WEBVTT - Tech News: The US Infrastructure Bill Introduces Some Tech Problems 0:00:04.400 --> 0:00:07.800 Welcome to tech Stuff, a production from I Heart Radio. 0:00:12.160 --> 0:00:15.160 Hey there, and welcome to text Stuff. I'm your host, 0:00:15.280 --> 0:00:18.280 Jonathan Strickland. I'm an executive producer with I Heart Radio 0:00:18.560 --> 0:00:21.680 and I love all things tech. And this is the 0:00:21.720 --> 0:00:28.240 tech news for Tuesday, August twenty one. As I record 0:00:28.320 --> 0:00:30.960 this here in the United States, the Senate has just 0:00:31.120 --> 0:00:36.360 passed a one point to trillion dollar infrastructure bill. Now, 0:00:36.440 --> 0:00:39.559 for a bill to pass in the United States, it 0:00:39.640 --> 0:00:43.640 has to pass by vote both the Senate and the 0:00:43.760 --> 0:00:48.440 House of Representatives in Congress before then being signed into 0:00:48.560 --> 0:00:51.880 law by the President. Now, the House is not set 0:00:51.920 --> 0:00:54.480 to vote on this until they come back from recess. 0:00:55.120 --> 0:00:57.440 That always makes me imagine a bunch of Congress people 0:00:57.440 --> 0:01:01.520 out on the playground. Anyway, my point is that, you know, 0:01:01.640 --> 0:01:05.320 with the Senate passing the bill, the House still has 0:01:05.440 --> 0:01:08.360 its part to play, so changes can still happen to 0:01:08.520 --> 0:01:12.600 the language in this bill, So nothing is set in 0:01:12.720 --> 0:01:16.080 stone until it actually passes and then the President has 0:01:16.120 --> 0:01:19.360 to sign it into law. Now, the bill that the 0:01:19.400 --> 0:01:23.800 Senate passed includes five fifty billion dollars of news spending, 0:01:24.319 --> 0:01:26.400 which is what puts it at the one point to 0:01:26.600 --> 0:01:31.399 trillion dollar mark total, and that includes stuff that you'd expect, 0:01:31.640 --> 0:01:36.720 like investment in roads and bridges that kind of infrastructure, 0:01:37.280 --> 0:01:40.160 or building out more rail systems for both freight and 0:01:40.280 --> 0:01:45.480 passenger trains, or building up more transit systems for various 0:01:45.560 --> 0:01:48.800 regional areas, that kind of thing. But there are some 0:01:48.840 --> 0:01:53.160 bits that we should mention beyond that. For example, there's 0:01:53.200 --> 0:01:56.440 a seven point five billion dollar budget for building out 0:01:56.560 --> 0:02:01.720 charging infrastructure for electric vehicles. Now you're probably aware there 0:02:01.840 --> 0:02:05.600 is a very hard push to transition away from internal 0:02:05.640 --> 0:02:09.959 combustion engine vehicles that depend upon fossil fuels. In fact, 0:02:10.000 --> 0:02:12.959 there are states that are really pushing to make sure 0:02:13.040 --> 0:02:17.679 that you know, no new internal combustion engine vehicles are 0:02:17.960 --> 0:02:21.680 sold in those states after a given year. But in 0:02:21.760 --> 0:02:24.519 order for this to be practical, we actually need the 0:02:24.560 --> 0:02:28.280 infrastructure in order to juice up electric vehicles that will 0:02:28.320 --> 0:02:31.680 replace all the internal combustion ones over time. Also, just 0:02:31.720 --> 0:02:35.960 as a reminder, these bills don't outlaw internal combustion engines. 0:02:36.080 --> 0:02:39.680 It's not like you know, in twenty thirty it'll be 0:02:39.720 --> 0:02:43.560 against the law to own a gas powered car. It's 0:02:43.600 --> 0:02:46.640 just that all new vehicles will need to be electric. 0:02:47.400 --> 0:02:49.400 And of course we also have to address how we 0:02:49.480 --> 0:02:52.720 generate that electricity. If the electricity comes from a power 0:02:52.720 --> 0:02:55.880 plant that's burning fossil fuels, if it's a coal powered 0:02:56.000 --> 0:03:01.240 power plant, then we're kind of shifting the problem to 0:03:01.320 --> 0:03:05.880 a different point along the chain. But anyway, that's one 0:03:06.000 --> 0:03:09.959 part of the bill. There's also a sixty five billion 0:03:10.240 --> 0:03:14.800 dollar budget to expand broadband internet access, something that is 0:03:15.000 --> 0:03:18.120 sorely needed in many parts of the United States. There 0:03:18.160 --> 0:03:20.600 are lots of communities here in the US that have 0:03:21.200 --> 0:03:24.920 little to no access to broadband and considering how much 0:03:25.040 --> 0:03:28.280 of our interactions with society now depend upon having a 0:03:28.320 --> 0:03:33.000 reliable connection to the Internet, particularly during times of crisis 0:03:33.120 --> 0:03:37.360 like the pandemic, that's a huge problem. It It creates 0:03:37.440 --> 0:03:43.160 an enormous divide that puts these communities at a severe disadvantage. 0:03:43.520 --> 0:03:46.480 But there are also a few items in that Infrastructure 0:03:46.520 --> 0:03:50.920 Bill that could surprise you. For example, one element one 0:03:50.960 --> 0:03:55.240 amendment in the Infrastructure Bill places are requirement on car 0:03:55.320 --> 0:03:58.840 manufacturers in the future, and according to this measure, car 0:03:58.880 --> 0:04:02.920 companies will have to in opreate a passive system capable 0:04:03.000 --> 0:04:07.280 of detecting blood alcohol levels in drivers. And the idea 0:04:07.320 --> 0:04:09.280 here is that the vehicle itself would be able to 0:04:09.280 --> 0:04:13.160 tell if a driver were under the influence, if their 0:04:13.240 --> 0:04:17.440 blood alcohol level exceeded the legal limit, and if that 0:04:17.480 --> 0:04:20.080 were the case, the car would not start, so the 0:04:20.080 --> 0:04:23.600 car would not allow the driver to actually operate the vehicle. Now, 0:04:23.600 --> 0:04:26.080 a lot of people have raised objections to this for 0:04:26.200 --> 0:04:29.680 various reasons, and I can I can get behind several 0:04:29.720 --> 0:04:33.240 of the reasons for those objections. Now, I am absolutely 0:04:33.279 --> 0:04:36.520 against drunk driving, but I also happen to know that 0:04:36.640 --> 0:04:40.120 tech isn't infallible. There are a lot of ways that 0:04:40.240 --> 0:04:45.159 this requirement could have unintended consequences. For example, car companies 0:04:45.200 --> 0:04:47.359 will have to figure out the system that can detect 0:04:47.400 --> 0:04:50.960 a driver's blood alcohol level and ignore everyone else in 0:04:51.000 --> 0:04:55.000 the car. So let's say you are a designated driver 0:04:55.400 --> 0:04:57.760 and you're going to give a lift to your friends 0:04:58.000 --> 0:04:59.920 and they've had a few too many. But you are 0:05:00.080 --> 0:05:03.760 designated driver, you are sober, and you're doing this so 0:05:03.800 --> 0:05:06.080 that your friends stay safe and they can still go 0:05:06.120 --> 0:05:09.440 out and you know, drink or whatever. The system in 0:05:09.440 --> 0:05:11.400 this car would need to be able to tell that 0:05:11.520 --> 0:05:14.640 you are sober, even if your friends in the car 0:05:14.800 --> 0:05:19.479 are all inebriated. So a passive system that's analyzing, say 0:05:19.520 --> 0:05:23.480 the alcohol content detected in the air would somehow need 0:05:23.520 --> 0:05:28.520 to differentiate between you and your passengers. False positives could 0:05:28.560 --> 0:05:33.520 cause frustrating and potentially dangerous consequences, and false negatives would 0:05:33.560 --> 0:05:36.279 be even worse, of course, if a system failed to 0:05:36.360 --> 0:05:39.359 pick up on the fact that a driver was intoxicated. 0:05:40.200 --> 0:05:42.920 I think this is a case where we see politicians 0:05:43.080 --> 0:05:47.560 leaning on a technological solution to a social problem, but 0:05:47.680 --> 0:05:51.200 the tech isn't quite up to the task to do it, 0:05:51.839 --> 0:05:56.200 and that means the solution really isn't a solution at all. 0:05:56.520 --> 0:06:01.120 And because the bill specifies a passive system, something like 0:06:01.240 --> 0:06:05.800 a breathalyzer wouldn't qualify because that's an active system. You 0:06:05.880 --> 0:06:10.600 must actively blow into a breathalyzer to use it, So 0:06:10.680 --> 0:06:14.320 a passive system would need to monitor alcohol levels without 0:06:14.400 --> 0:06:19.000 any like direct input or action from the drivers. It 0:06:19.200 --> 0:06:22.839 just seems like it's asking an awful lot of technology, 0:06:22.880 --> 0:06:26.440 and I'm not sure tech can actually deliver. The bill 0:06:26.839 --> 0:06:29.240 builds in a little time for car companies to actually 0:06:29.320 --> 0:06:31.800 do this, which is good. I mean, car companies have 0:06:31.920 --> 0:06:37.120 their production plans set out years in advance, so with 0:06:37.200 --> 0:06:39.360 the way the language of the bill is or the 0:06:39.400 --> 0:06:42.000 way the language of the amendment is formed, it would 0:06:42.000 --> 0:06:46.560 mean there'd be a range of at the earliest to 0:06:46.560 --> 0:06:51.520 to start incorporating this kind of technology. That date is 0:06:51.560 --> 0:06:54.839 there just in case the tech being mandated is found 0:06:54.960 --> 0:06:59.680 to not meet the criteria of being reasonable, practical, and appropriate. 0:07:01.200 --> 0:07:03.560 None of that. I mean, I guess practical covers it. 0:07:03.640 --> 0:07:06.719 But you know that doesn't sound like effective is among 0:07:06.880 --> 0:07:11.480 the three. Maybe practical is supposed to cover that. Another 0:07:11.520 --> 0:07:16.080 element in the Infrastructure Bill relates to cryptocurrency specifically, the 0:07:16.120 --> 0:07:20.520 bill includes requirements to report cryptocurrency transactions for the purposes 0:07:20.600 --> 0:07:24.440 of taxing them. After all, I mean, you've gotta pay 0:07:24.440 --> 0:07:27.960 for this one point to trillion dollar bills somehow. There 0:07:28.000 --> 0:07:30.400 are a few different ways of doing that. You could 0:07:30.480 --> 0:07:33.920 raise existing taxes, which tends to be a pretty politically 0:07:34.040 --> 0:07:38.360 unpopular move. People aren't crazy about that idea. You could 0:07:38.560 --> 0:07:41.560 shift money from some parts of the budget to a 0:07:41.560 --> 0:07:43.760 different part of the budget. But this also tends to 0:07:43.760 --> 0:07:47.200 be politically unpopular because it usually means some senator or 0:07:47.280 --> 0:07:50.480 other sees money that would go to businesses that are 0:07:50.720 --> 0:07:55.160 in their constituency diminish. So you'll see senators fight for 0:07:55.240 --> 0:07:58.720 certain things, not because they think that the thing they're 0:07:58.720 --> 0:08:01.720 fighting for is right, but rather the thing they're fighting 0:08:01.720 --> 0:08:05.360 for represents money that would be going to their state 0:08:05.640 --> 0:08:10.800 because the industries that represent that that particular element are 0:08:10.840 --> 0:08:15.520 located in their state. It does get very political. Another 0:08:15.560 --> 0:08:17.440 thing you could do is try and find loopholes that 0:08:17.440 --> 0:08:20.560 are in the tax code and close those loopholes. And 0:08:20.600 --> 0:08:22.880 that would allow the government to collect taxes on stuff 0:08:23.200 --> 0:08:27.520 that previously it couldn't but perhaps should have been able to. 0:08:28.000 --> 0:08:30.800 And that's kind of where the crypto part falls, this 0:08:30.880 --> 0:08:36.880 idea that there are these transactions that should fall into uh, 0:08:36.920 --> 0:08:40.800 you know, taxable income, but often don't unless they exceed 0:08:40.840 --> 0:08:45.720 ten thousand dollars. However, the language in the bill, in 0:08:45.880 --> 0:08:49.839 this crypto amendment is ambiguous, and that worries a lot 0:08:49.880 --> 0:08:53.240 of crypto investors. Now, some of that worry might be 0:08:53.240 --> 0:08:55.720 coming from folks who are trying to use cryptocurrency as 0:08:55.720 --> 0:08:59.120 a means of evading taxes, and uh, you know, you 0:08:59.160 --> 0:09:02.320 can't really take their arguments and good faith right there 0:09:02.320 --> 0:09:09.280 trying to circumvent an actual, you know, law, that's not fantastic. 0:09:09.320 --> 0:09:12.520 But others are concerned that because of that language, which 0:09:12.559 --> 0:09:15.839 talks a lot about brokers, and defines a broker as 0:09:16.040 --> 0:09:20.640 quote any person who, for consideration, is responsible for regularly 0:09:20.720 --> 0:09:24.760 providing any service effectuating transfers of digital assets on behalf 0:09:24.800 --> 0:09:28.320 of another person. End quote is pretty vague. It sounds 0:09:28.320 --> 0:09:32.360 like that could cover pretty much everyone in the crypto ecosystem, 0:09:32.400 --> 0:09:35.880 at least for proof of work cryptocurrencies like bitcoin. So 0:09:35.960 --> 0:09:39.559 let me explain. Let's say you've got a fraction of 0:09:39.600 --> 0:09:43.319 a bitcoin, right, You've got like point zero zero whatever 0:09:43.880 --> 0:09:46.640 one bitcoins, and you want to use this fraction of 0:09:46.640 --> 0:09:49.200 the bitcoin to purchase a brand new television from a 0:09:49.240 --> 0:09:53.000 merchant who happens to accept bitcoin is payment. So you 0:09:53.040 --> 0:09:58.280 make your transaction. You pay the required amount of bitcoin 0:09:58.400 --> 0:10:01.760 to the merchant, and you hand over your digital currency 0:10:01.840 --> 0:10:06.480 or essentially sending a digital certificate that transfers from you 0:10:06.559 --> 0:10:09.439 to the merchant. But this transaction has to be verified. 0:10:09.800 --> 0:10:12.800 So this way, the merchant knows for sure that the 0:10:12.840 --> 0:10:17.200 digital currency they're receiving is legitimate, that it's a real transaction, 0:10:17.520 --> 0:10:21.560 and that actual value is being exchanged. So the transaction 0:10:21.800 --> 0:10:27.240 process is part of bitcoin mining, transactions clumped together in blocks. 0:10:27.520 --> 0:10:30.120 Every ten minutes or so, a block is ready to 0:10:30.120 --> 0:10:35.200 be verified and then once verified that that verified block 0:10:35.480 --> 0:10:37.600 goes to the end of the blockchain and becomes the 0:10:37.600 --> 0:10:41.160 most recent block in the blockchain, and that's the chain 0:10:41.200 --> 0:10:43.920 of transaction blocks that stretch all the way back to 0:10:43.960 --> 0:10:48.520 the beginning of bitcoin. Bitcoin miners compete to essentially solve 0:10:48.600 --> 0:10:52.000 a very hard math problem that will then serve as 0:10:52.040 --> 0:10:56.800 the verification for that block of transactions. So in this scenario, 0:10:57.200 --> 0:11:00.520 would that mean every single bitcoin mining operation can count 0:11:00.559 --> 0:11:05.440 as a broker? Would it only be whichever bitcoin minor 0:11:05.559 --> 0:11:09.160 actually got the math problem correct first? I mean, that 0:11:09.200 --> 0:11:11.680 would kind of make sense. The reward for solving the 0:11:11.720 --> 0:11:17.280 math problem is some bitcoin, and I could definitely be taxed. 0:11:17.320 --> 0:11:19.360 But then you also have the issue that, you know, 0:11:19.440 --> 0:11:23.360 bitcoins a global cryptocurrency, the United States does not have 0:11:23.440 --> 0:11:26.600 jurisdiction over the entire world or anything like that. So 0:11:26.640 --> 0:11:29.400 it gets messy, is what I'm saying. And the fear 0:11:29.520 --> 0:11:31.760 is that the vague language in the bill will make 0:11:31.800 --> 0:11:35.000 it difficult, if not impossible, to execute the law properly 0:11:35.320 --> 0:11:38.800 without lots of unintended consequences. And this is the kind 0:11:38.840 --> 0:11:41.480 of stuff you see when after a survey you find 0:11:41.480 --> 0:11:45.000 out that more than all senators had never even heard 0:11:45.040 --> 0:11:50.240 of cryptocurrency before. Ignorance of technology often becomes an issue. 0:11:50.760 --> 0:11:52.640 Not Just to be clear, I'm not in favor of 0:11:52.679 --> 0:11:55.840 tax evasion, just like I'm not in favor of drunk driving, 0:11:56.160 --> 0:11:57.880 and I do think there needs to be some sort 0:11:57.880 --> 0:12:00.480 of system in place to close off loopholes that allow, 0:12:00.840 --> 0:12:03.560 you know, wealthy people to get even more wealthy at 0:12:03.600 --> 0:12:07.040 the expense of pretty much everyone else. And you really 0:12:07.040 --> 0:12:09.959 do have to be wealthy to be a cryptocurrency minor 0:12:10.040 --> 0:12:13.080 and effective one anyway, because the hardware you need in 0:12:13.200 --> 0:12:16.480 order to be competitive is very expensive, and the power 0:12:16.520 --> 0:12:20.400 bill to supply all that equipment with electricity is also expensive. 0:12:20.720 --> 0:12:24.800 So I don't necessarily disagree with the motivation for these 0:12:24.840 --> 0:12:28.000 amendments in this infrastructure bill, like I think they're coming 0:12:28.080 --> 0:12:31.760 from generally a good place, but I do worry about 0:12:31.800 --> 0:12:34.760 the actual language itself and how that will manifest in 0:12:34.800 --> 0:12:37.400 the real world. And it's always kind of chaotic when 0:12:37.400 --> 0:12:40.680 politics tries to catch up with tech. We've got a 0:12:40.720 --> 0:12:43.200 couple more stories to cover, but before we get to that, 0:12:43.440 --> 0:12:53.679 let's take a quick break. We're back, and if you've 0:12:53.720 --> 0:12:55.920 been listening to tech stuff for a while, you've heard 0:12:55.920 --> 0:12:58.640 me talk about the issue of bias. When it comes 0:12:58.679 --> 0:13:03.080 to artificial intelligence. Bias influences AI to favor one set 0:13:03.120 --> 0:13:07.000 of outcomes over others. Now, it's not always a bad thing, 0:13:07.400 --> 0:13:10.560 mind you. It really depends upon the application. But when 0:13:10.559 --> 0:13:13.400 it comes to stuff that relates to say, someone's appearance, 0:13:13.600 --> 0:13:17.600 or their ethnicity or something along those lines, bias usually 0:13:17.640 --> 0:13:19.520 turns out to be a pretty bad thing to have 0:13:19.600 --> 0:13:23.280 in your system. Well, Twitter recently held a competition among 0:13:23.520 --> 0:13:26.440 developers and hackers, and at the heart of it was 0:13:26.480 --> 0:13:30.800 Twitter's photo cropping algorithm. Now, if you've ever browsed Twitter, 0:13:31.320 --> 0:13:34.720 you've likely seen that there were photos in the Twitter 0:13:34.800 --> 0:13:38.160 feed are cropped and clicking on the photo opens up 0:13:38.200 --> 0:13:40.920 the full image. So the crop to photo access kind 0:13:40.960 --> 0:13:43.760 of like a preview and it gets all images to 0:13:44.000 --> 0:13:48.240 fit within Twitter's you know, established parameters for what photos 0:13:48.240 --> 0:13:51.480 should look like in the Twitter feed. But these photos, 0:13:51.480 --> 0:13:53.880 obviously they have to be auto cropped because there's just 0:13:53.960 --> 0:13:57.200 no way that Twitter could employ humans to do it manually. 0:13:57.720 --> 0:14:00.600 That means Twitter needed to design and now rhythm that 0:14:00.640 --> 0:14:04.280 would make decisions on where and how to crop photos 0:14:04.320 --> 0:14:07.760 so that they fit within Twitter's interface. And that meant 0:14:08.000 --> 0:14:10.920 there was the possibility that this algorithm would have certain 0:14:10.960 --> 0:14:15.720 biases that might favor, say, white faces over black faces. 0:14:16.160 --> 0:14:19.960 In fact, users in twenty suggested that this was happening. 0:14:20.240 --> 0:14:23.760 So Twitter's competition was all about inviting people to examine 0:14:23.760 --> 0:14:27.320 the photo cropping algorithm and look for signs of bias 0:14:27.440 --> 0:14:30.200 in a bug bounty, and it would reward people for 0:14:30.280 --> 0:14:34.200 the discovery of flaws in the system. This, by the way, 0:14:34.440 --> 0:14:37.400 is a pretty responsible approach in my opinion. I mean, 0:14:37.520 --> 0:14:40.400 of course, it would be best if the algorithm didn't 0:14:40.440 --> 0:14:44.040 have bias to start off with, but failing that, opening 0:14:44.080 --> 0:14:47.560 things up to a wider audience of critics and hackers 0:14:47.600 --> 0:14:53.160 creates the opportunity to isolate and address bugs quickly. At 0:14:53.200 --> 0:14:56.840 def Con, which is an information security conference held each year, 0:14:57.000 --> 0:15:00.200 lots of hackers go to it. The award for best 0:15:00.280 --> 0:15:02.600 performance in the bug bounty went to a person who 0:15:02.640 --> 0:15:07.160 found that Twitter's algorithm preferred faces that are quote slim, young, 0:15:07.680 --> 0:15:11.440 of light or warm skin color and smooth skin texture 0:15:11.760 --> 0:15:15.840 and with stereotypically feminine facial traits end quote. And I 0:15:15.880 --> 0:15:18.240 love how this hacker actually put this to the test. 0:15:18.480 --> 0:15:21.960 The hacker created a again, a g N, a generative 0:15:22.120 --> 0:15:26.840 adversarial network to create computer generated images of human faces. 0:15:26.840 --> 0:15:31.400 So again does this well? Again does anything by essentially 0:15:31.400 --> 0:15:36.040 having to AI systems two machine learning systems. One of 0:15:36.080 --> 0:15:39.320 those is in charge of generating something, so in this 0:15:39.360 --> 0:15:43.880 particular instance we're talking about computer generated images of faces. 0:15:44.640 --> 0:15:48.680 The other one attempts to figure out which stuff that's 0:15:48.720 --> 0:15:51.920 being fed to it is real versus computer generated, So 0:15:51.960 --> 0:15:55.520 in this case, what images were created by a computer 0:15:55.680 --> 0:15:59.160 and which images are just photos of people? And over time, 0:15:59.360 --> 0:16:03.200 both systems get better at their jobs. So if the 0:16:03.280 --> 0:16:07.360 detector is figuring stuff out much more frequently, then you 0:16:07.400 --> 0:16:12.440 feed that information over to the generator, which then adjusts 0:16:12.480 --> 0:16:17.000 how it goes about generating stuff and tries again. And 0:16:17.040 --> 0:16:20.760 then as the detector's success rate falls, it tries to 0:16:20.880 --> 0:16:25.720 adjust how it, you know, analyzes images and gets better 0:16:25.760 --> 0:16:30.200 at it that way. So, over thousands of trials of 0:16:30.440 --> 0:16:33.920 pitting these two systems against each other, you start to 0:16:33.960 --> 0:16:40.480 create more convincing computer generated images of faces um And 0:16:40.520 --> 0:16:43.720 in fact, you can create them so convincingly that if 0:16:43.800 --> 0:16:46.240 you were to look at a series of pictures, you 0:16:46.320 --> 0:16:48.600 probably wouldn't be able to tell which ones were real 0:16:48.680 --> 0:16:52.120 photos of actual people and which ones were just computer generated. 0:16:53.240 --> 0:16:57.080 Some other hackers also found other elements of bias. A 0:16:57.200 --> 0:16:59.960 one found that the algorithm had a negative bias towards 0:17:00.040 --> 0:17:03.040 people with white or gray hair, which means, before too long, 0:17:03.080 --> 0:17:05.000 you all won't be seeing any photos of me on Twitter, 0:17:05.680 --> 0:17:09.640 and that another one found that the algorithm would favor 0:17:09.760 --> 0:17:14.439 English text over Arabic text whenever those would appear in 0:17:14.440 --> 0:17:17.040 an image, like it would crop out Arabic text, but 0:17:17.080 --> 0:17:20.720 it would keep in English text. So that was an issue, 0:17:21.200 --> 0:17:23.920 and Twitter really took the right step here, in my opinion. 0:17:24.040 --> 0:17:27.479 We've seen other companies like Amazon really doubled down on 0:17:27.560 --> 0:17:30.400 denying that there's any kind of bias problem in systems 0:17:30.440 --> 0:17:35.800 like facial recognition technology, but Twitter instead accepted the criticism 0:17:35.920 --> 0:17:38.000 and then invited people to really drill down and find 0:17:38.000 --> 0:17:43.080 the issues, which ultimately, you know, could lead to Twitter 0:17:43.240 --> 0:17:46.600 creating better algorithms with less bias. So in my opinion, 0:17:46.640 --> 0:17:48.720 it was a great job for Twitter and an even 0:17:48.760 --> 0:17:52.679 better job to the hackers who found those bugs and 0:17:52.720 --> 0:17:56.720 really gross. News NBC reported that a company called Teleperformance, 0:17:56.800 --> 0:18:00.560 which supports tons of other companies including Apple and Amazon 0:18:00.560 --> 0:18:06.080 and Uber, has instituted a particularly draconian practice for employees 0:18:06.200 --> 0:18:11.479 in Colombia, so Teleperformances employees essentially do contract work for 0:18:11.560 --> 0:18:16.000 these other companies, often handling sensitive data in the process. 0:18:16.080 --> 0:18:19.920 And because of COVID, Teleperformance has shifted many of its 0:18:19.960 --> 0:18:23.440 offices around the world from going into the office to 0:18:23.640 --> 0:18:28.120 a work from home policy. But in Colombia, employees were 0:18:28.160 --> 0:18:31.960 required to sign new contracts that included clauses that allowed 0:18:32.000 --> 0:18:38.000 Teller Performance to install cameras inside the homes of those employees. 0:18:38.880 --> 0:18:41.880 Now why would it do this well to make sure 0:18:41.920 --> 0:18:44.720 that the employees were doing their jobs right, and also 0:18:44.760 --> 0:18:47.959 to make sure that no unauthorized individuals had access to 0:18:48.240 --> 0:18:51.439 the company's computers. But think about this for a second. 0:18:51.480 --> 0:18:54.000 I'm sure most of you would object to your boss 0:18:54.080 --> 0:18:56.720 telling you that they want to set up a camera 0:18:56.800 --> 0:18:59.439 in your home so that the boss can monitor you 0:18:59.520 --> 0:19:01.920 as you work from home. Add to that the fact 0:19:01.960 --> 0:19:05.720 that for many people, living space is limited, so their 0:19:05.920 --> 0:19:09.720 office might also be their bedroom, and you have serious 0:19:09.760 --> 0:19:13.600 privacy violations going on right there. Generally speaking, I am 0:19:13.640 --> 0:19:17.720 opposed to all the monitoring software and hardware that is 0:19:17.760 --> 0:19:21.880 in use to quote unquote guarantee productivity because I don't 0:19:21.920 --> 0:19:26.159 think it serves much purpose beyond being oppressive and invasive. 0:19:26.240 --> 0:19:29.919 I think it's far more demoralizing to employees than it 0:19:30.080 --> 0:19:34.000 is helpful to employers. And in this case, this issue 0:19:34.000 --> 0:19:36.400 goes above and beyond. I don't think there's any real 0:19:36.480 --> 0:19:40.600 defense for it. According to Interesting Engineering, a website where 0:19:40.600 --> 0:19:43.159 I read about this, Apple and Amazon have said that 0:19:43.240 --> 0:19:45.800 they did not request this particular kind of scrutiny for 0:19:45.840 --> 0:19:49.520 their work. But Uber was a different story. Uber did 0:19:49.600 --> 0:19:53.280 opt for this level, with the company defending the decision 0:19:53.359 --> 0:19:57.320 saying that contractors are working with sensitive customer information and 0:19:57.359 --> 0:19:59.720 the risk of that info getting out to someone else 0:19:59.800 --> 0:20:03.240 is prompted Uber to request the higher level of employee surveillance, 0:20:03.480 --> 0:20:08.359 to which I say, this system is broken. I have 0:20:08.560 --> 0:20:12.480 one other big story to cover, but before I jump 0:20:12.520 --> 0:20:23.280 on that, let's take another break, all right. Finally, in 0:20:23.359 --> 0:20:28.680 an ongoing story that I really wish would just conclude already, 0:20:29.400 --> 0:20:33.920 we're seeing misinformation campaigns about COVID surge here in the 0:20:34.000 --> 0:20:37.000 United States, while at the same time we're seeing case 0:20:37.119 --> 0:20:40.240 numbers on the rise. And why is this a tech story? Well, 0:20:40.280 --> 0:20:45.560 it's because the misinformation campaigns are heavily relying upon social 0:20:45.600 --> 0:20:49.880 networking platforms like Facebook and Twitter in order to spread 0:20:50.160 --> 0:20:55.720 the misinformation and to convince people to not follow the 0:20:55.840 --> 0:21:00.720 guidelines that scientists and doctors have come up with, which 0:21:00.760 --> 0:21:06.960 is making a bad problem monumentally worse. We have variants 0:21:07.040 --> 0:21:10.719 like Delta and Delta plus and Lambda, and these are 0:21:10.840 --> 0:21:15.760 virulent enough and resistant enough to vaccinations to occasionally cause 0:21:15.880 --> 0:21:20.919 breakthrough COVID cases. Now that's not always the case, and 0:21:21.000 --> 0:21:25.800 you are better protected if you're vaccinated, but the point 0:21:25.840 --> 0:21:31.880 being that these variants can break through vaccinations in certain cases, 0:21:32.119 --> 0:21:35.359 and yeah, on the whole, these cases tend to be 0:21:35.520 --> 0:21:39.520 far less severe than those for people who are unvaccinated 0:21:39.600 --> 0:21:43.040 and then get COVID. The vaccinated cases tend to be 0:21:43.160 --> 0:21:46.159 much more mild with their symptoms, but still not good, 0:21:46.359 --> 0:21:49.600 and it's still creating more opportunities for the virus to 0:21:50.000 --> 0:21:53.760 mutate further and more variants to arise as a result 0:21:53.800 --> 0:21:57.480 of it, and vaccination rates in some parts of the country, 0:21:57.680 --> 0:22:02.080 like my own state of Georgia, are still far behind 0:22:02.400 --> 0:22:05.399 where they should be. The New York Times reports that 0:22:05.440 --> 0:22:12.360 misinformation campaigns about COVID and COVID vaccines increased significantly over 0:22:12.400 --> 0:22:15.800 the last couple of months. According to Signal Labs, false 0:22:15.880 --> 0:22:21.320 information about vaccine effectiveness increased by four hundred thirty seven 0:22:21.359 --> 0:22:25.720 percent over the summer. Now, some of these campaigns originate 0:22:25.880 --> 0:22:29.760 from Russian troll farms, which are not just pushing out 0:22:29.800 --> 0:22:34.560 they're not spreading misinformation, they're pushing out disinformation outright lies. 0:22:35.520 --> 0:22:39.320 Then you've got people who become banner carriers. They pick 0:22:39.440 --> 0:22:43.520 up this this falsehood, and then they spread that falsehood further, 0:22:43.680 --> 0:22:46.760 and they promoted and they give it more credence, and 0:22:46.840 --> 0:22:53.600 the consequences are beyond awful. Obviously the worst consequence is 0:22:53.640 --> 0:22:57.800 that more people are getting very sick, and some of 0:22:57.800 --> 0:23:02.359 these people are dying. Right For the unvaccinated, it's awful 0:23:02.400 --> 0:23:05.440 because they're being they're they're getting information that are telling 0:23:05.480 --> 0:23:10.480 them not to get vaccinated, which drastically increases the chance 0:23:10.520 --> 0:23:13.040 that if they get COVID it will be a serious 0:23:13.119 --> 0:23:17.159 case and that they could in fact die. A lot 0:23:17.160 --> 0:23:20.080 of the ones, even the ones who who will survive, 0:23:20.280 --> 0:23:24.320 are ending up filling up hospitals, particularly here in the Southeast. 0:23:24.640 --> 0:23:27.480 That means that in regions like my state, where again 0:23:27.520 --> 0:23:32.560 at a point where hospitals are at capacity or over capacity. 0:23:32.640 --> 0:23:37.159 That means anyone experiencing any kind of medical emergency is 0:23:37.200 --> 0:23:40.480 in much greater danger because of a lack of medical 0:23:40.520 --> 0:23:43.840 capacity to treat them. So even if you're totally safe, 0:23:43.920 --> 0:23:46.680 let's say that like I'm staying home, I'm isolating, I'm 0:23:46.760 --> 0:23:50.720 not going out. I'm not you know, uh, exposing myself 0:23:50.760 --> 0:23:53.359 to the possibility of getting COVID. But I take a 0:23:53.400 --> 0:23:55.360 spill and I fall down my stairs and I break 0:23:55.440 --> 0:23:58.480 my legs, there might not be any way for me 0:23:58.560 --> 0:24:00.800 to get to a hospital in good time. I could 0:24:00.800 --> 0:24:04.800 be suffering for quite some time before anyone can treat 0:24:04.840 --> 0:24:08.840 me because all of the local facilities are over capacity. 0:24:09.480 --> 0:24:12.200 That's a real issue. And that's that's a minor one, 0:24:12.320 --> 0:24:15.520 right breaking your legs bad, but clearly there are much 0:24:15.560 --> 0:24:19.320 more serious medical emergencies that could arise where people would 0:24:19.320 --> 0:24:21.359 not be able to get the treatment they need. It 0:24:21.440 --> 0:24:23.880 also means that for all the people who did their 0:24:23.880 --> 0:24:26.440 best to stay safe, you know, the folks who went 0:24:26.480 --> 0:24:29.520 and got vaccinated, the people who are wearing masks and 0:24:29.560 --> 0:24:32.280 who are socially distancing, they're going to have to continue 0:24:32.280 --> 0:24:35.760 to do that or potentially become part of the problem. 0:24:35.840 --> 0:24:40.159 And there is so much quarantined fatigue out there that 0:24:40.200 --> 0:24:43.359 a lot of people who up to this point have 0:24:43.480 --> 0:24:48.080 been being super careful are just not willing to do 0:24:48.119 --> 0:24:50.680 that anymore because they feel like it's a punishment, right, 0:24:50.680 --> 0:24:53.679 Like they did everything right, but they're being punished for 0:24:53.760 --> 0:24:56.400 it by being told, hey, you have to keep doing this. Meanwhile, 0:24:56.560 --> 0:25:00.640 there are these other people who refuse to accept any 0:25:00.680 --> 0:25:04.360 kind of accountability, who are going out and doing all 0:25:04.400 --> 0:25:07.639 the fun things without wearing a mask or anything. And 0:25:07.680 --> 0:25:12.560 I understand that that feels monumentally unfair, and it is 0:25:12.600 --> 0:25:16.840 hard for me to condemn people who have that reaction. 0:25:16.880 --> 0:25:19.920 I mean, in full disclosure, I have even gone out 0:25:20.400 --> 0:25:24.439 to be with friends and a few isolated cases, and 0:25:24.480 --> 0:25:29.159 we did take precautions, but it still was more risky 0:25:29.200 --> 0:25:31.800 than just staying at home. Like I could have just 0:25:31.840 --> 0:25:34.200