Welcome to Tech Stuff, a production from iHeartRadio. Hey therein Welcome to Tech Stuff. I'm your host, Jonathan Strickland. I'm an executive producer with iHeart Podcasts and how the tech are you. It's time for the tech news for the week ending on Friday May seventeenth, twenty twenty four. And this was a pretty big week in tech news. There were a couple of large companies that made major announcements, and then there were tons of other things going on.
So we're going to start off with open ai, the company that is behind the chat GPT chatbot, among many other things. I mentioned last week that the rumor mill said open ai was on the verge of announcing an AI powered search tool, which would pretty much be akin to putting Google dead in the center of open AI's crosshairs, because Google was just about to have it's Google Io
Developers conference, which we'll talk about in a second. And open ai did release an announcement for a new AI model that's called GPT four O and I think that's the letter oh, not a zero. Anyway, this model is meant to allow for actual voice conversations between human and AI eventually, and open Ai says it'll also work across image and video related tasks. So this is what they
call a multi modal approach to AI. It's not just text based, it's other stuff too, And the concept is kind of like you could have an image of something or a video of something. You could be looking at things through say your camera app on your phone, and then talk to the AI assistant and try to get
information about what it is you're seeing. So we've seen stuff like this or related to this in the past, like holding your phone up to look at a sign that's written in another language and have real time translation to translate that into whatever your native language is. We've seen that in the past. This would be more advanced,
much more advanced than that, you know. The example I would think of is I would imagine like, say you're starting up your car and the engines making a weird noise, and so you turn your car off, you pop the hood of your engine, and you're turning the car on and you're trying to take a look and see what it is that's possibly going wrong. Well, if you're like me, you may only have a passing familiarity with that particular engine.
So an app that could hear and see the same things that you do, but have access to an entire world of information would be really useful. It could tell you, hey, this is something you don't need to worry about right now, or maybe it would tell you you need to get your vehicle checked right away, or maybe even give you a quick fix, if such a thing as even possible now.
In some ways, you could argue open ai is doing kind of similar stuff to what we're seeing with other personal digital assistants like Google Assistant or Siri or Alexa, but just turbocharged. Open Ai did a real time demonstration of the model's capabilities, something that would loom over Google a little bit for their io event, but we'll get
to them in a second. Anyway. Chat GPT four oh is rolling out now, but not all features are available rail of the gate, so for example, it is not able to talk to you just yet, or at least not to me. I actually check to see if I have access to Chat GPT four oh. I have a chat GPT account that I never use, and so I think I logged onto it for the first time in maybe a year and found out that sure enough, I've got access to it. But it doesn't talk to me yet.
The rollout is also going to be happening in batches, So if you don't have access to it yet, even though you have a chat gpt account, it is coming. You can actually sign in now by going to chatgpt dot com. They created that URL, so that's where the service lives now. I think that kind of is a sly indicator that they are transitioning from something that is largely talked about as kind of a test product into
something that's an actual product. Existing chat gpt account holders will receive access to four O capabilities over the next
several days if they don't have it already. And I want to echo a warning that I've seen a lot of other people make, because I think it's wise to remember be careful if in your interactions with chat gpt it suggests links to stuff, because apparently at least some malicious folks have found ways to get chat gpt to suggest links that ultimately lead to malicious website it's and malicious software. So don't trust everything Chad Gbt says, particularly when it comes to links. I think it's a really
interesting interaction that you can have with this thing. I don't fully like it. I have concerns, and we'll talk more about that in just a second. But beyond the announcement that open ai made, they also managed to close a deal with Reddit, which will give open ai permission to train AI models on Reddit and subreddit posts. That's something that folks had suspected open ai was already doing, just you know, without the part about having permission to
do so. Reddit will in return get some access to AI powered tools that can then be incorporated into the platform in various ways. So Reddit is saying, hey, don't worry, you're gonna have enhanced ways to interact with Reddit. Though what those will actually be haven't, you know, been fully disclosed,
if you recall. Open Ai similarly sign to deal with stack overflow not too long ago, which caused a bit of a tiff in that community because a lot of folks said they preferred their work not be used for training AI. But the terms of service for stack overflow are such that essentially, once you post something to the platform, it doesn't belong to you anymore. It's the platforms. So while you might object to your work being used to train AI, you don't have any authority to deny it
from happening. Some people tried to deface or delete old posts, but that's exactly that's against the terms of service on stack overflow. So many of those people subsequently found themselves banned, at least temporarily from that site. It'll be interesting to see if there will be a similar backlash in the Reddit community. That community is already kind of predisposed to
not liking the corporate leadership of Reddit. There's already pretty low opinion of the company's leaders, largely because of changes that Reddit made to its API like a year ago, and apparently that's not the case with investors. Investors on the flip side are pretty darn happy. Reuters reported that shares were trading fourteen percent higher today than they were yesterday in the wake of this news. Not all the
news is good news for open ai. Between announcements and partnerships, two prominent individuals in the organization resigned from the company this week. That would be Ilia Sutzkiver, who is a co founder of open ai, one of the original members of the non profit version of open ai, and the other one is Jan Leika, an executive who gave a pretty straightforward announcement on x their announcement. Just read quote
I resigned end quote. Now the alarming thing is both Laikah and Sutzkiver worked for a department within open ai called super Alignment, and essentially they were assigned a task that was the need to make sure AI evolves in a way that is helpful but not harmful. This was kind of the central premise, the reison deetre of open ai when it was first founded as a nonprofit organization.
The founders wanted to make certain that AI would be developed in a responsible and accountable and transparent way in order to maximize its benefits to humanity and minimize or eliminate any risks. But it turns out AI is really expensive.
Like it's expensive to develop, it's expensive to implement and to run and to maintain, and without a near constant cash injection, open ai just wouldn't be around for very long, which is why the organization ultimately spun off a for profit arm, which is the version of open ai most of us talk about today, like, it's not the nonprofit version, it's the very aggressive for profit version that makes the
news and that created conflict within the organization. It led up to the brief period in which the board of directors removed Sam Altman as CEO of Open Ai, before being pressured to turn around and put him right back in charge again, and then subsequently the board of directors essentially dissolved and reformed with new people. Sutzkever was actually part of the effort to remove Altman, although he subsequently apologized for that, so I don't think it's really a
big surprise that he has now left the company. But I think anyone who is concerned about ethical approaches to AI should really be aware that a lot of red flags are going up right now. Okay, with that fun note, let's take a quick break to thank our sponsors. We're back, and now let us shift to Google, which, as I mentioned earlier, held its annual IO Developer conference this week, and, to the surprise of literally no one, much of the
conference touched on artificial intelligence. Now there are do you have great headlines out there like Google aio or even Google aiaio, et cetera. So there's so many out there, I'm kind of sad that none are left for me. But enough about that. Google announced an update to its AI model Gemini. The company revealed that both the basic version, called Gemini one point five Flash and the pro version
are available for public preview now. The primary function of these specific AI models is to help developers create stuff, so if a developer is working on an app, they can make use of these models to help boost those efforts, whether it's encoding or whatever. Google also revealed an updated
version of its assistant called Project Astra. Google showed a pre recorded demonstration of the assistant's ability, and a Googler was taking her phone around an office and asking questions and having the camera app activated on her phone so she could point the camera at something and ask a question about the thing that's on screen, and the an
Astra would answer. So one point, for example, she asks the assistant to point out anything that makes sound, and she scans a desk, and when she brings the camera in view of a speaker, the assistant says, that's a speaker, and then she actually uses a markup feature to circle the tweeter on the speaker, and she asks a follow up question saying what is this part of the speaker called,
and the assistant says, that's the tweeter. Then it explains what the tweeter does, which is to playback you know, higher pitched sounds, higher frequency sounds, so it was a cool demonstration. She pointed out the window and asked, you know,
essentially like what's this neighborhood? And the assistant answered, and she at one point even said that she had misplaced her glasses and asked the assistant if it knew where her glasses were, And because she had been walking through the office with the camera on, the app was able to reference that information and it was able to pick out a moment where the camera saw some glasses on a desk, and it directed her to those glasses, which,
from what I understand, that's what got the biggest reaction from the crowd, which was kind of cool. I mean, goodness knows, I misplaced things all the time, so it'd be good to have, you know, an assistant that could tell me, yeah, no, that's in your hand. You'd do
fas anyway. Later on, Googler Michael Chang somewhat cheekily shared on X now known as you know X it used to be known as Twitter that he had watched the Open AI announcement on Monday with Astra, which is kind of funny because you know, the whole thing was that Open AI was doing this sort of to undermine Google and that Michael Chang saying, yeah, I watched it along with Project Astra, and Project Astra made a real time
transcript of that presentation, complete with attributions for each speaker, so identifying who was speaking in the transcript as it was going on. That's pretty impressive, and also giving some commentary about the presentation as it went on. The demonstration of the AI interpreting images and giving relevant responses was really interesting, but again, it was pre recorded. It wasn't like it was a live demo. Google also demonstrate an
updated image generation AI model called imagen three. On top of images, Google showed off a music AI sandbox to let folks use AI to put musicians and composers out of a job. No, I'm sorry, that's not a messaging they wanted. I mean to help artists elevate their creativity through the use of AI. I looked at these I just got to see demos. I don't have access to the tools themselves because I'm not a notable musician. I'm not really a musician at all, so I don't have
access to these things. But it's both fascinating and concerning. Because I have a lot of respect for actual musical artists, and I don't like the idea of people who are really talented and creative potentially getting pushed aside for something that is perhaps easier to access and maybe cheaper. That's not great. I think people's talent and time is worth money, like I think it's worth paying for and that you know, you can't just expect something for free, and anything that
undercuts that really concerns me a lot. Google also announced video FX, which you guessed it uses generative AI to create video clips, which again raise lots of concerns, like the more you're seeing of AI generated content, the more you're worried about how reliable is the stuff you are seeing.
Whether it's you know, a listing for a real estate property, like who's to say it wasn't a picture that was then enhanced by AI to make the house look way nicer than it really is, Or maybe it's something where it's a it's a video clip that seems to put a public figure in a bad light, and you start wondering, Okay, well is this real or is this something that was created by someone who just has an axe to grind for whatever reason these are things that are getting more
and more relevant. Arguably they are incredibly relevant right now. Now. The thing that is probably going to have the largest impact, at least the earliest for most folks, is Google implementing more AI into its search tool, which is of course what Google is primarily known for, even though it's arguably
really an advertising company, it's known for search. The company announced that AI overviews in search is going to be rolling out nationwide in the United States this week, so if you live in the United States, you probably have this incorporated in your search right now. Google also said that multi step reasoning capabilities are coming soon, which means that you'll be able to ask more complex questions in
your search query and to get relevant answers. So, like, here's an example I just made up, how do I make a lasagna? And what cheese is a good substitute if I don't have ricotta? Right, Because that's a two part question. It's one asking what are the steps to make a lasagna? And I know I already know I don't have an ingredient that's in a lot of lasagna recipes, So what is a good substitute? That's the sort of thing that you can expect to be able to access
in the not too distant future. The various keynotes and presentations stretched across multiple days, so there was so much more brought up during the IO event. But I think the real message for most of us, the people who aren't like actively developers, that we should take home is just that AI is going to be all up in
your business and soon. Google hasn't just been talking to developers, however, the company has also been pushing out updates to its browser, Google Chrome, and this is not part of its all AI, all the time strategy. Instead, Google is addressing some zero day vulnerabilities that have come to light over the last couple of weeks, three of them in fact, recently. These vulnerabilities are severe. Potentially hackers would be able to quote perform an out of bounds memory right via a crafted
HTML page end quote that's from Google. That's not good. It could mean that you could have a vulnerability activated and then you visit a malicious website and then either get malware downloaded to your computer or worse. So this is a gentle reminder to all of y'all that it's never a good idea to just push off installing updates on stuff like web browsers and operating systems, because doing
that puts you at risk for malware or worse. I know it's a hassle to have to update, especially if you get to your machine and you've got like a specific thing you need to do, and then the first thing you see is a notification saying that you should run an update. I know that's a hassle. It happens to me too, and I hate it too, But it's a good idea to go ahead and do the update
ninety nine times out of one hundred. One time out of one hundred, it ends up being the wrong thing to do, But when it's a ninety nine percent success, right, you should probably do it. Also, hey, fun concept AI can now help hackers discover and exploit vulnerabilities, So it's just gonna get worse. Brave New World and all that kind of stuff. Apple iOS beta testers say that iOS build seventeen point five has a disturbing habit of bringing the dead back to life. And so I'm not talking
about reanimated goules here. I'm talking about deleted material, like deleted photos or deleted voicemails. According to a Reddit thread, some users have reported iOS digging up old and sometimes deleted photos and then showcasing them as if they are recent images. So you pull up your app and say, huh, according to this, this picture I took seven years ago is a new one. Apparently in some cases this has included images that people once took but long since deleted,
including nudes. Nude photos. That should be a massive concern because Apple's stated practice is that once you delete a photo, then you have thirty days in which you can recover the deleted photo. But after thirty days, that photo is supposed to be deleted for good. So if, as some people have claimed, some of these photos have reappeared reportedly years after they were first deleted, that raises serious privacy questions.
Right Like, if you took a nude photo seven years ago, then like six and a half years ago, you deleted it. You do not expect that seven year old photo to come bouncing back today. That would be really shocking, and it does again raise real questions about data retention and privacy and security. At least one Reddit user claimed that they had sold an old iPad to a friend of theirs, but they had gone through Apple's process to wipe their
own personal data off of the iPad before selling it. So, in other words, just making sure that you're reverting back to like factory settings and deleting all personal information, but that some of their old not safe for work photographs have been popping back up on their friend's device. Now surprise friend that you didn't expect to see your buddy in their glory. This is not good news at all,
if in fact it is true. So I have to keep on stressing if it is true, because I haven't seen confirmation from anyone other than people commenting on Reddit. I don't know if you know this, but sometimes people on Reddit say things that are perhaps not one hundred percent true. A lot of subreddits are almost exclusively made up of fiction anyway. Wes Davis of The Verge extends some grace to Apple and points out that this might not be a case of Apple doing something insidious or
being really negligent. That deleted data is never really gone until some other information gets overwritten on top of what once was there. And that's true if you delete something off your hard drive, but you don't overwrite information on top of where that drive was, you know, storing stuff it can be possible to recover the deleted information. So maybe that's what's happened at Apple, just you know, on
a much much larger scale. Now, personally, I find that a little hard to swallow because the trend has been for more folks to use more Apple services over time. So you would think if more people are using those services, that stuff would be overwritten all the time, because otherwise Apple would just have to keep on adding additional servers to handle incoming images and voicemails and all the other stuff,
because you know the otherwise. The implication is that, well, we haven't overwritten everything that you know has been freed up since people have deleted things. I can't even be sure that the claims that are being made are accurate. So it's possible that some of these quote unquote deleted files weren't really deleted at all. Maybe they were just forgotten about, because goodness knows that happens too. But the story about the wipe dipad in particular is trouble if
it is true, So just something to be aware of. Again, that's for iOS seventeen point five, and it's in beta, so if you're not beta testing, then this isn't likely to affect you if you're an iOS user, but I'm sure it's something that Apple is looking into to figure out before iOS eighteen goes out. All right, we got more stories to cover, but before we can get to
all of those, let's take another quick break. We're back, and here's a story that hit like deja vus this week at centers around the company game Stop, and it's stock price. So you might remember that back in twenty twenty one, a collection of folks decided to stick it to some hedge funds by investing in game Stop because word got out that some massive hedge funds were short selling game Stop. By driving the price of the stock up,
the investors could do two things. One, they could see a return on their own investment, right because if the stock goes up, then you're making profit. And two, they could cause serious financial hardship for these massive hedge funds that typically hold a tremendous amount of power in the stock market. Now, if you're not familiar with how short selling goes, let me give you a very basic description,
and I'm glossing over a lot of details here. Let's say that you believe a particular stock is going to decline in value. So you decide, hey, I'm going to make some money with stock going down. I'm going to sell stock that I don't actually own, and i'm going to sell it at what it is currently trading for. So let's say that it's currently trading for ten dollars
per share. So you sell one hundred shares at ten dollars per share, but you don't own those shares, So you're literally selling stock you do not own, but you have the promise of delivering that money, or rather the stock, in a certain amount of time. Let's say it's one month. Now, a month has gone by, and it's time for you to deliver the stock that you've promised. At this point, the stock has gone down to one dollar per share.
Now you previously sold it for ten dollars per share, so you get to buy up the amount of stock you agreed to handover in the transaction, and you get to pocket the difference between the ten dollars that you accepted for this sale and the one dollar that you actually purchased everything for. So you make money by stock going down. But if the stock price doesn't go down, if it goes up, that's bad. News for you, because now you're gonna have to pay out a pocket the
difference of that. So let's say that the stock actually rose up to twenty dollars per share at the end of the month, Well, then you're gonna have to pay an extra ten dollars on every share of the stock that you sold, So you're gonna lose that money instead of gaining extra Well, these investors that ended up driving up the price of game Stop were led by a guy named Keith Gill, who's better known by his online handle, which is Roaring Kitty, and they did a massive screw
you to major investment companies that are trying to short sell game Stop. This is what we call a short squeeze. This week, it happened again briefly because on Tuesday, game Stop shares rose to a high of sixty four dollars eighty three cents per share, and in fact, there were companies that were putting on a short sale of game Stop stock. However, in this case, the rally didn't last long. It has subsequently dropped to around twenty seven dollars per share.
And this roller coaster followed some posts by the aforementioned Roaring Kitty who seemed to be getting back into the game after being quiet for a couple of years. But again, unlike the first go round when this happened in twenty twenty one, this particular boost appeared to just be a blip, so we're not seeing a full repeat of the phenomenon. So a lot of hedge funds have gotten wise to this and have measures in place to limit their liability so they're not caught out quite as badly as they
were the last time. Because boyotty, that was a doozy. Okay, let's talk about Meta. They are ending the Workplace app for external customers. At the very least, some elements of Workplace may remain as internal Facebook systems. So Meta first launched Workplace in twenty sixteen. It's a business collaboration tool. Essentially, it lets companies make interconnecting page networks with their partners and helps with collaborative projects that span different departments or
different companies entirely. They can have like dedicated message boards, that kind of stuff. It's a business to business product. Meta charged four dollars per user. That adds up pretty quickly if you're talking about a really big company. But now Meta is going to sunset the Workplace app and shift assets to two other massive projects, which are AI and of course the Metaverse. Because Meta is still determined to make that happen. I think it's going to be
like Fetch. It's not gonna happen, but maybe it will. I mean, maybe people will love it. I still am skeptical, but I'm also not a genius and I'm not an entrepreneur. I'm just a grouchy guy who talks into a microphone. The company says it will work with current Workplace customers to transition them over to work Vivo, which is a product from Zoom and the one partner they work with that makes something similar that they actually recommend. Okay, here's
an interesting story about suppressing tech. Seven hundred and fifty miles to the east of Australia is New Caledonia. In the mid nineteenth century, France claimed New Caledonia and has remained the ruler of New Caledonia since it is a French territory. But recently the government of New Caledonia enacted changes to election laws that really sparked protests across the territory, mostly in indigenous populations that have been fighting for independence.
For quite some time, these protests grew and became more chaotic. Four deaths have resulted as the protests have grown more violent, and since then the government has cracked down on the populace. The government has banned public protests and demonstrations and also blocked access to TikTok all in an effort to squash
protests online and off. The government says that people were using TikTok in order to organize demonstrations, and since demonstrations are now banned, so is TikTok, so people can't use it to do that. That's really why I'm including the story here, because it's a social and political event that
has spilled into suppressing access to online platforms. And as to the heart of the matter, like what this is all about again, it's largely about the question of independence for New Caledonia and the recent election laws that were changed now allow people who are more recent transplants to New Caledonia to vote. Previously, in order to be eligible to vote in matters of great import in New Caledonia, you had to have a long and deep connection to
the territory. You couldn't just move there and then influence the vote so a lot of people who support independence say this is a blatant attempt to dilute their influence in public matters, because the people who are moving to New Caledonia are largely French citizens, who are of course not voting in favor of independence. So that's what's really the heart of the conflict. But yeah, like I said,
it's spilled out now with a ban on TikTok. And while there is a twelve day state of emergency currently in effect for New Caledonia, it is not yet clear if the ban on TikTok will lift once that state of emergency expires. A lawsuit targeting Tesla alleging that the company made false claims regarding full self driving capability these in its car may move forward. Tesla had attempted to
get the case thrown out. A California resident by the name of Thomas Lesavio alleges the company engaged in fraud by claiming that its vehicles would have full self driving capability and that's why Lasavio purchased a Tesla vehicle, only to subsequently find out that those claims were perhaps exaggerated or false. The judge overseeing the matter has decided that at least some of Lasavio's case has merit and can go to trial, but some other parts have been dismissed
because of the way they were presented. Now, Lasavio originally was going to be part of a class action lawsuit, but the other plaintiffs of that lawsuit had to drop out because they had signed an agreement which would require them to go through arbitration with Tesla before they could bring a case to court. Losavio did not sign such an agreement, so was able to move forward with his own lawsuit, and he has opened it up to be a class action so other Tesla owners can join if
they so wish. Now, Lasauvio is going to have until June fifth to amend his complaint in order to get parts that were dismissed reincorporated into it. Tesla has until June nineteenth to respond. So there's still no guarantee that this is actually going to go to trial, but it is a step closer. Reuters has a piece titled the inside story of Elon Musk's mass firings of Tesla's supercharger staff. Obviously,
this relates to a story I talked about. A couple of weeks ago in which the company laid off a five hundred person department responsible for the development and deployment of supercharger stations, which seemed more than a little weird because Tesla has made great strides in establishing its supercharger technology as a kind of de facto industry standard. You know, other car manufacturers have bought into that, and they've made
their own vehicles compatible with it. The Reuter's piece tells a pretty disturbing tale in my opinion, because the former head of the supercharger department, a woman named Rebecca to Nucci, had already gone through a round layoffs like she had already held layoffs within her division at the direction of Elon Musk, and Musk wanted to do more. But she was meeting with Musk to present her vision that involved the company expanding the supercharger network. Instead, Musk demanded more layoffs.
To Nucci reportedly said that if she were to lay off anyone else in her department, that would be counterproductive, it would hurt their business objectives. And that's when Musk just went nuclear. He pulled the trigger and he fired everyone or practically everyone in that department. And it's very hard for me not to get the feeling that this is a leader who, when he perceives a challenge to
his absolute authority, just goes with that nuclear option. And now Tesla's having to deal with the consequences because people from other departments, notably the Solar and battery divisions, have been given the task to work with the various partners and contractors who had been engaged in building out supercharging stations. You know, these are folks who already had a lot of work on their plate, and now they're to take
on the work of an entirely different department. It seems like the general strategy right now for Tesla is to stall for time, which isn't exactly welcome news if you happen to be a contractor who's expecting to get paid. Musk has said the plan is still to grow the supercharger network, but that things are gonna have to happen at a more methodical pace, which seems like that's a
given considering how many people were fired. There have also been reports that Tesla has subsequently tried to hire back at least some of the staff that were previously laid off. I'm sure for some of those folks coming back is a pretty hard decision. I mean, you do need to make a living, so there is that element. But if you've already been unceremoniously laid off for arguably irrational reasons, it's not really great to think about coming back to
that company. It's also not a great development for electric vehicles in general, because if Tesla fizzles out, then some other entity or multiple entities will have to step forward to provide the infrastructure necessary to support a migration from internal combustion engine vehicles to evs. So that's one of the big reasons why people say they're they're reluctant to adopt evs is that they fear being stranded someplace because they weren't able to find a charging station and recharge
at a reasonable limit and cost. So, yeah, this is not great news for the EV industry as a whole. Now, when he's not gutting his EV company, Elon Musk is sparring with his old nemesis, the US Securities and Exchange Commission, or SEC. He's had a few run ins with the SEC, one stemming from his inability to not post stuff that could be seen as market manipulation. When he intimated that
he was going to take Tesla private. Anyway, now the SEC wants to chat with Musk about his purchase of Twitter, of course, now known as X. The SEC did this once before. Musk initially told him to pound sand and that he wasn't going to listen to all their subpoenas because they were just trying to bully him. After that, some judges weighed in to say, no, no, no, that's not how the law works, and Musk tried again get the whole thing swept away because you know, it's such
a hassle to have to answer for your actions. It turns out that's not a good enough reason for judges to throw things out, even for billionaires. So it looks like Musk is going to have to have a chat with the SEC, which says that it came into possession of new documents about the purchase of X and it raises questions about Musk's involvement with Twitter in the months leading up to his announced plans to acquire the company.
So we'll have to see where that goes. Considering the decline of Twitter since Musk's announcement, this feels like it's
largely a lot of fuss over a rotting corpse. Honestly, One story that I don't think I covered before is that the first human recipient of a Neurlink implant, which is the brain computer interface chip that's made by Neurlink, another company owned by Elon Musk It has experienced some issues as several of the very thin threads that connected to the device have lost contact with the neurons that they were previously attached to in this human subject, So
this affected the chip's ability to serve as a conduit between the recipient's brain and the computer systems he was interacting with. Engineers addressed the problem by making tweaks to the algorithms that run along on the chip, and that helped compensate for this issue. But now some unnamed sources that are familiar with the testing say that this was a known thing that the Neurlink had been observed to
have wires retracting in animal test subjects as well. Considering the fact that this implant requires invasive and potentially dangerous surgery, some are now questioning the ethical implications of moving forward
with a human test subject. While this retraction thing has been a known issue, there's a concern that should more wires retract, the neurolink will hit a wall with however much adjustment you can do to the algorithms to deal with it, and that in turn, that'll mean the chip will gradually lose functionality and potentially become largely useless, which could lead to an eividuality in which it becomes necessary
to remove the implant entirely. As for finding a way to fix this problem in future procedures, that's pretty challenging. You could create a system where you have more wires or they're more firmly anchored, but that would mean they would be harder to dislodge by definition, and if you wanted to do something like an upgrade, or if you needed to remove the implant, it could cause complications. So this is not an easy fix one way or the other.
And in my mind, it's suggest to me that perhaps this was a bit too aggressive for neuralink, but we don't know that yet. Like it may turn out that some wires retracting is just a normal course of action and that the rest of it is fine. And there's no denying that the technology does provide an incredible quality of life boost to people who otherwise don't have use of their limbs. That is something that is really important.
So if in fact we don't see further degradation of the implant, I think that this issue, while it does raise concerns and raises questions about why the FDA would allow things to move forward, those are questions that I think need to be answered. I think ultimately we could say it's for the best if the device actually does work and doesn't degrade to a point where it must be removed. I mean, it's in arguably an improvement to
the life of the people who have it. Touching back on AI for a moment, Sony Music Group has joined countless artists, writers, and other creators and expressing concern over AI using existing human created works as training material, and Sony Music is saying, don't do that. We don't want
you to do that. So the company has reportedly sent out several hundred letters to various AI startups and companies saying, hey, don't you dare use our catalog of music without our express permission, And at least for some of those recipients, the company went on to say, you know, we know that you're doing it now, so you've got to knock it off Buster, except, of course, they worded it differently.
Sony specifically said, quote, due to the nature of your operations and published information about your AI systems, we have reason to believe that you and or your affiliates may have already made unauthorized uses, including TDM, of SMG content, in relation to the training, development, or commercialization of AI systems. End quote. For the record, the initialism TDM stands for
Text and Data Mining. SMG is Sony Music Group. According to jem Oswad of Variety, a recent EU law called the AI Act prompted Sony to send these letters, as the Act compels AI companies to reveal what their training data was like where did they get their data to train their AI models? So Sony, at least in part, is saying, Hey, you used our information without our permission.
Don't do that now. I usually hold off on reading recommendations to the very end of the episode, but I want to mention Ashley Bellinger's incredible piece for Ours Technica titled MIT Students stole twenty five million in second by exploiting etch blockchain bug. DOJ says, now, as that title implies, this is a heck of a story. It's about a pair of brothers who learned how blockchains work at MIT.
They arguably learned a little too well how they work, and they took this incredible education that they received at MIT and they turned it into kind of an Ocean's eleven style heist. Except this heist took less than fifteen seconds, and it involved finding a way to access pending transactions on the blockchain. So these were transactions that have not yet actually been cemented in the Ethereum blockchain, and they were able to divert cryptocurrency meant for these transactions to
go into their own accounts. So the way blockchains work is that once a transaction gets added to a block and the block is added onto the chain, it's secure because going into the history of transactions and making a change would cause a ripple effect for every block that follows, and so any tampering immediately becomes clear and is reversed because the entire network is able to see that someone
is messing around. The brothers, however, figured out how to tap into transactions that had been initiated but not yet assimilated into the blockchain, and they would have gotten away with it too if it hadn't been for you know the fact that their victims made a really big fuss about this, and it prompted an investigation from the irs, which then they said was incredibly easy to conduct because the brothers were not particularly good at covering their tracks.
They tried to cover their tracks, but they didn't try to cover their tracks in a smart way. Apparently the brothers spent a lot of time searching for stuff that was a little bit suss, like how to create shell companies, how to launder money, and you know who's a good crypto lawyer, which is all, I guess, arguably circumstantial, but it's not a good look. As the kids used to say, This is just a gentle reminder for all of y'all
out there. If you're going to commit a crime, and I strongly urge you not to do so, maybe don't search stuff while you're logged into a browser on a personal device you know, or logged in at all to a browser. You want as many layers of obfuscation between what you're doing and you as you can manage, and your browser history shouldn't be a treasure map that leads
authorities to your criminal behavior, because you just brought them there. Man. Now, I just did some episodes about live service games this week, and I mentioned the company rock Star in those shows. Rockstar is the video game company behind the Grand Theft Auto series. Jennifer Mos of Variety says that that Rockstar had a two point nine billion dollar loss last quarter.
That is a heck of a princely sum to have as a loss, and it's a here bit more money than what was previously estimated because the original estimate shortfall was going to be one hundred and seventy million dollars, So, you know, just two point eight billion dollars off. Anyway, about two billion of that was a quote unquote goodwill chart. That seems like an oxymoron. It's kind of like the convenience fees you see on services like Ticketmaster, and you're like,
who is this convenient for? But no goodwill charge just refers to cases in which a company spends more money acquiring some other entity than the value of that entity's assets and whatnot. So it's like saying, this money represents us paying more than what stuff was worth. But in good news, gamers now know they can expect the long awaited next game in the GTA franchise in the fall of twenty twenty five. Okay, I got a couple of recommended articles for y'all before I sign off. First up
is Jacob Stern's piece in The Atlantic. It's titled The Dream of Streaming is Dead, and it goes into the recent trend of bundling various streaming services together and how this is essentially reverting to the older model of cable television packages rather than the dream of freeing people from that experience with the prospect of streaming. Next up is Alex Heath's article in The Verge titled Goo and Open
AI are racing to rewire the Internet. It goes into more detail about how the implementation of AI into search can and will have a huge impact on the Web. In particular, a lot of content sites out there are wondering what the heck is going to happen because of all that, So check those out. I hope you are all well, and I'll talk to you again really soon.
Tech Stuff is an iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or wherever you listen to your favorite shows.