 This 10th year of Daily Tech News show is made possible by its listeners, thanks to all of you, including Pillay Glendale, Dr. X-17, Dustin Campbell, and our new boss, Dario. Dario just started backing us on Patreon. Thank you and welcome, Dario! Coming up on DTNS, AI regulation standards are coming, but can we all agree on what is best? Can electronic brain implants help reverse parallelization? And how might algorithms affect the future of human resources? This is the Daily Tech News for Thursday, May 25th, 2023. From Studio Redwood, I'm Sarah Lane. From Lovely Cleveland, Ohio, I'm Rich Schroffalino. From Petaluma, I'm Megan Maroney. And on the show is producer Roger Chang. Megan Maroney, so glad to have you on the show and we're going to talk all about AI and what the future means. But first, before we get to the quick hits, MoviePass, you might remember was the $9.95 per month subscription for one movie per day. It went away but has re-emerged. It's now starting at $10 a month for one to three movies, up to $40 a month for up to $30. A little bit different of a deal, but it has re-emerged. Let's now start with the quick hits. Microsoft has made good on its promise to officially appeal the UK Competition and Markets Authority's decision to not approve Microsoft's acquisition of Activision Blizzard. 37 countries, including China and the EU, have approved the acquisition, but the US FTC is also suing to block the merger, so more to come on that. On Saturday, Neva announced it was shutting down its consumer-facing search engine, which didn't have ads, but also charge a subscription. Kind of a different take on search for some ex-Googlers. Wednesday afternoon, Cloud Data Management company Snowflake announced it acquired Neva. This was rumored for a while and now it is official. Snowflake intends to use some of Neva's generative AI tech to help enterprise users search their data. TikTok is testing a chatbot in its app called Taco. That's Taco with a K, like the Japanese word for octopus. For some users in the Philippines, it shows up on the right-hand side of the TikTok interface. Users can ask it questions about the video that they're watching or ask recommendations for new content. While it is still in testing, TikTok says it reviews all chats for safety reasons. The chatbot is provided by an outside company, which TikTok has not revealed what that company is. Amazon announced it's ceasing operations for its app store in China as of July 17th. Amazon has operated its app store there since 2011, so going on a dozen years now. It has previously announced its ebook service will stop operating in the country as well on June 30th. The company will continue to operate its retail site as well as Amazon Web Services in China. Vietnamese electric car company Vinfast has recalled all 999 of its SUVs in the US. This was because of a software glitch that made the main screen go blank in some cases, which could be a big deal because pedometer and safety notices show up on that screen. Vinfast is pushing an over-the-air update. Of the 999 vehicles, you might say, who has them? Well, 735 haven't been sold yet. 153 are in fleet service, and the other 111 were bought by customers. Those customers will get an email notification about this by May 29th. All right, well, the rumor is now in announcement. That's right, Sony made official with its game streaming handheld that they're calling Project Q. Sony says it will be able to stream games from a PlayStation 5 over Wi-Fi, similar to what you could do with the PlayStation app on Windows, Mac, iOS, and Android. Pretty much any platform you really want to. Project Q looks like, I guess, what would happen if you split a PS5 controller in half, like King Solomon style, and then mounted each half on either side of an 8-inch screen. I guess no one could determine which one loved the controller the most. Sony did not give a price or release date, but insider gaming's Tom Henderson sources say Project Q should come out this November. But that's not all. Sony also announced PlayStation earbuds, which promise lossless audio with low latency. Not a lot of info really about the earbuds right now, but the company promises more details on the earbuds and the Project Q handheld in the coming months. We talked about Sony's other announcements with Scott Johnson in yesterday's show. Rich, what stands out to you about Project Q? That it's weird. When a new product comes out from a company, the company is very forward with what they will tell you the product story is. But it's also easy to imagine. You're thinking how the company wants you to use this. Project Q to me is weird given the other solutions that Sony has out there readily available for remote play from a PlayStation 5, from streaming from PlayStation 5 to another device. Tons of options out there, and yet they're still coming with this Project Q. And what it kind of says to me is that I think PlayStation remote access and this kind of remote play is more popular than I guess I had assumed because by being only PlayStation 5 specific, that's their market. It's like the number of PlayStation 5s that are out in the public like less a certain amount are the amount of people that can buy Project Q. It has no other utility outside of that. It makes me think that this is a much more popular feature to use than I would have thought. Does that make sense? I would add in a way that the product is kind of a trial balloon. It's there to kind of test the waters to see if this is something that is a sustainable device. A while back, and by a while back I mean a couple of weeks ago, a month ago we were talking about Nintendo. Nintendo's video gaming strategy looks very logical if you take the point of view of the fact that Nintendo sees itself as a toy company. Sony is a consumer electronics company and so they come from the position of we have things that are devices that have circuits in it and that we sell that to consumers as opposed to Microsoft, which is services company. We sell software, really you don't care what you play it on as long as you use our services to do it. And so for Sony, I think it's just sort of ingrained in them that you have a product that you put onto the public as a device to engage people and then to gauge to see how open they are to that product category and if it's something you could milk ideally further down the line. But right now this is one of those products that you scratch your head because it's not a standalone device. It's not a device that you can use away from a Wi-Fi like if you had a data connection, mobile data connection. So it's kind of in this nebulous realm where we want to play games on it but it's going to be tethered to something much like a VR headset. Well, are you scratching your head because of the not knowing the price? For example, the Logitech G Cloud Gaming handheld, $150, Steamed X 400. A Switch can be bought for under $200 at this point. What would be the price point that would make it seem like it's worth the test? It's a peripheral. So regardless of at this stage, you still need the PS5 to go with it until they announce it's going to be a game streaming device. Not unlike the way that Nvidia's Shield TV originally was. You can connect to your PC and you can remotely play in your local area network and you can play all your PC games through your Shield TV on your big screen TV. Eventually they came out with GeForce Now and now you could do cloud gaming exclusively instead of having your PC on. I mean, Megan, I'm curious from your perspective. We're definitely in this everybody's cutting costs across tech here. We're trying to come to efficiency. I guess on the basis that this is a trial balloon for the service, does it make sense to you to invest in a whole new product? I mean, even though this is, yes, it looks like a PlayStation 5 controller, Sony makes displays and stuff like that. This is still a lot of R&D to put in something that seemingly replicates a software solution that's had for years now. Does that make any sense to you? Yeah, I mean, it doesn't make sense. There's so many devices we have so much stuff and things are lasting longer than they used to. I mean, I am not a gamer, but just having another device that does something that you already have a device that does the same thing doesn't make a lot of sense. I mean, this is the thing. It's catering to people already with a PS5. And so if you're one of those individuals that sees a PS5 as the center of their entertainment universe, this device actually makes a lot of sense. However, I don't think it's a very broad audience. I mean, it can't be. Like I said, it's literally the number of X-Tant PlayStation 5s less than that. Where the Steam Deck is all of PC, like anyone that could want to play any kind of game. The Logitech G-Cloud, which by the way, I haven't heard is burning things up in terms of sale at a decent enough price point, $3.50, I don't know. Not burning it up, but that can play like stream any cloud-based game, any Android game, like a huge addressable market, not burning it up. My only thought is that maybe this is Sony. This is a popular enough application for Sony, right? This remote play use case, right? That they're like, we're going to take every complaint that everybody has and we're going to knock out everything that we can kind of take care of. Obviously, latency is always going to be an issue. You know, your network is always going to be the issue, but like let's make top-notch control interface. Top, like a big display, 8 inches is bigger than you get on pretty much any other of these portables, right? Like let's take away everything that consumers could possibly complain about and then maybe that makes sense in this category. Honestly, real quick is I think, again, the trial balloon, they're going to test it out with the hardcore PS5 gamers. If they seem to like it, they could very easily broaden out the appeal by, hey, we now have a game streaming service. They'll need to plunk down $400, $500 for, you know, console system. You spend $2,000, $2,500, $300. You play this anywhere. You have a decent data connection or Wi-Fi connection. All right. Well, the European commissioner for the internal market, Thierry Breton, said the European commission will work with Alphabet to develop voluntary regulations for AI ahead of more formal regulations, which will take longer to develop the AI Act that's kind of been in circulation out there. The EC's commissioner of competition, Margaret Vestiger, said the EC is also working with the U.S. on minimum standards for AI regulation by the end of the year. But that's not the end of the regulatory story, is it, Sarah? No, it isn't. In the U.S., Microsoft president, Brad Smith, said that new U.S. federal agency, a new U.S. federal agency to oversee AI development is most sensible. Many people would agree. Open AI's Sam Altman expressed similar sentiments to the U.S. Congress last week. However, on Wednesday, Altman said the current draft of EU regulations for AI would be over-regulating and its transparency and safety requirements might be technically impossible for open AI to comply with. Sounds like we are gearing up for regulatory mishmash of all sorts of things. Megan, wanted to ask you, where do you land on the idea of AI regulation versus AI over-regulation? Well, I think I probably land right in the middle because I think there's not a good answer at all. So I feel like we're definitely in the AI hype cycle right now, mostly because of generative AI, because of things like chat GPT. And now you're putting this into the hands of people. It's free. It's now on your iPhone. And so people are thinking like, oh, I guess I need to be concerned about AI. Obviously, we've been using AI for years and years and it's already been in everything. So I think that bias is real in AI. We all know putting bad data in if bad data comes out. So yes, I think we need regulation for sure. I don't know what it'll look like. Does regulation limit innovation? Of course it does. But avoiding regulation could mean, you know, just headed in the same place we were when we didn't regulate social media. I like just move fast, break things again. So that's probably going to happen. Yeah, I mean, Rich, my feelings of somebody like Sam Altman, you know, heading up open AI saying, we need regulation, but we don't want to over-regulate because the technology shouldn't be, you know, you know, circumvented. But I don't know. Well, here's the thing. Sam Altman, good move, copying the talking points of Microsoft president Brad Smith, because this is classic Microsoft, like get ahead of regulation. We saw this with facial recognition. We've seen this already with kind of more early AI systems and that kind of stuff. Microsoft, especially in the Sachin Nadella, you know, era of the company is extraordinary. Like knows the timing for when they need to go to the regulators and say, hey, this is really big and important. Now AI, we did not take a genius to figure out, hey, we need to figure out an AI regulation strategy. But Sam Altman, I would say stick with the Brad Smith playbook of we want to, we want to hand in the table when it comes to this regulation. And you'll notice that that is seemingly where he is more having his issues with EU regulations that maybe haven't been made in consultation with AI companies, such as OpenAI, which he is obviously overseeing and running and founding. The other interesting point here is that the European Commission working with Alphabet, definitely again, kind of very early on in this generative AI hype cycle. Megan, definitely agree with you there. Seen as kind of, you know, coming from behind in a lot of ways to open AI and what Microsoft are doing seems doesn't seem coincidental that they would perhaps be more likely to talk to regulators informing these voluntary regulations, which might have the unfortunate side effect of delaying innovation in a field where for once they are not like number one with a bullet or leading there. I don't think that is coincidental. Well certainly unfortunate for a company like AI saying if you over-regulate us we can't actually do our job the way that we would like to. But you know what we saw in Italy? We saw that there were GDPR complaints and that they had to shut down their service to come into compliance and OpenAI didn't seem to have a problem doing that, whether you could call that regulatory overreach or not, what that regulation was meant to intend didn't apply to how OpenAI was doing their data sets. They still complied to it to the satisfaction of Italian regulators. Now I know there's a whole task force that's going on there, but you know, I'm a little suspicious with that. It may be that they can't do business exactly as they are doing it right now and run their systems exactly as they're doing right now, but I don't necessarily think, you know, we will see OpenAI leave the market or anything like that. Well, if you would like to hear us talk about more of this on this show, one way to let us know is with our subreddit. If you haven't hung out on our subreddit lately, a lot of smart folks are submitting stories and voting on others. You can do the same at dailytechnewshow.reddit.com. All right. The U.S. Equal Employment Opportunity Commission, or EEOC, is an independent federal agency tasked with enforcing federal laws involving discrimination in work situations. So hiring, firing, promoting people, you get the idea. So EEOC guidance carries weight with employers. Yeah, so it's important to have that context because last Thursday, the EEOC issued new guidance clarifying companies that may, excuse me, clarifying that companies may be held responsible for disparate treatment by algorithmically driven HR tools under Title VII of the Civil Rights Act, even if these tools were designed or run by a third-party vendor. So you're using a SaaS service or a platform or something like that, you would be responsible. As a reminder, Title VII prohibits employer discrimination, including using tests that exclude someone based on race, color, nationality, sex, or religion. So if you have an HR tool, software as a service, social media ad, or something like that, and you have these algorithms found to be discriminatory, the employer using it may be held responsible. So, Megan, I'm curious, what kind of responsibilities does this guidance put on employers that they haven't had before, I guess? Well, so they did, this is just a little bit different than what they've already had, but again, it has to do with the AI hype cycle because everybody is using this technology, HR people especially, there's a huge market in HR tech and usually HR people aren't necessarily tech people. So they're just like, yes, just choose a vendor from, you know, close my eyes, I'll pick it out, and I'll find that system that will help me hire or that system that will help me recruit or what a lot of people are using now, the technology that they're using now is surveillance technology because we're all working at home and you want to make sure that people are doing their work. So instead of just like, I don't know, trusting them, you could use this bossware software that, you know, knows when you've gotten up from your desk or anything like that, and all of that is using AI and the employers are now responsible to make sure that that AI is not biased and isn't treating one employee differently than another employee. So even if the HR person doesn't understand how the software works, then they could still be liable and, you know, for fines. I mean, isn't that inherently the issue here is a person might now be tasked with trying to understand AI that they don't understand was not part of their job description and could get them in trouble with somebody else that's working at the company. Yeah, exactly. I mean, and so it's like just a warning, like if you are in charge of choosing a vendor, like know what questions to ask them and, you know, you can Google it. You could probably ask chatGPT to tell you what you should ask the vendor about what they're doing. Exactly. But yeah, because you just, I mean, there's so many ways that this software is used. So for example, like let's say like you are Muslim and you want to get up from your desk five times a day to pray in the past, if you were working in an office or, you know, you would just tell your boss like, hey, this is what I do and your boss would say great, like religious freedom, I love it. Like get up from your desk five times a day and pray. But like if you're using an AI, if you're using technology to monitor all of your 1,000 employees and so all you know is that they're like, they've gotten up from their desk five times. They must be like binge watching, you know, whatever. Yeah, going out for smoke breaks or, you know, all sorts of things like, you know, assume somebody's doing if they're not at their desk. Right. And so like in this particular instance, if that person is treated differently and fired, then that's going to be, I mean, that's discrimination against your religious affiliation and that company will be held libel for that, even if they have no idea how that works, even if like it's a completely an algorithm that's like, okay, well, that person didn't meet their standard because they took too many breaks and I don't know why the AI knows why I don't need to know. Well, you do need to know. So that's why it's really interesting and, you know, we're, and that's sort of like this is kind of the problem with all of AI. Like I just happened to like follow HR tech because it's what I'm doing right now, but like it's true that everybody needs to know how this works as best that they can or they need to know the questions that, you know, nobody knows exactly how it all works, but they need to know the questions to ask to make sure that it is not being discriminatory or, you know, not being biased that it's accurate. It's all the things that we were talking about that like why we would need regulation for this. Now, one of the interesting things about this also is it's not just at the time of, you know, we acquired, we're going to decide to use this HR tool, but it's also doing regular and continuous impact assessments to make sure, you know, nothing's changing. My question though is, are there, did this guy set out any like benchmarks to, to kind of weigh like performance of these services on or is there any case law that would kind of point to a standard that these companies have to meet when, you know, using these algorithms? Well, the general rule of thumb is the four fifth rule is what, so it's, it's sort of like, if, you know, let's say like, you can't just say like, hey, you hired more men than you hired women, so you must, you know, you must be fine. It's sort of like, you have to say like you, you hired substantially more men than you did women. And so that's the four fifths. That's what the four fifths rule means. I mean, it's complicated. So it, but it's like, basically like 80% of the time if you did that, it's like, that's, that's a disparate impact is what they call it. But what this new guidance says is that when you're talking about algorithms and we're when you're talking about numbers, this big and complicated, you might, the four fifths rule might not be a benchmark. It could be much lower than that and you could be fine. So that they're, they're basic. They didn't give a new, they didn't give new guidance to replace the four fifths rule, but just basically a warning that like you better make sure that all this software in hiring, firing, recruiting, monitoring isn't isn't treating protected classes differently than everyone else. And do you think we'll see like a vendor response to this, you know, because right now we're talking obviously this puts the onus on the companies using these tools, but do you think we will see Hey, this, this tool is you know, certified to to, you know, you know, be title seven compliant or something, you know, like title seven compliant or either that or show a flat, you know, give a warning screen of you know, please ensure that, you know, under title seven, blah, blah, blah, you know, you to claim all liability. Do you think we'll see like, like the vendors kind of push this on companies to let that like raise awareness in that way. Yeah, I mean, I think I mean, I would hope so, like what the EEOC is saying that the companies are the employers themselves are being held accountable. So it's kind of like the vendors like, well, you know, but so I do think that it will probably I mean, if people know to ask the right questions. And HR team should ask how models are trained, like how whether the programmers whether the beta testers are diverse themselves, like, I mean, that's what we talked about for like bad data and bad data out, like if you, you know, just have a bunch of, you know, very, I don't know white guys, let's just say working as beta testers and programmers. You know, you're going to you're going to miss some things. So like those so companies when they're looking for vendors, they should make sure that that the vendors prioritize inclusivity and the people that they're hiring and the way they work and you know, just and you could ask them, like, did you use the fourth fifth rule? Like, even if you don't want to know, if they don't know what the fourth fifth rule, if they just look back at you and say then maybe find another vendor. Definitely definitely a red flag for sure. Well, this is sort of a feel good story. Gerd, John Oscom a Dutch man who was paralyzed in a cycling accident 12 years ago has been given the ability, at least some part of the ability to walk again simply by thinking about it. This is thanks to electronic brain implants. He had surgically installed in 2021 that wirelessly transmit his thoughts to his legs and his feet using a second implant on his spine instructing his leg muscles to work as needed. The progress was published in the journal Nature by Swiss scientists. In the if you're wondering how did this work in the operation the surgeon two surgeons actually cut two five centimeter circular holes on each side of Oxam skull then inserted to disc shape implants which wirelessly transmit brain signals to two sensors attached to a helmet on his head. Now, you know he has had quite a bit of success with this not everybody who is a paraplegic or otherwise would have the same results but this is science done good. When we're talking about brain-computer interfaces a lot of times we're like oh man this sci-fi future but this is the stuff that excites me this real world stuff obviously this is a long way from being a standard surgery but the fact that it's working for Gert Jan is just tremendous really exciting advancement hopefully we can see that become like a standard option depending on the level of paralysis. Indeed. Love to see a good algorithm. Indeed, speaking of speaking of that it is an algorithm that actually is part of this that's translating the signals into instructions to move leg and foot muscles using a second implant that's inserted around Gert Jan's spinal cord so yeah I mean hey that is that is promising stuff love to see it thanks to the BBC for letting us know about it all right Rich let's check out the mail bag Indeed in there we find that Victor wrote in on the option of booting a PC from directly from Windows 365 skipping over that old hard drive getting a more powerful PC with the hardware we talked about it yesterday on the show he says I used to work for a large company to find the PC models you were allowed to get based on your function but there were gray areas that caused engineers in certain functions to get quote, Fisher price PCs that didn't support resource intensive software things like CAD applications this was a big issue add to that traveling with a huge PC if got the workstation powerful enough for those resource intensive tasks just made everything much more complicated this definitely sounds interesting and yeah that's the use case I feel like that Microsoft is hoping get some more use out of your old PC send it right to that thin client right to the cloud in yesterday's show we were talking about hoteling stations and then said do everyone know what hoteling stations are this is if you go into an office and you have to share a PC with a bunch of other people and you know you have your login and that kind of thing this is sort of that but I think you know Victor's point to sometimes we had to do intensive stuff and we were traveling and in order to you know I don't know pack all that stuff to do our job was cumbersome and this would be a good solution for that fun fun stuff alright well thank you so much Megan Maroney for being on the show dropping the HR knowledge the AI knowledge all the good stuff making it exciting where can people find more of your great stuff if they want to keep following you online I am currently working at morning brew and specifically editing the HR Brune newsletter so you can subscribe to that it's HR dash brew.com and HR is everywhere and yeah Megan Maroney I'm also everywhere Megan Maroney on Twitter and I go to blue sky where people follow me and they don't like my tweets and also T2 that also happens where people just follow me and don't like my tweets so I'm all the places. Skeets tweets you know what Megan I think you're the best. Follow Megan and like her skeets tweets and toots or whatever is out there so yeah all the things also patrons do stick around for the extended show good day internet we are going to be talking about bug killing AI maybe not the bug that you're thinking of but just a reminder DTNS is live you can catch our show live Monday through Friday at 4 p.m. Eastern that's 2100 UTC you can find out more at dailytechnewshow.com slash live we are back doing it all again tomorrow talking about digital twins with Rob DeMillo and Len Peralta drawing the top tech stories talk to you then show is part of the frog pants network get more at frogpants.com hope you have enjoyed this program