:31:49,000 --> 00:31:53,480 highly recommend it. But they the racing cars, Yeah yeah, and they 442 00:31:53,599 --> 00:31:59,119 show a lot of like the behind the scenes like decisions about you know, 443 00:31:59,160 --> 00:32:02,640 if it's raining, do you decide to put your raining tires on now or 444 00:32:02,759 --> 00:32:06,799 later? Or try and extend the things, or do you right? You 445 00:32:06,839 --> 00:32:09,240 know, when do you move maneuver and all that stuff. There's just so 446 00:32:09,319 --> 00:32:15,319 much you know, intricacy that goes into that. And it's the same type 447 00:32:15,359 --> 00:32:17,240 of thing. It's like when do I put my car out or when do 448 00:32:17,319 --> 00:32:22,200 I hold it back. Could we then surge ahead and some of it pans 449 00:32:22,200 --> 00:32:24,319 out and some of it doesn't. Yeah, it's getting a three months jump 450 00:32:24,400 --> 00:32:29,440 on a market. Give me more advantaged in doing three more months of research? 451 00:32:30,119 --> 00:32:34,680 Right for at least it's a great question, And obviously we're seeing it 452 00:32:34,799 --> 00:32:37,599 play out right now and two different positions. I think to a large degree, 453 00:32:37,640 --> 00:32:42,160 I think, oh, it'll be the source of books in a couple 454 00:32:42,200 --> 00:32:45,039 of years, right, There'll be so many books about it and what worked 455 00:32:45,039 --> 00:32:49,720 and what didn't and what you can learn from your for your own company in 456 00:32:49,759 --> 00:32:52,359 the future. Absolutely the race for AI, No, I see it, 457 00:32:52,400 --> 00:32:55,599 absolutely, And Amber, I'm gonna interrupt for one moment if it's very important 458 00:32:55,640 --> 00:33:02,440 message and we're back. It's done in rocks. I'm Richard Gamble. Let's 459 00:33:02,480 --> 00:33:07,680 Carl Franklin. You talking to our friend Amber Mackenzie a little bit about Well, 460 00:33:07,720 --> 00:33:09,920 I just called this the ethics of large language models. But I think 461 00:33:10,359 --> 00:33:15,559 there is some ethical conversation going on here as we're trying and figure out the 462 00:33:15,640 --> 00:33:19,519 consequences of what's being built and being basically played out in front of us. 463 00:33:19,759 --> 00:33:22,359 I wonder how much of this also is just the culture of modern software development. 464 00:33:22,359 --> 00:33:25,920 Where it's like just sticking in front of the customers. We have telemetry 465 00:33:27,279 --> 00:33:30,480 and we'll fix it via the internet going forward. Yeah, you don't. 466 00:33:30,519 --> 00:33:34,720 It's almost like you don't have to be responsible anymore because it's all fixable. 467 00:33:34,880 --> 00:33:38,440 It's true, but also like you put this thing out and oh you don't 468 00:33:38,440 --> 00:33:40,720 like how it works. Do you think it's getting worse? Well, I've 469 00:33:40,759 --> 00:33:45,400 got all these other people lined up to use it, you know that sort 470 00:33:45,400 --> 00:33:49,200 of thing. It's like, well, we'll get the feedback. The feedback 471 00:33:49,279 --> 00:33:54,519 is oftentimes more valuable than anything else, right, Yeah, well that's certainly 472 00:33:54,559 --> 00:34:00,000 where chat GPT started, was just a feedback mechanism for more prompts and more 473 00:34:00,319 --> 00:34:02,880 data on how it was behaving and how people were reacting to it. So 474 00:34:04,920 --> 00:34:07,920 it's just, you know, this sudden flip to make it make it a 475 00:34:07,960 --> 00:34:12,679 product. It's interesting to see if it's going to be good and what that 476 00:34:12,760 --> 00:34:19,440 even looks like. I'm still not convinced that this isn't a dead end because 477 00:34:19,480 --> 00:34:22,280 my experience with software has generally been if if the only way you know how 478 00:34:22,320 --> 00:34:24,960 to make it better is to make it bigger, it's probably not that good. 479 00:34:25,960 --> 00:34:31,519 Yeah, I've been saying that about LM's lately. They're just I'm one 480 00:34:31,519 --> 00:34:37,079 of those, uh when I non jumpers is what I call myself. So 481 00:34:37,199 --> 00:34:40,840 I turned to wait on a lot of things, like if my company rolls 482 00:34:40,880 --> 00:34:44,719 out a new software that we're going to be using, I'm like, are 483 00:34:44,719 --> 00:34:46,119 we I'm not going to dive all into it. I'm going to wait to 484 00:34:46,159 --> 00:34:50,079 see if we're really going to use that, or in a couple of months 485 00:34:50,079 --> 00:34:52,079 they're going to be like, ah, no, we made the wrong decision, 486 00:34:52,440 --> 00:34:58,119 right. And so I actually have not been as all in on it, 487 00:34:59,320 --> 00:35:00,440 you know. I think that there's probably, yeah, I could I 488 00:35:00,480 --> 00:35:05,119 spend some time and really find ways to help my productivity or that sort of 489 00:35:05,159 --> 00:35:07,719 thing. Yeah, but it hasn't been this thing that I've been like, 490 00:35:07,800 --> 00:35:13,000 oh my gosh, that makes all the difference. I've got to use it, 491 00:35:13,480 --> 00:35:15,679 you know, and I'm waiting to see where it goes. I Mean 492 00:35:15,719 --> 00:35:20,440 a lot of times that's true too, that if you're the second mover, 493 00:35:21,199 --> 00:35:23,199 you don't get the arrows in the back and you get the advantages of their 494 00:35:23,320 --> 00:35:25,559 hits to say, Okay, this is a better way to use it. 495 00:35:27,000 --> 00:35:30,079 Yeah, and maybe some better tooling, Like there's a lot of second mover 496 00:35:30,199 --> 00:35:35,639 advantage. There is English the only language that CHAT Gypt and GPT for uses 497 00:35:36,320 --> 00:35:39,119 and is that an issue? So I don't know the answer, would I 498 00:35:39,159 --> 00:35:45,280 would assume not. I mean, the machine translation space is pretty solid and 499 00:35:45,320 --> 00:35:47,960 we have a lot of data and other languages, but I actually don't know, 500 00:35:50,039 --> 00:35:52,440 right, But is it translation at that point? I mean, because 501 00:35:52,800 --> 00:35:57,400 language models are that right, it's trained on language. It's a yeah, 502 00:35:57,440 --> 00:36:00,440 you know, there's actually been a way CHAT is available in fifty languages. 503 00:36:00,559 --> 00:36:06,239 Okay, that's first and foremost right, But there have been some great papers 504 00:36:06,360 --> 00:36:13,280 about the tokenization strategy and how it doesn't map symmetrically to all languages because different 505 00:36:13,360 --> 00:36:16,320 languages have different architectural portrays, and because it was trained on English, wasn't 506 00:36:16,320 --> 00:36:22,800 it Well, fundamentally the tokenizations technique was built by English speakers. Yeah, 507 00:36:22,800 --> 00:36:25,599 like I wouldn't even say, it's just it's the training set, per se 508 00:36:25,880 --> 00:36:30,280 in that of the adeomatic structure of the tokens is based on people who were 509 00:36:30,280 --> 00:36:36,199 thinking in English. And so there was a really great piece I read about 510 00:36:36,280 --> 00:36:39,639 its mistakes in Urdu, not that I speak or do, and I know 511 00:36:39,679 --> 00:36:45,199 it's in Pakistan. It's a very old language, but architectural fundamental different language, 512 00:36:45,199 --> 00:36:47,639 and the token system just didn't work as well for that one appen. 513 00:36:47,719 --> 00:36:51,199 Yeah, and it just begs the question like if Grant we're here, he 514 00:36:51,239 --> 00:36:53,239 would be saying, what's language anyway? You know, it changes all the 515 00:36:53,280 --> 00:37:00,400 time, So what's formal English language today or Urdu language today? We'll change 516 00:37:00,440 --> 00:37:05,800 in ten years. I'm reminded of that awesome show night Rider. Do you 517 00:37:05,840 --> 00:37:09,639 remember that night Rider with the kit kit the car, the AI car, 518 00:37:09,760 --> 00:37:14,440 right, And I thought one of the funniest things I ever heard about that. 519 00:37:14,440 --> 00:37:16,960 One of the funniest bits was from Eddie Murphy who's like, you know, 520 00:37:17,679 --> 00:37:21,360 the door is a jar. He says, you know, if that 521 00:37:21,440 --> 00:37:24,360 were in the hood, the like by yeo, man, someone stole your 522 00:37:24,360 --> 00:37:32,360 battery him. It's true though, Yeah, that's actually a big I mean 523 00:37:32,480 --> 00:37:38,159 that speaks back to the data bias situation, where, right, the amount 524 00:37:38,199 --> 00:37:43,320 of data that you have in different languages or different dialects or whatever, it 525 00:37:43,480 --> 00:37:49,400 really does dictate sort of the outcome that you're going to get from any of 526 00:37:49,400 --> 00:37:54,719 these models, and most of it isn't you know, more English oriented or 527 00:37:54,760 --> 00:38:00,320 white oriented English? Right? Yeah, white English yep. And part of 528 00:38:00,320 --> 00:38:04,599 this also is now this film. It seems to be the philosophy that a 529 00:38:04,679 --> 00:38:07,719 larger data set is better. Yeah, and so like and I look at 530 00:38:07,719 --> 00:38:09,559 it, go, I go over and look at Wikipedia. Say, there's 531 00:38:09,559 --> 00:38:14,360 far more English articles than anything else. Are they more accurate because there's more 532 00:38:14,519 --> 00:38:17,800 and more participation or are there other languages with fewer articles that are actually higher 533 00:38:17,880 --> 00:38:23,280 quality. Well, the argument that more data is that are it has nuances? 534 00:38:23,400 --> 00:38:28,480 Right? So if you are doing a particular thing, having more data 535 00:38:28,639 --> 00:38:31,159 that speaks to the thing that you need to do, Yes, that is 536 00:38:31,199 --> 00:38:37,400 accurate. But say you're doing in classification, you've got five classes. Having 537 00:38:37,519 --> 00:38:40,559 more data in one of your classes that you have the most data for is 538 00:38:40,599 --> 00:38:45,199 actually it degrades the model for the other ones. So having more data in 539 00:38:45,239 --> 00:38:52,480 that first bucket, you know, say white English, having more data there 540 00:38:52,519 --> 00:38:57,679 does actually not help. It actually will bias the model towards those things. 541 00:38:57,840 --> 00:39:01,719 Right. So you need more data in a diverse area. You need more 542 00:39:01,800 --> 00:39:06,360 data in the places that you don't have a lot of data, or if 543 00:39:06,360 --> 00:39:08,639 you don't have data about a certain subject, or if you don't have it 544 00:39:08,960 --> 00:39:13,159 you know, data and a certain thing you're trying to do. That helps, 545 00:39:13,239 --> 00:39:15,639 but not across the board. Yeah, I mean, I guess you 546 00:39:15,719 --> 00:39:21,039 have to when tokenizing, you have to determine, Okay, does this apply 547 00:39:21,159 --> 00:39:25,360 to every culture, every language or is this a language or culture specific thing? 548 00:39:25,679 --> 00:39:29,119 Right? Yeah, I don't know the answer. I'm not smart enough 549 00:39:29,119 --> 00:39:31,960 to do that to know that well. And it's it's twofold. It's the 550 00:39:32,159 --> 00:39:37,599 linguistic part that is sort of the translation or the how do I you know, 551 00:39:37,920 --> 00:39:42,000 when I'm setting up questions, when I'm setting up prompts and I'm translating 552 00:39:42,000 --> 00:39:45,519 those into other languages, or even if I have a model that I'll take 553 00:39:45,559 --> 00:39:47,639 it from other languages, if I don't have enough data to support that, 554 00:39:47,679 --> 00:39:51,960 then it's not going to be as good. But then the cultural differences, 555 00:39:52,320 --> 00:40:00,760 So if if even if it's the same English, if other cultures questions in 556 00:40:00,800 --> 00:40:02,840 a different way, if they speak in a different way, if they provide 557 00:40:02,840 --> 00:40:07,000 their thoughts in a different way, then it's not going to be as effective. 558 00:40:07,199 --> 00:40:09,119 Yeah, yeah, yeah, And I think this is where like you 559 00:40:09,159 --> 00:40:15,320 can go to the platform tokenizer on open Ai and actually see how it tokenizes 560 00:40:15,360 --> 00:40:22,159 English GPT three just to get an idea of what they're talking about, and 561 00:40:22,199 --> 00:40:29,039 then recognize how English centric that actually is like, again, I don't know 562 00:40:29,159 --> 00:40:31,679 enough language to say now if I was fluent in and I would prefer not 563 00:40:31,719 --> 00:40:37,800 even a romance language, something very different. Do the tokens actually makes sense? 564 00:40:37,400 --> 00:40:39,599 I don't know how much, really, how much work has been done 565 00:40:39,599 --> 00:40:44,440 in that area, Richard. When you say you can see how it's tokenized, 566 00:40:44,559 --> 00:40:49,400 is it tokenizing your questions or the data that it's trained on? It 567 00:40:49,559 --> 00:40:52,480 you can literally just write some text and it'll show you what about the data 568 00:40:52,519 --> 00:40:55,719 that it's trained on. That's that's my No, it doesn't do that, 569 00:40:55,840 --> 00:40:59,599 But I'm just saying, you know, folks are asking about tokenization, and 570 00:40:59,599 --> 00:41:02,800 it's look, I'll clue the link in the shows, but you can literally 571 00:41:04,159 --> 00:41:07,119 take text, plug it into that, and it'll show you how the tokenizer 572 00:41:07,440 --> 00:41:12,320 represents it well. And this ends up being a problem. Actually, the 573 00:41:13,000 --> 00:41:19,800 productization of it becomes a problem there because products are oriented towards their users. 574 00:41:20,880 --> 00:41:23,559 So you know, if you're making it a product and you're trying to sell 575 00:41:23,599 --> 00:41:28,920 it, you're going to make it work best for the majority of the people 576 00:41:29,079 --> 00:41:34,320 who want to buy it, and that oftentimes is not the marginalized communities. 577 00:41:34,920 --> 00:41:38,039 Well, and yeah, and you get into now you have all the market 578 00:41:38,079 --> 00:41:43,000 forces, which is what's the appropriate demographic for me to optimize return especially in 579 00:41:43,079 --> 00:41:45,719 the early days of a product like this where they're desperate to get some returns 580 00:41:45,760 --> 00:41:50,400 quickly. Right, They're not looking at diversity or any of those things. 581 00:41:50,440 --> 00:41:52,559 They're like, who's got the money that we know how to sell to now? 582 00:41:52,719 --> 00:41:55,320 Right? Absolutely? I mean, and again I go back to M 583 00:41:55,360 --> 00:41:59,519 three sixty five, which copile, which seems to be the first one really 584 00:41:59,559 --> 00:42:07,559 product since Gehub Coopilot, where they're targeting large American corporations. Yep, because 585 00:42:07,679 --> 00:42:12,039 they because they're already they're already customers, they're going to use it on mass 586 00:42:12,039 --> 00:42:15,039 like there's a lot of reasons. I totally get why they would do that, 587 00:42:15,239 --> 00:42:19,360 but has nothing to do with any kind of sense of diversity to the 588 00:42:19,440 --> 00:42:22,760 data. That's probably a bigger concern in my mind than you know, some 589 00:42:22,800 --> 00:42:28,000 of the misinformation stuff. The misinformation has happened, and it will happen, 590 00:42:28,119 --> 00:42:31,519 and people will make things to combat that, but there isn't a market to 591 00:42:31,840 --> 00:42:37,719 combat the lack of diversity and bias. There are smaller groups who work towards 592 00:42:37,719 --> 00:42:42,360 it. But again, if it's a product and it's somebody keeps it, 593 00:42:42,840 --> 00:42:45,599 you know, privately held and that sort of thing. There isn't an impetus 594 00:42:45,679 --> 00:42:51,519 for them to to fix that, and that widens the widens the tech gap, 595 00:42:52,039 --> 00:42:55,639 widens the gap for people having you know, access to things that other 596 00:42:55,679 --> 00:43:00,320 people have access to. It sort of heads us in a further in a 597 00:43:00,360 --> 00:43:05,760 direction that we have been going and as where you really have the argument about 598 00:43:05,800 --> 00:43:10,280 making a product in this particular scenario is that because of the need to show 599 00:43:10,360 --> 00:43:16,119 returns quickly, you're going to to bias towards where those returns are likely to 600 00:43:16,239 --> 00:43:20,360 come from. Yeah, although and I'm in our past conversation you said like 601 00:43:20,480 --> 00:43:22,000 step one of dealing with any biases, No, it's there. Yeah, 602 00:43:22,000 --> 00:43:25,599 what's step two? Well, in this case, the app to care about 603 00:43:25,639 --> 00:43:30,880 it. Yeah, but I think we always did, but or we always 604 00:43:30,880 --> 00:43:34,880 supposed to. But that's again where it comes into the product side, is 605 00:43:34,920 --> 00:43:37,800 like, yeah, when you take it out of the research right, research, 606 00:43:38,559 --> 00:43:44,320 the research arena oftentimes will penalize for bias for things like that because you 607 00:43:44,400 --> 00:43:47,119 have a diverse group of researchers who were like, hey, did you think 608 00:43:47,119 --> 00:43:51,039 about this thing? And we care about that, We're not going to you 609 00:43:51,079 --> 00:43:53,039 know, approve your paper or whatever it is, if you haven't done your 610 00:43:53,079 --> 00:43:57,440 due diligence, many you take it out of that realm and put it into, 611 00:43:58,000 --> 00:44:01,440 you know, a commercial product. What do they get for increasing the 612 00:44:01,559 --> 00:44:06,679 diversity there? What benefit does it do them? In fact, it's probably 613 00:44:06,920 --> 00:44:09,559 worse for them because they have to put more time and resources into something that 614 00:44:09,599 --> 00:44:15,039 they're not going to get a return from. So yeah, it's to me, 615 00:44:15,320 --> 00:44:17,639 you know, and again I've done enough product work to be pretty comfortable 616 00:44:17,639 --> 00:44:22,480 with these terminologies. It is like you go get your pioneering group and you 617 00:44:22,599 --> 00:44:24,400 kind of want them to all be the same because you've only got so many 618 00:44:24,480 --> 00:44:32,119 resources. That diversity quotient only comes in when we're trying to broaden the customer 619 00:44:32,159 --> 00:44:38,840 race. It's literally the second half before that. He becomes a question absolutely 620 00:44:38,840 --> 00:44:43,320 where it's like we are now limiting our market by not being broad enough. 621 00:44:43,559 --> 00:44:46,360 And if you think about the people who are going to pay for chat GPT, 622 00:44:46,519 --> 00:44:51,320 it's often not going to be, you know, a diverse group. 623 00:44:51,360 --> 00:44:52,960 There's a whole bunch of this world that is not going to be able to 624 00:44:53,000 --> 00:44:59,039 pay for Chat GPT. It's already priced at the monthly income of a true 625 00:44:59,119 --> 00:45:02,039 non trivial chunk of the this world. So uh, and that and that's 626 00:45:02,039 --> 00:45:06,039 a whole other can of worms, right, is that you know we already 627 00:45:06,039 --> 00:45:09,719 have multinationals where pricing varies by country by country, and that we could get 628 00:45:09,800 --> 00:45:14,039 there. I don't know that that's true. And I would also say we're 629 00:45:14,079 --> 00:45:17,119 arguing over a product that's I think yet to show profound value. Yeah, 630 00:45:17,159 --> 00:45:22,960 it's certainly created profound hype, but show me that, you know, the 631 00:45:23,079 --> 00:45:29,079 company with a huge competitive advantage because they've been using this technology. Lots of 632 00:45:29,079 --> 00:45:32,239 people are talking about it. They're certainly using excuses to lay people off over 633 00:45:32,320 --> 00:45:36,400 it. But you know, I want to see two quarters after that and 634 00:45:36,519 --> 00:45:38,719 say, did you actually get better? Did you were your customers happier? 635 00:45:38,880 --> 00:45:43,079 Like, I don't know that any of us know the answer to that those 636 00:45:43,119 --> 00:45:45,639 things yet. Yeah, I think there's going to be opportunities, especially in 637 00:45:45,679 --> 00:45:51,440 the sort of data science and machine line consulting room, to really speed up 638 00:45:52,000 --> 00:45:58,440 those consulting opportunities, right, to take them from something that you know you 639 00:45:58,440 --> 00:46:01,639 you have to kind of train a new thing every time and tailor it to 640 00:46:01,679 --> 00:46:05,760 your client. But then to get ahead of some of that competition, say 641 00:46:05,800 --> 00:46:07,760 hey, we can significantly cut down on that time. But for all the 642 00:46:07,760 --> 00:46:13,719 companies that don't necessarily have in house data science, machine learning, that sort 643 00:46:13,719 --> 00:46:16,840 of thing, you know, are they suddenly going to be able to pick 644 00:46:16,920 --> 00:46:22,320 up chat GPT and have it make a big difference in that area. Yeah, 645 00:46:22,440 --> 00:46:25,360 I don't know about that. Yeah, I think it's a great question. 646 00:46:25,440 --> 00:46:30,039 And funny thing is, like, this is what Microsoft's has been known 647 00:46:30,159 --> 00:46:35,599 for, is the commoditization of certain technologies, sure, making it very easy 648 00:46:35,679 --> 00:46:40,360 to get into things so that typically needed experts. You could use these tools 649 00:46:40,639 --> 00:46:45,800 without the experts, and you wouldn't get a superior experience. Like arguably we'd 650 00:46:45,800 --> 00:46:51,239 constantly get an infior experience, but perhaps good enough that it was worthwhile. 651 00:46:51,639 --> 00:46:58,320 Yeah. I think it goes back to them not anticipating how big it would 652 00:46:58,320 --> 00:47:02,960 get immediately, right. Yeah. I don't want to present the idea that 653 00:47:04,039 --> 00:47:07,480 anyone had a plan here because I don't think it's true. I think they're 654 00:47:07,519 --> 00:47:12,880 all making it up as they go. Right now, that the hundred million 655 00:47:13,039 --> 00:47:17,760 users wiled them, so they pointed at Bying of all places, and Bying 656 00:47:17,920 --> 00:47:22,480 got a hundred million users, you know, like this is one hundred million 657 00:47:22,599 --> 00:47:27,960 user machine. Let's go. Yeah, it's an interesting space right now. 658 00:47:28,159 --> 00:47:32,239 Well, and I'm excited to talk to you because you come from the science 659 00:47:32,280 --> 00:47:37,519 side of this and the academic knowledge side of this. And I can see 660 00:47:37,000 --> 00:47:42,760 because we could see each other while we're recording this, Like the productization of 661 00:47:42,800 --> 00:47:46,519 this stuff is the worst I think for someone from your spaces, like you're 662 00:47:46,559 --> 00:47:51,800 doing what with my thing? Where? How? And literally it's like that 663 00:47:51,960 --> 00:47:54,039 sort of grotesque. And then they took the sledgehammer to it to get it 664 00:47:54,079 --> 00:47:59,800 to fit into the square, like just hammering it through a box to get 665 00:47:59,800 --> 00:48:05,039 it to customers. There have been some musical artists who are demanding that or 666 00:48:05,159 --> 00:48:10,119 who are who are offering um royalties or something like that or demanding royalties. 667 00:48:10,119 --> 00:48:14,920 What's the story there, Richard. I'm confused about this, but I remember 668 00:48:14,960 --> 00:48:17,679 talking to you about it. I mean, there's there's certainly a question about 669 00:48:19,760 --> 00:48:22,480 uh yeah, rights to data, rights to the data used in the training 670 00:48:22,519 --> 00:48:29,400 models. And you can see the thought that as a developer running these models, 671 00:48:30,679 --> 00:48:32,920 because they're all you believe, they're all behind the scenes, so it 672 00:48:32,960 --> 00:48:37,320 doesn't really matter. I'm not taking a copy of your thing, I'm not 673 00:48:37,480 --> 00:48:42,360 using your thing per se. It's just being munged in a big neural net, 674 00:48:42,760 --> 00:48:46,039 like why should I care? And everything was fine until Dolly started spitting 675 00:48:46,039 --> 00:48:51,079 out Getty logos. Then it wasn't fine. Yes, right, this is 676 00:48:51,199 --> 00:48:53,000 there's there was no intent here, There was no plan here. I think 677 00:48:53,000 --> 00:48:58,039 a devs have it always been doing what devs were doing until the logo, 678 00:48:58,400 --> 00:49:01,400 until the logo started appearing in the pictures, and then it's like that, 679 00:49:02,039 --> 00:49:08,320 huh. And I love that it was Getty because Getty is the greatest pursuer 680 00:49:08,400 --> 00:49:14,599 of copyright you've ever seen. Heaven help you if you use a Getty picture 681 00:49:14,599 --> 00:49:17,199 without permission. Ah, it's all a big I mean, it just opened 682 00:49:17,280 --> 00:49:21,039 up a lot of stuff that I think people have been talking about for a 683 00:49:21,079 --> 00:49:27,079 while, but that the like need wasn't so great that people had to really 684 00:49:27,239 --> 00:49:30,320 pursue, you know, some of these I mean, we had to do 685 00:49:30,360 --> 00:49:34,559 some of that, right, Napster days and then there's you know, there's 686 00:49:35,519 --> 00:49:39,199 Spotify days and all of that stuff. Everything's always changed and it's just one 687 00:49:39,239 --> 00:49:45,960 more time that those things are gonna have to change. But they just the 688 00:49:45,159 --> 00:49:50,400 speed with which it came out, it's going to take a little bit for 689 00:49:51,039 --> 00:49:55,039 the extra things to catch up. And be like, all right, we 690 00:49:55,119 --> 00:50:00,199 have been having this conversation for years, including the three of us. Yeah, 691 00:50:00,360 --> 00:50:05,199 Like, in some ways this is a useful forcing function. Okay, 692 00:50:05,199 --> 00:50:08,920 I've now finally made a thing that's showing up in the public gestalt. Now 693 00:50:08,960 --> 00:50:16,119 we've been talking about treating data maturely, respecting copyright, considering bias, like 694 00:50:16,239 --> 00:50:20,960 all of those things. Now ya gotta because we're now starting to stick it 695 00:50:20,960 --> 00:50:25,440 in front of regular people who don't know how problematic these data collections and utilization 696 00:50:25,480 --> 00:50:30,199 methods have been. Yeah, I mean that's what normally drives innovation, right, 697 00:50:30,440 --> 00:50:35,800 every time, other innovation. There's no driver's licenses until cars started killing 698 00:50:35,880 --> 00:50:40,159 people. Yeah, true, right, Just hopefully we can get we can 699 00:50:40,199 --> 00:50:44,760 get to a point of some moderation with this before folks die. I would 700 00:50:44,840 --> 00:50:49,800 rather not have folks die over this crazy bit of software. Yeah, crazies, 701 00:50:49,880 --> 00:50:52,199 right man, people are going nuts. It's true. I had the 702 00:50:52,199 --> 00:51:00,000 same thing comes to fruition on the environmental impact side, because these models are 703 00:51:00,000 --> 00:51:05,119 are very extensive and very impactful to train. Yeah, and now that they 704 00:51:05,119 --> 00:51:08,039 are out there and people are using them, that likely means that more of 705 00:51:08,079 --> 00:51:10,360 them are going to be trained, and more of them are going to be 706 00:51:10,400 --> 00:51:14,039 deployed, and it's like, okay, how are we going to then also 707 00:51:14,159 --> 00:51:16,639 mitigate that impact as well? Well. This one good thing that comes out 708 00:51:16,639 --> 00:51:20,760 of it is now people won't be complaining so much about how much water it 709 00:51:20,760 --> 00:51:25,719 takes to grow a stake. That's true. I always thinking more in terms 710 00:51:25,760 --> 00:51:30,840 of you know, Jess Wing Crypto was winding down and giving back those resources. 711 00:51:30,840 --> 00:51:34,320 We found a new place to put them. Yeah, exactly, absolutely. 712 00:51:34,800 --> 00:51:37,639 Yeah. And if you want to make an AI person angry in a 713 00:51:37,679 --> 00:51:40,719 big hurry, just call it the new Crypto. That's a good way to 714 00:51:40,760 --> 00:51:45,239 do it. Oh yeah, we will not get a Christmas card from that 715 00:51:45,280 --> 00:51:49,519 person. That's not a happy thing to say. That's awesome. And I 716 00:51:49,559 --> 00:51:52,519 also think it's not true, like it doesn't seem to be trying to separate 717 00:51:52,519 --> 00:51:57,559 people with our money quite as vociferously. But the bill, you know, 718 00:51:57,559 --> 00:52:00,559 the rates are the prices are starting to come. Man. I guess we'll 719 00:52:00,559 --> 00:52:04,239 see what happens. Yeah, it's true. It'll be an interesting back half 720 00:52:04,280 --> 00:52:07,199 of the year. Yeah. So Amber, what's next for you? What's 721 00:52:07,199 --> 00:52:13,280 in your inbox? Ah? Well, I am doing good things at Kauma. 722 00:52:13,440 --> 00:52:17,679 I'm happy to be there. And for me, I'm finally just in 723 00:52:19,320 --> 00:52:24,880 a steady state, which is great, and looking forward to kind of just 724 00:52:24,920 --> 00:52:29,480 seeing where things go. I've been throwing a lot of energy into sort of 725 00:52:29,519 --> 00:52:34,199 mentoring and growing the next data science leaders. I've noticed that there's a lot 726 00:52:34,239 --> 00:52:38,199 of they teach a lot about the data science and then they find people who 727 00:52:38,199 --> 00:52:42,159 are like, oh, well, you look like you could lead, We're 728 00:52:42,159 --> 00:52:44,920 going to throw you into leadership, but they don't do a lot of teaching 729 00:52:45,079 --> 00:52:52,679 about data science for business value or data science you know, in the bigger 730 00:52:52,719 --> 00:52:57,760 picture, and how to lead a team and that sort of things. I've 731 00:52:57,760 --> 00:53:00,960 been doing a lot of mentoring some of the younger data scientists, which I've 732 00:53:00,960 --> 00:53:04,360 been really enjoying. Wow, that's great. Well yeah, yeah, you're 733 00:53:04,400 --> 00:53:07,400 doing You're doing the people's work for sure, because we goodness knows, we 734 00:53:07,519 --> 00:53:12,000 need more knowledgele people in this space. Lots of people using data, not 735 00:53:12,159 --> 00:53:15,840 enough data scientists. Yep, that's true. Yeah, well, Amber, 736 00:53:15,000 --> 00:53:19,480 thank you so much for coming on the show and talking about this. It's 737 00:53:19,559 --> 00:53:22,719 not a topic that's going away soon, and I'm sure it's going to change 738 00:53:22,719 --> 00:53:27,960 by the next time we have you on. Absolutely. Thanks for having me. 739 00:53:28,320 --> 00:53:30,000 Great to have you about Yeah you bet, and we'll talk to you, 740 00:53:30,159 --> 00:53:55,519 dear listener next time on dot net rocks. Dot net Rocks is brought 741 00:53:55,519 --> 00:54:00,679 to you by Franklin's Net and produced by Poop Studios, a full service audio, 742 00:54:00,800 --> 00:54:06,519 video and post production facility located physically in New London, Connecticut, and 743 00:54:06,639 --> 00:54:12,480 of course in the cloud online at pwop dot com. Visit our website at 744 00:54:12,559 --> 00:54:17,719 dt n et r ocks dot com for RSS feeds, downloads, mobile apps, 745 00:54:17,880 --> 00:54:22,000 comments, and access to the full archives going back to show number one, 746 00:54:22,360 --> 00:54:27,079 recorded in September two thousand and two. And make sure you check out 747 00:54:27,079 --> 00:54:30,400 our sponsors. They keep us in business. Now go write some code CE 748 00:54:30,599 --> 00:54:32,199 next time you got