WEBVTT - Tech News: AI Versus The Pope 0:00:04.440 --> 0:00:12.280 Welcome to tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.280 --> 0:00:15.920 and welcome to tech Stuff. I'm your host, Jonathan Strickland. 0:00:15.960 --> 0:00:18.960 I'm an executive producer with iHeartRadio. And how the tech 0:00:19.000 --> 0:00:22.160 are you. It's time for the tech news for Tuesday, 0:00:22.320 --> 0:00:28.320 August eighth, twenty twenty three. And despite my voice catching there, 0:00:28.400 --> 0:00:31.200 I feel a lot better than I did yesterday. Thanks 0:00:31.200 --> 0:00:31.800 for asking. 0:00:32.159 --> 0:00:33.640 Let's get to the tech news. 0:00:34.159 --> 0:00:37.120 And we've got a lot of AI related stories today. 0:00:37.320 --> 0:00:40.879 No big surprise there. It has been the topic of 0:00:41.000 --> 0:00:46.479 twenty twenty three, at least whenever Elon Musk isn't demanding 0:00:46.520 --> 0:00:50.680 all the headlines. And it's also sadly no surprise that 0:00:50.720 --> 0:00:53.199 one of the AI stories we're covering today has to 0:00:53.240 --> 0:00:57.360 do with faulty facial recognition technology and the mistake of 0:00:57.440 --> 0:01:01.040 relying on that tech for the purposes of law enforcement. 0:01:01.760 --> 0:01:06.200 Porsche Woodruff, a black woman in Detroit, found herself arrested 0:01:06.280 --> 0:01:10.560 and attained for eleven hours when police acted on an 0:01:10.680 --> 0:01:14.680 incorrect facial recognition match while seeking a suspect in a 0:01:14.720 --> 0:01:19.600 carjacking and robbery crime. Not only that Portiae is eight 0:01:19.680 --> 0:01:23.200 months pregnant, and the surveillance footage from the crime in 0:01:23.280 --> 0:01:27.319 question showed a woman who very much was not eight 0:01:27.360 --> 0:01:30.160 months pregnant. So not only was this a case of 0:01:30.200 --> 0:01:34.200 facial recognition software giving a false positive, it's also a 0:01:34.240 --> 0:01:37.840 case of cops apparently working under the assumption that a 0:01:37.880 --> 0:01:41.839 woman can go from not visibly pregnant to eight months 0:01:41.840 --> 0:01:45.600 pregnant in a very short amount of time, which is wild. 0:01:45.840 --> 0:01:50.040 So here's kind of what unfolded. A man reported being 0:01:50.080 --> 0:01:53.920 the victim of a robbery and carjacking. The police were 0:01:53.920 --> 0:01:57.080 able to secure surveillance footage, and they used a tool 0:01:57.200 --> 0:02:02.720 called data works plus to unmatches against mug shots that 0:02:02.800 --> 0:02:06.680 were stored in a police database. Woodroff had been arrested 0:02:06.680 --> 0:02:09.840 back in twenty fifteen, so her mugshot was one of 0:02:09.880 --> 0:02:13.440 the images in that database. This tool pulled a match 0:02:13.600 --> 0:02:16.959 between her mugshot and the surveillance footage, and the guy 0:02:16.960 --> 0:02:20.880 who was robbed also mistook a photo of Woodroff her 0:02:20.919 --> 0:02:24.320 mugshot as that as the same person as the perpetrator. 0:02:24.720 --> 0:02:27.840 So the police go and they arrest Woodroff, and of 0:02:27.840 --> 0:02:30.400 course she was not involved in the crime, could not 0:02:30.560 --> 0:02:33.080 have been, and just the fact that she was, you know, 0:02:33.160 --> 0:02:36.919 eight months pregnant should have been the immediate giveaway that 0:02:36.960 --> 0:02:40.000 this is not the same person. The New York Times 0:02:40.040 --> 0:02:44.040 subsequently reported that Woodroff's case was the third in the 0:02:44.080 --> 0:02:48.800 city of Detroit alone that resulted in a wrongful arrest 0:02:49.040 --> 0:02:53.360 due to incorrect facial recognition matches, and that all three 0:02:53.400 --> 0:02:56.400 of those cases, as well as three other cases that 0:02:56.480 --> 0:03:01.359 were not in Detroit, all involved black people. I think 0:03:01.400 --> 0:03:05.040 it's safe to say that even the faulty facial recognition 0:03:05.080 --> 0:03:08.520 technology is able to see a pattern emerging here, and 0:03:08.560 --> 0:03:13.320 it's one of racial bias in surveillance and identification tools. Now, 0:03:13.360 --> 0:03:16.720 as we have covered on this podcast, several cities and 0:03:16.840 --> 0:03:20.760 jurisdictions have banned the use of facial recognition for law 0:03:20.840 --> 0:03:25.239 enforcement purposes. Personally, I think that is merited. I can't 0:03:25.280 --> 0:03:29.440 help but imagine what being wrongfully arrested must be like. 0:03:29.560 --> 0:03:33.440 It's got to be incredibly traumatic and disruptive and. 0:03:33.360 --> 0:03:36.520 Potentially cause lots of issues. 0:03:36.040 --> 0:03:39.480 In your life, and you at no point were at 0:03:39.520 --> 0:03:42.720 fault for any of it. And if a technology is 0:03:42.800 --> 0:03:49.000 disproportionately leading to that kind of thing, we should not 0:03:49.320 --> 0:03:54.880 be using that technology for those purposes. I don't know 0:03:54.920 --> 0:03:58.160 how there's any argument against that. If the tool is 0:03:58.480 --> 0:04:01.640 leading to innocent people getting arrested and their lives getting 0:04:01.680 --> 0:04:02.840 up ended in the process. 0:04:03.320 --> 0:04:04.800 You got to stop using the tool. 0:04:05.560 --> 0:04:09.400 British researchers showed how a deep learning algorithm, once trained, 0:04:09.400 --> 0:04:13.640 would be able to decode keystrokes just from the sound 0:04:14.040 --> 0:04:18.400 of typing. This isn't totally new. I've heard of these 0:04:18.480 --> 0:04:22.040 kinds of attacks before. But imagine for a moment that 0:04:22.520 --> 0:04:25.880 you are set up in some public space and someone 0:04:25.880 --> 0:04:29.160 else happens to have their phone out and you know, 0:04:29.240 --> 0:04:31.800 you don't know this, but what they're actually doing is 0:04:31.839 --> 0:04:35.240 activating the phone's microphone, and they're picking up on the 0:04:35.279 --> 0:04:39.359 sound of you tippity tappany typing away, and you're oblivious 0:04:39.360 --> 0:04:41.400 to any threats. So like maybe you're you're like being 0:04:41.400 --> 0:04:43.479 careful with your screen or whatever, but you're not thinking 0:04:43.480 --> 0:04:48.039 about the actual keystrokes. And meanwhile, a computer program on 0:04:48.080 --> 0:04:51.240 the other end of that microphone is effectively transcribing everything 0:04:51.240 --> 0:04:55.880 you've typed, potentially including your login information. This is an 0:04:55.920 --> 0:04:59.919 acoustic attack, and I bet that makes all those mechanical 0:05:00.120 --> 0:05:03.600 keyboard clackie clackie types out there a little nervous. So 0:05:04.040 --> 0:05:06.400 is it possible that some hacker out there could get 0:05:06.440 --> 0:05:09.920 your log in credentials just by having a computer listen 0:05:09.960 --> 0:05:14.040 to you type. Technically, yes, it is possible. It is 0:05:14.080 --> 0:05:18.120 not necessarily straightforward or easy to do. It is possible. 0:05:18.760 --> 0:05:23.320 So if you're at a location, then you need to 0:05:23.360 --> 0:05:26.240 know that is a possibility. But like a lot of 0:05:26.240 --> 0:05:28.479 locations have a lot of other noise, it's hard to 0:05:28.480 --> 0:05:31.400 set up a microphone in such a way that you're 0:05:31.440 --> 0:05:34.599 going to get a very clear recording of that sound. 0:05:35.160 --> 0:05:36.960 Maybe if you're at a coffee shop, you should just 0:05:36.960 --> 0:05:39.480 start making CLACKI clacky noises with your mouth at the 0:05:39.520 --> 0:05:41.960 same time as you type, you know, to throw off 0:05:42.000 --> 0:05:45.320 any potential batties who are trying to listen in on you. 0:05:45.800 --> 0:05:46.280 I don't know. 0:05:46.720 --> 0:05:49.919 It's a wild world out there. Reuter's reports that the 0:05:49.920 --> 0:05:53.560 Walt Disney Company has created a task force to research 0:05:53.600 --> 0:05:57.560 how artificial intelligence could be used within that company, which 0:05:57.600 --> 0:06:01.159 of course encompasses lots of different divisions. You've got the 0:06:01.279 --> 0:06:05.080 entertainment division, You've got theme parks, You've got merchandising, you've 0:06:05.120 --> 0:06:10.080 got advertising. There's tons of businesses underneath the umbrella of 0:06:10.120 --> 0:06:14.719 Walt Disney. Well, the company currently has nearly a dozen 0:06:14.839 --> 0:06:19.200 job openings that mention artificial intelligence research and development. So 0:06:20.040 --> 0:06:21.960 this does look like it's a big push and could 0:06:21.960 --> 0:06:25.960 include things like imagineering, but it also ranges to other stuff, 0:06:26.000 --> 0:06:30.240 you know, theme parks to advertising to also Disney television. 0:06:30.440 --> 0:06:33.400 And that last one is pretty notable because many of 0:06:33.400 --> 0:06:36.320 the contentious elements that are at the center of the 0:06:36.600 --> 0:06:41.120 ongoing strikes in Hollywood, which if you weren't aware, involves 0:06:41.160 --> 0:06:45.280 both writers and actors. They're both on strike in Hollywood 0:06:45.320 --> 0:06:47.880 right now. Well, one of the things they're striking about 0:06:48.560 --> 0:06:51.760 is all about how studios should or should not use 0:06:51.880 --> 0:06:57.880 artificial intelligence moving forward. Reuter's cites some unnamed folks connected 0:06:57.920 --> 0:07:00.560 to Disney, no people who didn't want to have this 0:07:00.640 --> 0:07:03.400 get back to them, but they said that the company 0:07:03.440 --> 0:07:06.400 really has little choice here, that if it doesn't incorporate 0:07:06.480 --> 0:07:10.760 AI into its strategy, it runs the risk of becoming obsolete. 0:07:11.840 --> 0:07:13.840 Maybe that's true. 0:07:13.960 --> 0:07:17.400 But my knee jerk reaction kind of is triggered because 0:07:17.440 --> 0:07:20.640 back in two thousand and four, Disney famously shut down 0:07:20.680 --> 0:07:25.520 its two D animation studios, and it was unthinkable for 0:07:25.600 --> 0:07:30.400 a company that had built its reputation on traditional two 0:07:30.480 --> 0:07:34.000 D animation to suddenly turn its back on it. It 0:07:34.040 --> 0:07:37.520 has subsequently changed that, But for a while, it looked 0:07:37.560 --> 0:07:39.880 like two D animation and Disney were just things of 0:07:39.920 --> 0:07:42.880 the past, and it just seemed like there was this 0:07:43.000 --> 0:07:47.120 attitude among Disney executive leadership that computer animation was somehow 0:07:47.480 --> 0:07:50.840 not just different from traditional hand drawn two D animation, 0:07:51.320 --> 0:07:55.600 but innately better than two D animated films. Like, audiences 0:07:55.640 --> 0:07:58.680 didn't want to see two D animation, they just wanted 0:07:58.720 --> 0:08:01.720 computer animation, and for proof of that you would look 0:08:01.760 --> 0:08:06.720 at Pixar. Here's the problem. Pixar was investing heavily in 0:08:06.880 --> 0:08:11.720 developing great stories to tell, and yes, the computer animation 0:08:12.240 --> 0:08:15.040 was getting more and more impressive with every single film, 0:08:15.800 --> 0:08:19.960 but they were really putting story first, whereas the animated 0:08:20.040 --> 0:08:23.480 side over at Disney had fallen a long way since 0:08:23.520 --> 0:08:26.560 the early days of the so called Disney Renaissance, which 0:08:26.600 --> 0:08:29.560 included movies like The Little Mermaid and Beauty and the 0:08:29.560 --> 0:08:32.640 Beast in Aladdin. You know, if you look at nineteen 0:08:32.679 --> 0:08:35.319 ninety five, for example, I would say I don't think 0:08:35.360 --> 0:08:38.160 Toy Story, which came out in nineteen ninety five, was 0:08:38.640 --> 0:08:41.920 a better movie than Pocahontas, which also came out in 0:08:42.040 --> 0:08:45.679 ninety five, just because Toy Story was computer animated and 0:08:45.720 --> 0:08:48.680 Pocahontas was hand drawn. I think Toy Story was a 0:08:48.679 --> 0:08:51.839 better movie because the script was better. But you know, 0:08:52.040 --> 0:08:54.280 Pocahontas also had to follow up on the amazing work 0:08:54.280 --> 0:08:56.800 of Ashman and Mancoln and Ashman had passed away in 0:08:56.880 --> 0:08:59.680 nineteen ninety one, so there are a lot of other 0:09:00.080 --> 0:09:03.920 to getting factors there. Anyway, Disney hasn't really commented on 0:09:04.240 --> 0:09:08.000 how it plans to incorporate AI into its processes. It's 0:09:08.440 --> 0:09:11.560 planning on doing it, but it hasn't talked about what 0:09:11.720 --> 0:09:14.600 that might look like. So you know, you could have 0:09:14.679 --> 0:09:18.720 AI incorporated into very mundane stuff, right, like automating things 0:09:18.720 --> 0:09:22.640 like scheduling and finding the most efficient means to do that, 0:09:23.200 --> 0:09:28.199 which isn't necessarily impacting the creative side of the business 0:09:28.200 --> 0:09:31.439 that much. Right, if you're using it to handle stuff 0:09:31.960 --> 0:09:35.520 that is otherwise tedious and takes up a lot of 0:09:35.559 --> 0:09:40.359 time but is easy enough to automate, that's not necessarily 0:09:40.400 --> 0:09:42.840 a bad thing. That can end up making a company 0:09:42.840 --> 0:09:47.439 more efficient without also displacing employees in the process, or 0:09:47.440 --> 0:09:50.839 at least freeing those employees up to do more rewarding 0:09:50.960 --> 0:09:54.760 work instead of something that's really tedious. The concern is 0:09:54.800 --> 0:09:57.959 whether or not that use of AI could end up 0:09:58.920 --> 0:10:04.120 replacing very important and creative roles for people, whether it's 0:10:04.200 --> 0:10:09.600 actors or writers or imagineers or whomever. So yeah, big 0:10:09.640 --> 0:10:15.480 concern in AI with Disney. Yesterday, Zoom made a change 0:10:15.520 --> 0:10:18.280 to its terms of service after receiving some pretty harsh 0:10:18.280 --> 0:10:23.040 criticism from customers over the weekend. So around Sunday, people 0:10:23.080 --> 0:10:26.160 who were paying attention to zoom terms of service started 0:10:26.200 --> 0:10:29.560 to post screenshots of those terms, and they included a 0:10:29.600 --> 0:10:32.600 passage saying the company has the right to collect, store, 0:10:32.920 --> 0:10:37.400 and use quote unquote service generated data. Now that alone 0:10:38.120 --> 0:10:41.360 seems a bit concerning. Right. Let's say your company uses 0:10:41.440 --> 0:10:44.800 Zoom for business meetings. You probably don't like the idea 0:10:45.280 --> 0:10:49.480 of Zoom potentially snooping in on your calls. On top 0:10:49.520 --> 0:10:51.400 of that, the term said that Zoom could use that 0:10:51.520 --> 0:10:55.839 data essentially to train AI, and that really got people upset. 0:10:56.200 --> 0:10:58.640 The thought that their video calls could be used as 0:10:58.720 --> 0:11:03.559 material to train another aimodel seemed invasive. Now, a Zoom 0:11:03.600 --> 0:11:06.559 rep explained that the collection features are part of an 0:11:06.559 --> 0:11:10.960 opt in system users can choose to enable generative AI 0:11:11.040 --> 0:11:14.840 features such as transcription services, and the company does not 0:11:15.120 --> 0:11:19.959 use any customer content without first gaining the consent from 0:11:20.000 --> 0:11:23.559 the customer. To that end, now, Zoom has updated its 0:11:23.640 --> 0:11:26.880 terms of use to clearly say it will only collect 0:11:26.920 --> 0:11:31.319 audio and video data with consent from the user first. 0:11:31.920 --> 0:11:35.560 In related news, Zoom leadership has also called on Zoom 0:11:35.559 --> 0:11:39.240 employees who live within fifty miles of a Zoom office 0:11:39.960 --> 0:11:43.760 to actually attend work in person at least two days 0:11:43.800 --> 0:11:44.120 a week. 0:11:44.320 --> 0:11:44.920 So that's right. 0:11:45.000 --> 0:11:48.600 The company that made the tool talented as one of 0:11:48.600 --> 0:11:53.600 the most important during lockdown, the one that enabled remote work, 0:11:54.520 --> 0:11:59.720 is now restricting remote work at the corporate level at Zoom, 0:12:00.280 --> 0:12:04.240 which is you know how the tables have turned. I 0:12:04.280 --> 0:12:07.800 guess okay, we're going to take a quick break. When 0:12:07.800 --> 0:12:19.400 we come back, we've got some more tech news. We're back, 0:12:19.480 --> 0:12:23.720 and actually have a couple of other AI stories to 0:12:23.880 --> 0:12:26.880 finish up with before we move on to other tech news, 0:12:27.320 --> 0:12:29.680 so we're not done with AI just yet. AP News 0:12:29.720 --> 0:12:32.520 reports that the folks over at Dungeons and Dragons, you know, 0:12:32.559 --> 0:12:34.840 Wizards of the Coast, which in turn is part of Hasbro, 0:12:35.520 --> 0:12:40.000 are now telling their artists not to use artificial intelligence 0:12:40.360 --> 0:12:43.920 as part of the generative process to create fantasy arts, 0:12:43.920 --> 0:12:47.439 so they're telling the artists, hey, don't use AI when 0:12:47.480 --> 0:12:50.440 you're making art for us. This comes after several D 0:12:50.480 --> 0:12:53.480 and D fans raised questions about an illustration that included 0:12:53.520 --> 0:12:57.520 a giant that they said looked a little weird, like 0:12:57.840 --> 0:13:00.400 perhaps it had not been made by a human being, 0:13:00.960 --> 0:13:03.360 and they asked, hey, was this made by a robot? 0:13:03.600 --> 0:13:04.959 So D and D reached. 0:13:04.679 --> 0:13:08.240 Out and contacted the artist, talked with them, found out 0:13:08.240 --> 0:13:12.280 that yes, there was some use of AI generative features 0:13:12.320 --> 0:13:16.880 to collaborate and make this art, and the company now 0:13:16.960 --> 0:13:19.800 is clarifying rules on what can and cannot be used 0:13:19.800 --> 0:13:23.920 to make fantasy art for the games. The particular piece 0:13:24.160 --> 0:13:28.240 that triggered you folks to ask, is this AI generated 0:13:28.600 --> 0:13:32.079 is actually going to be appearing soon in an expansion 0:13:32.080 --> 0:13:35.560 book called Big Bie Presents Glory of the Giants. I 0:13:35.559 --> 0:13:39.520 think it comes out next week in fact, so maybe 0:13:39.520 --> 0:13:41.640 collectors will rush out and grab a copy in case 0:13:41.679 --> 0:13:45.680 future editions will remove the AI generated giant person. But 0:13:45.760 --> 0:13:48.680 you might ask why is D and D, Why is 0:13:48.720 --> 0:13:52.640 Wizard of the Coast, And perhaps by extension, Hasbro saying 0:13:53.240 --> 0:13:56.080 don't use AI to generate fantasy art. Why is that 0:13:56.120 --> 0:13:59.640 a big deal? Well, part of it is about copyright 0:13:59.720 --> 0:14:03.880 quest like who owns the copyright to a machine generated 0:14:03.920 --> 0:14:07.640 piece of work. Obviously, Wizards of the Coast want to 0:14:07.640 --> 0:14:10.640 be able to copyright their their stuff and not have 0:14:10.760 --> 0:14:14.720 some other party claim ownership of something that's featured in 0:14:14.960 --> 0:14:19.080 a Wizards of the Coast work. But also there's the 0:14:19.240 --> 0:14:25.320 issue of you know, copying another person's style. Right. If 0:14:25.440 --> 0:14:29.600 artist A produces a ton of fantasy art and then 0:14:29.720 --> 0:14:34.800 artist B uses a generative tool that coincidentally is referencing 0:14:35.120 --> 0:14:39.840 artists a work and it's doing so extensively, and then 0:14:39.880 --> 0:14:43.000 creates a new piece in the style of artist A, 0:14:43.200 --> 0:14:46.040 well then it's almost like artist B is copying Artist A. 0:14:46.720 --> 0:14:49.400 And you could argue, well, then that means artist A 0:14:49.480 --> 0:14:52.320 could have landed this gig and gotten a paying gig 0:14:52.320 --> 0:14:56.680 out of it, but instead their work was then you know, 0:14:56.720 --> 0:15:00.920 sort of repurposed and reimagined without their consent. This is 0:15:00.920 --> 0:15:05.560 an ongoing issue with generative AI in the visual arts realm. 0:15:05.800 --> 0:15:09.120 There's also a very similar effect, the same problem that's 0:15:09.160 --> 0:15:13.800 going on within the written word. Right. There are authors 0:15:13.840 --> 0:15:17.640 and poets who are arguing that AI models being trained 0:15:17.720 --> 0:15:24.600 on published works are effectively copying these folks without their consent. So, 0:15:25.400 --> 0:15:29.200 if you'll excuse me, I need to roll to see 0:15:29.200 --> 0:15:33.040 if my AI generated image of an elf will deceive 0:15:33.280 --> 0:15:41.640 Wizards of the Coast and oh, critical fail. Okay, well, 0:15:42.480 --> 0:15:45.640 I guess the fingers are all nudely and there's like 0:15:45.720 --> 0:15:48.440 fifteen of them, so I guess that's a dead giveaway. 0:15:49.120 --> 0:15:49.480 Okay. 0:15:49.960 --> 0:15:52.560 You know a lot of folks have voiced concerns about 0:15:52.600 --> 0:15:56.680 AI and the possible dangers that it could bring, and 0:15:56.760 --> 0:16:00.480 we can now count the Pope as one of those voices. 0:16:00.560 --> 0:16:05.200 Pope Francis has called on a global reflection today that 0:16:05.400 --> 0:16:09.680 is all about how AI could be really dangerous. Pope 0:16:09.680 --> 0:16:12.720 Francis has said in the past he is largely unfamiliar 0:16:12.800 --> 0:16:16.240 with modern technology, including stuff like computers and the Internet, 0:16:16.680 --> 0:16:19.160 but that he also sees that these tools can be 0:16:19.320 --> 0:16:23.400 incredibly helpful when they are put to appropriate use, which 0:16:23.680 --> 0:16:26.560 you know, is a refreshing take from someone who is 0:16:26.680 --> 0:16:31.760 unfamiliar with technology that they recognize tools aren't necessarily good 0:16:31.880 --> 0:16:35.400 or bad in of themselves. It's really all in how 0:16:35.440 --> 0:16:37.840 we make use of those tools, and if we commit 0:16:38.040 --> 0:16:41.400 to using them. In ways that aren't harmful, we can 0:16:41.440 --> 0:16:44.960 see great benefit. But some of these tools like AI, 0:16:45.160 --> 0:16:48.480 for example, have the potential to be very dangerous if 0:16:48.520 --> 0:16:51.960 we are using them improperly or if we don't have 0:16:51.960 --> 0:16:55.040 a full understanding of the consequences before we use them, 0:16:55.280 --> 0:16:58.600 so we have to take extra care when we're working 0:16:58.680 --> 0:17:02.440 with them. It's not that AI is not worthwhile or 0:17:02.640 --> 0:17:07.800 could never do anything positive. That's clearly not the case. 0:17:08.440 --> 0:17:12.000 We just have to be very very methodical in our 0:17:12.080 --> 0:17:14.919 approach to using AI, and right now you could argue 0:17:14.960 --> 0:17:19.199 that is not what we're seeing. Apple has reportedly struck 0:17:19.240 --> 0:17:23.880 a huge deal with chip manufacturer TSMC out of Taiwan 0:17:24.480 --> 0:17:29.040 that will see Apple purchase essentially all of TSMC's chips 0:17:29.080 --> 0:17:33.240 made with their next generation manufacturing process, which is called 0:17:33.240 --> 0:17:38.000 the three nanometer manufacturing process. Just as a reminder, once 0:17:38.080 --> 0:17:41.680 upon a time, when we would use things like nanometer 0:17:41.840 --> 0:17:46.280 to describe the chip manufacturing process that actually referenced the 0:17:46.320 --> 0:17:49.399 size of individual components found on the chips, that you 0:17:49.400 --> 0:17:53.119 would actually see stuff on the chips that measured at 0:17:52.800 --> 0:17:57.240 that unit. But these days it's really just a naming convention. 0:17:57.440 --> 0:18:00.920 It's really just to designate the generation of the chip 0:18:00.960 --> 0:18:04.399 manufacturing process, the individual elements on the chips are not 0:18:04.800 --> 0:18:08.000 three nanometers in size. That would end up being a 0:18:08.040 --> 0:18:12.440 big disaster because of the way quantum physics works. So yeah, 0:18:12.520 --> 0:18:15.879 just a reminder that the whole nanometer thing, it's just 0:18:15.920 --> 0:18:19.160 a naming convention now. It doesn't actually reference anything other 0:18:19.280 --> 0:18:21.560 than this is the newest one and it has to 0:18:21.640 --> 0:18:25.879 keep getting smaller. So it does raise questions of do 0:18:25.960 --> 0:18:29.720 we go down to the atomic scale once we get 0:18:29.760 --> 0:18:35.280 past one nanometer. Anyway, according to the information, Apple has 0:18:35.359 --> 0:18:40.920 essentially ordered every single TSMC three nanometer chip at least 0:18:40.920 --> 0:18:43.600 in the short term, and by short term I mean 0:18:43.760 --> 0:18:48.119 Apple will have exclusive use of chips made by that 0:18:48.160 --> 0:18:52.280 manufacturing process from TSMC for about a year, and that 0:18:52.359 --> 0:18:55.959 definitely gives Apple a leg up on the competition that 0:18:56.080 --> 0:18:59.800 wants to use TSMC's chips. There are other fabricators out there. 0:19:00.119 --> 0:19:02.720 TSMC is not the only game in town. It's just 0:19:02.800 --> 0:19:07.080 the biggest one. Meanwhile, there is a political battle surrounding 0:19:07.119 --> 0:19:11.800 TSMC's planned fabrication facility that would be here in the 0:19:11.880 --> 0:19:16.320 United States, in Arizona. So the Taiwan based company plans 0:19:16.400 --> 0:19:20.000 this fabrication plant in Arizona, but just last month announced 0:19:20.040 --> 0:19:22.800 that there was going to be a construction delay that 0:19:22.920 --> 0:19:27.000 would last until twenty twenty five. The reason being, according 0:19:27.040 --> 0:19:29.639 to the company, is that there are a lack of 0:19:29.800 --> 0:19:32.840 skilled workers here in the US who would be needed 0:19:32.960 --> 0:19:36.760 to prepare and open the facility, not to work once 0:19:36.800 --> 0:19:39.879 it is open, but to actually get everything in place 0:19:39.960 --> 0:19:44.400 as well. They're just the talent isn't here, and instead 0:19:44.480 --> 0:19:48.200 TSMC wants to bring around five hundred employees from Taiwan 0:19:48.800 --> 0:19:52.000 to the United States to do that work instead. That 0:19:52.080 --> 0:19:55.159 has led to US politicians weighing in and they have 0:19:55.359 --> 0:19:59.359 argued that these jobs should go to US workers. Now 0:20:00.359 --> 0:20:03.240 there's a lot going on here, and it gets very 0:20:03.280 --> 0:20:07.119 messy and it gets very political. But from a high level, 0:20:07.520 --> 0:20:12.639 the US decades ago seeded as in got rid of 0:20:12.800 --> 0:20:18.199 pretty much all major chip fabrication because it's expensive. It 0:20:18.280 --> 0:20:21.679 is very expensive to build chip fabrication plants. You have 0:20:21.720 --> 0:20:24.919 to update them constantly because as we were just talking about, 0:20:25.000 --> 0:20:30.320 you're always evolving the technology to make more powerful chips, 0:20:30.880 --> 0:20:33.879 which means you got to retool everything. Sometimes you have 0:20:33.920 --> 0:20:38.240 to build totally new facilities and that's a huge investment 0:20:38.280 --> 0:20:40.439 and a lot of US companies got out of that 0:20:40.520 --> 0:20:44.760 game ages ago, and instead that work went to places 0:20:44.840 --> 0:20:51.800 like Taiwan and TSMC in particular. So because America got 0:20:51.920 --> 0:20:56.239 rid of didn't totally get rid of chip fabrication, but 0:20:56.400 --> 0:20:59.200 largely pushed that out to other places in the world, 0:21:00.280 --> 0:21:03.159 you might say that TSMC could at least have a 0:21:03.200 --> 0:21:06.679 partly legit point to make that the US lacks the 0:21:06.720 --> 0:21:12.040 experts needed to open an advanced fabrication facility simply because 0:21:12.520 --> 0:21:15.320 the US hasn't really been focused on that part of 0:21:15.440 --> 0:21:19.159 chip manufacturing for a while. Yes, in the US you 0:21:19.200 --> 0:21:22.760 have a lot of people developing the next chips, designing 0:21:22.800 --> 0:21:27.159 the next generation of chips, but the actual fabrication is 0:21:27.200 --> 0:21:31.240 taking place elsewhere. So while the expertise is definitely in design, 0:21:32.200 --> 0:21:36.320 the argument is it's not in creating the fabrication facilities, 0:21:36.480 --> 0:21:40.080 and that's where the problem is. On the flip side, 0:21:41.040 --> 0:21:44.520 there's a concern that the reason TSMC really wants to 0:21:44.520 --> 0:21:47.879 bring in Taiwanese workers to the US is not just 0:21:47.960 --> 0:21:51.280 because they have expertise in the area, but also because 0:21:51.280 --> 0:21:55.040 they're less likely to resist a push to work really 0:21:55.119 --> 0:21:59.879 long hours, including working weekends and stuff, whereas US workers 0:22:00.200 --> 0:22:04.240 have this pesky habit of arguing that they need to 0:22:04.280 --> 0:22:07.840 be you know, fairly compensated and have work life balance. 0:22:08.320 --> 0:22:11.239 So in other words, there's a concern that TSMC is 0:22:11.280 --> 0:22:14.560 really looking to exploit a workforce in an effort to 0:22:14.640 --> 0:22:18.520 keep costs down and to speed up building out the facilities. 0:22:19.400 --> 0:22:22.760 As for what happens once the facility opens, TSMC says 0:22:22.800 --> 0:22:26.359 it's committed to providing around twelve thousand jobs and that 0:22:26.520 --> 0:22:29.879 US employees will fill those roles, so that like the 0:22:29.920 --> 0:22:34.119 actual jobs of working at the facility will go to 0:22:34.280 --> 0:22:38.280 US citizens, not you know, it won't be Taiwanese workers 0:22:38.280 --> 0:22:42.119 brought over to do that particular work. So yeah, like 0:22:42.160 --> 0:22:45.240 I said, it is political. There is a technical side 0:22:45.240 --> 0:22:49.480 to it too, but it's messy. And this is why 0:22:49.560 --> 0:22:53.080 you can't just leave politics out of discussions of technology, 0:22:53.440 --> 0:22:57.199 because politics affects us and it affects the tech sector 0:22:57.640 --> 0:23:01.080 a lot. In fact, you know, we can often see 0:23:01.119 --> 0:23:04.120 it most acutely in the tech sector. Not that it's 0:23:04.160 --> 0:23:07.359 not impacting other sectors as well, it's just when it 0:23:07.440 --> 0:23:12.040 hits tech, people take notice because it's high profile stuff. 0:23:12.200 --> 0:23:15.080 So we can't really avoid it here. I don't know 0:23:15.800 --> 0:23:18.080 what the actual story is back here. I mean, it 0:23:18.119 --> 0:23:21.720 may very well be that TSMC could not find the 0:23:21.800 --> 0:23:25.359 talent they needed in order to prepare the fabrication facility 0:23:25.359 --> 0:23:29.040 properly here in the US. Maybe that's true, or maybe 0:23:29.040 --> 0:23:32.800 it's not. I just don't know, but I do know 0:23:32.880 --> 0:23:36.000 that it is an ongoing issue right now. So we'll 0:23:36.040 --> 0:23:40.320 have to see how that plays out in the short term. Okay, 0:23:40.880 --> 0:23:44.080 I got a few more stories to cover, but before 0:23:44.119 --> 0:23:47.080 I can get to that, let's take another quick break. 0:23:56.560 --> 0:23:57.240 We're back. 0:23:57.960 --> 0:24:00.960 Ours Technica has an article about how scientists at the 0:24:01.040 --> 0:24:04.920 Lawrence Livermore National Lab in California have, for the second 0:24:05.000 --> 0:24:10.080 time now, produced a fusion reaction that generated more energy 0:24:10.680 --> 0:24:14.920 than was needed to initiate the reaction. We're gonna put 0:24:14.920 --> 0:24:17.320 an asterisk on that because we're gonna come back to 0:24:17.359 --> 0:24:20.240 it now. I have talked about fusion a lot on 0:24:20.280 --> 0:24:23.479 the show, but just as a quick reminder, fusion occurs 0:24:23.480 --> 0:24:28.000 when you take two lightweight atoms, like hydrogen atoms, for example, 0:24:29.280 --> 0:24:32.000 and then you blast those atoms with enough pressure and 0:24:32.400 --> 0:24:36.040 or energy like heat, in order to. 0:24:35.960 --> 0:24:38.159 Fuse them into a new. 0:24:38.040 --> 0:24:41.919 Atom, helium. In this case, this is what happens in 0:24:42.000 --> 0:24:46.280 the sun. You know, hydrogen is forged into helium at 0:24:46.320 --> 0:24:50.920 a temperature of millions of degrees as how does or 0:24:50.960 --> 0:24:54.480 why does the sunshine would tell us anyway, in this process, 0:24:55.040 --> 0:24:57.480 you also end up with a release of energy, right, 0:24:57.480 --> 0:25:00.320 That's why the sun actually does shine. It's releasing it energy, 0:25:00.359 --> 0:25:03.679 it's not just doing this process. So if you end 0:25:03.760 --> 0:25:07.880 up with more energy than you used to start the reaction, 0:25:08.000 --> 0:25:11.359 you've got a potentially viable alternative to other kinds of 0:25:11.359 --> 0:25:14.200 power facilities, you know, like coal power plants, or even 0:25:14.280 --> 0:25:17.800 things like solar and wind farms, or traditional nuclear power 0:25:17.800 --> 0:25:22.159 plants which rely on nuclear fission. That's the process of 0:25:22.240 --> 0:25:26.840 splitting heavy atoms into lighter atoms, and that also releases 0:25:26.840 --> 0:25:28.720 a huge amount of energy in the process, but that 0:25:29.720 --> 0:25:31.960 also creates things like nuclear waste, which you have to 0:25:31.960 --> 0:25:34.639