Thanks for tunion to text stuff. If you don't recognize my voice, my name is Ozvoloshian and I'm here because the inimitable Jonathan Strickland has passed the baton to Cara Price and myself to host Tech Stuff. The show will remain your home for all things tech, and all the old episodes will remain available in this feed. Thanks for listening. Welcome to Tech Stuff, a production of iHeart Podcasts and Kaleidoscope.
I'm Os Voloshian, and today will bring you the headlines this week, including the UK government's demand for a back door into all iPhone users encrypted accounts. Then, on today's Tech Support, a new finding that using AI might make human cognition quote atropeed and unprepared in our final segment, and look at Elon Musk's government contracts. All of that
on this Week in Tech. It's Friday, February fourteenth. Happy Valentine's another week, another lonely studio, especially painful on Valentine's Day. Cara Price, we miss you, but hosting solo means I do get to share a few extra headlines today, and producer Eliza Dennis is here to go through them with me. Eliza are you ready, absolutely, So let's start with a story that's in your backyard. Lithium ion batteries in La.
Right, Lithium ion batteries are in all kinds of tech products, so I'm positive there are just tons of abandoned batteries across LA because of the wildfires.
Yes, in fact, the most in history, and the LAist has a fascinating story on the cleanup. One of the big issues is how easy it is for damage batteries to ignite or even explode, and if they do catch fire, they may take up to ten times more water to contain, so you can imagine this is really not helpful in a wildfire situation.
So the Palisades and Eaten fires out in LA are one hundred percent contains now, thank god. But yes, this definitely is a concern for them future.
Yes, but even more immediately once the fires are over. The batteries can still explode weeks after they were initially overheated, which makes cleaning up and disposing of them pretty difficult. And also, even if they look undamaged, it turns out they can be releasing toxic gas into the air or leaching chemicals into the ground. And the EPA which is overseeing the cleanup, said these fires burn more electric vehicle batteries and home energy storage systems than any fires that
ever burnt before. According to Weiss, as of October last year, there were over four hundred and thirty thousand teslas alone in the LA area, and a number of these vehicles burn. So the scale of the cleanup is pretty huge, and
it's a delicate operation, right. The EPA workers collect the batteries and seal drums, bring them to staging areas, and if they still have charge, the batteries are put in this saltwater brine solution to decharge them, and then you have to find a recycler who can salvage the minerals from the batteries, which remain in high demand.
Yeah, this is so interesting because I just never thought that an electric vehicle would kind of change the way that we do disaster recovery. But here we are. So you have another story for.
Me, Yes, Well, I can't resist stories coming out of the UK, and I saw this headline the Washington Post and I was pretty shocked. Actually, so security officials in the UK have demanded that Apple create a backdoor that would allow authorities to access any Apple user's data uploaded to the cloud. Now, this wasn't a request to access a specific user account, which law enforcement quite often does. This was a request to have an unlimited back door.
Wow. I mean this feels like a real invasion of privacy.
Not only that the order was actually issued secretly. So thanks to the Washington Post, democracy dies in darkness. While you may think you're safe outside of the UK, that's just not true if this order goes ahead, because anyone around the world can be subject to these demands from the UK government. So if for some reason the British government want to look at your encrypted phone day and this order goes through, they would just be able to do that.
Can you tell me, like why now? iPhones have sort of been around for a while.
Yeah. So this backdoor is for a specific type of cloud storage that Apple started offering in twenty twenty two called Advanced Data Protection. Most iCloud storage can be unlocked by the user and by Apple, but with Advanced Data Protection, only the user can access their data stare in the cloud. Basically, it extends the end to end encryption of I messages and FaceTime to all data on iCloud, but it's a
setting that users have to opt into. It's not turn on my default and advanced data protection obviously offers more security protections from hacking, but also from law enforcement who often serve Apple with search warrants to access iCloud storage and backups, often without the user ever knowing.
Right, And I do believe that this is a tension that a lot of tech companies feel. There's definitely a friction between law enforcement and the data privacy of their users.
Absolutely. You probably remember back in twenty six when the FBI tried to order Apple to unlock the phone of a dead terrorist. Apple fought back and the two groups eventually came to a compromise letting the government scan the phone for illegal material. But there's a similar tension around
this story. The Washington Post, who broke the story noted that when Apple was warned about this order in March, they threatened to quote publicly withdraw critical security features from the UK market, depriving UK users of these protections.
Wow, I mean sounds like fighting words.
Yeah. And there's this brewing larger fight between Britain and various EU countries and the American tech giants, which is super interesting and something I'm sure we'll cover soon, but that's a side point on this story. The key thing here is that Apple knows, as I'm sure we all do by now, that any backdoor or security workaround is also,
by definition, available to bad actors. On a brighter note, this week, I'd like to shout out the latest web telescope images which show the most beautiful image of a new star forming in vivid definition.
Okay, yes, for our listeners, if you haven't seen any of these images from the web telescope, please stop what you're doing and go look at them. They are absolutely unbelievable.
They really are. So the Web is this space telescope that takes stunningly sharp images and it looks deeper into space than any previous telescope, which means actually looking further back in time, around thirteen point five billion years into the past.
Yeah. That will never stop blowing my mind. That looking at the space pins looking back in time. So just keep that in mind when you're looking at these images.
Exactly Well, this week's images really areline blowing. How to describe an audio that, think of a vertical collision of pain and blue translucent light, almost kaleidoscopic and what's truly unbelievable is that this gorgeous array of colours contains newly forming stars, and the image may help us understand a little more about the environment in which planets are born.
This story makes me so happy. I absolutely love the web telescope.
METO and there's actually this really beautiful documentary I saw a few months ago called Deep Sky, all about the Web Telescope. I saw it in Imax and it was an incredible experience that I really do recommend. But that's it for the headlines this week. Thanks for joining me, Eliza,
Happy to do it. Coming up on this week's Tech Support, we discussed the potential impacts of AI on our cognition stay with us, so today for our Tech Support segment, we're turning to four or four Media's Immanuel Myberg to discuss a question and a lingering fear I and I think many of us have about the rise of generative AI, which is will this make us dumber? That's definitely a reductive statement, but it's not a million miles from what a study from Microsoft recently found, which shows that AI
might affect our critical thinking skills. It's something that's been asked every time a new technology is invented. For example, when the radio came out, some people worried that reading skills would diminish. In more recent generations, it's not been uncommon to hear concerns about TV or video games or smartphones rotting our brains. But in the case of this study, that might actually be happening. So, Emmanuel, thanks so much for joining us this week.
Yeah, thank you for having me.
You wrote a story with the headline Microsoft's study finds AI makes human cognition quote atropheed and unprepared. With the subhead, researchers find that the more people use AI at their job, the less criticals thinking they may use I found this absolutely fascinating. Can you tell us a bit about the study, who conducted it, why, what it found.
So this was a study that was conducted primarily by researchers at Microsoft, who collaborated with some researchers at Carnegie Mellon University. And what they did for this study is they recruited three hundred and nineteen knowledge workers. Knowledge workers in this case refers to most people with an office job, right, most people who work on computers who generate reports data things like.
This to bankers, journalist, teachers.
Yeah, anything that, like most work that is not manual labor,
you can think of as knowledge workers, right. And they collected nine hundred and thirty six first hand examples of how these people used generative AI in their jobs and then asked them to complete a questionnaire where they asked them how they used THEAI tool, how they felt about using the AI tool, how confident they were in the AI tool's ability to give them a reliable, good output, and how confident they were in their own ability to do the job with or without the AI tool.
So, in Manuel, what were the conclusions of this study that was partly led by Microsoft.
So the bottom line here is that the researchers found that the more these knowledge workers were confident that the generative AI tool was doing the job well enough, the less likely they were to use critical thinking to make sure it's correct.
We did a story on tech Stuff a few weeks ago about how London taxi drivers have bigger hippocampuses than the London bus drivers because they saw that they had to kind of make out the route as they go along versus following a preset route. I found that super fascinating. You're smiling. How do you see this story in relation to that story.
I share the personal anecdote in the story that I
think many people can relate to. So, when I first moved out to college in San Francisco, it was a little bit before the iPhone became very popular and everyone had access to Google Maps on their phone, and I would leave my house with a little mini map of the city and all the bus routes, and doing that for a year, you learn the city, and then you just have an intuitive understanding of where you are and relations just to the rest of the city, the rest of the bus lines, and you know how to get
places without looking at your phone, which wasn't available to me at the time. And ever since then, every time I moved to a new city, I never learned the city again in the same way because I relied on the phone. Same thing with I'm old enough to remember remembering my friend's phone numbers and my family's phone numbers. Still remember some of those numbers, but ever since I had a contact list on my phone, like, never remember another phone number again.
This is one of the things I think about. Why I'm grateful that I'm an elder millennial is because My education was basically, you know, reading tons of things, memorizing facts and then regurgitating them, probably in a way which was below the standard of any LM today, well below the standard. But the process of learning to do that made me who I am today, and it's I find it quite concerning to think that so many people don't get the experience of having to train their own brain.
Yeah, it's a really interesting question that I think is maybe bigger than the study, which is what do we feel comfortable outsourcing to technology? Which parts of our brain do we feel comfortable outsourcing in this way? So do I need that muscle in my brain to remember ten phone numbers? I don't know. Maybe that gray matter is better used for another task. I feel fine about that, but I feel less fine about the fact that I'm not as good as I used to be at navigating
cities and so on. Obviously, to go back to the fact that this comes from Microsoft, Microsoft is the biggest investor in open AI. It's a giant tech company. Like all the other giant tech companies, it is at this point heavily invested in generative AI tools being widely adopted and used by people in their personal lives and their professional lives and so on.
So why are they funding and publishing research that drives a whole news cycle about how AI might be bad for us?
Yeah, the same reason that a lot of these AI companies have these self reflective studies about the negative side effects of AI tools. Right. So, I think last year I wrote a story about Google doing a study about how AI tools make it much easier to produce disinformation and spread it online. But the purpose of that study was for Google to say, we developed these tools, we
recognize this is a problem. Here's our idea for how we can mitigate that problem to put you at ease so you can feel more comfortable about buying our AI tools. And similarly, Microsoft here is recognizing that this is a problem and offering a solution so they can move forward and developing and deploying these AI tools.
And how can it be mitigated? I mean, did any of the users or the subjects that the study talk about how AI is beneficial to them in their work or what's the what's the optimistic takeaway here?
So in general, they note that there is a shift between what they call people doing task execution to oversight, and when you shift into that oversight role. That's when things can get lax and you use less critical thinking and mistakes can slip through, and that's bad. And what they suggest is developing AI tools with this problem in mind and sort of constantly encouraging the person who is using the AI tool to use critical thinking to examine
the output. Like they say, they could offer guided critiques for you know, how to approach an output from AI. Something that I thought was really interesting and we've seen maybe your listeners have heard about deep Seek, which was this popular LLLM from China, and one of the things that it does, and I think increasingly we're going to see other AI tools do, is that you presented with the problem and it doesn't just spit out the final answer.
It kind of shows you its quote unquote thought process.
Right, it's just so called reasoning models, right, yes.
Right, So it's showing you how it's getting to the conclusion. It's presenting you at the and that's another thing that the researchers here say might help make people continue to think critically about the output by showing how the AI is arriving at that conclusion. And when you look at that, maybe you'll you'll be able to note, oh wait, this is a bad process and this is a bad conclusion.
Well, it brings me back to my school days and that very frustrating phrase show your workings. Yeah, do you have any tips? I mean, it's somebody who's you know, a journalist who covers this, but also as a person in the world who spends more time thinking about this than the average person, about how we should as regular people think about finding the right balance and using AI tools to help us versus kind of corroding our own abilities on the past.
Ooh, that's a big question, I would say. I think this is easy to forget. We're kind of early in the game here about the widespread adoption of these tools, and from what I see, a lot of our stories are basically about the outputs being bad and the outputs being wrong and ways that are not immediately obvious. So today I would approach any output from any AI with a lot of skepticism. Can it be useful? For sure? Is it making people more efficient at work? For sure?
But like open aya has a new research tool, I would never just trust the output blindly. I would have to double check. Maybe it's a good jumping off point, right, I remember in school they told us it's like Wikipedia might be a good thing to look at, but it's a jumping off point. It's not the end of your research or something similar to that. As for in the future, when these tools maybe get more competent, I think then it's a question of personal choice. I enjoy the process
of writing. I'm going to continue to do it the way that I've done it because that's what I like to do, even if chat Gupte makes it instantaneous.
And Matanuel, thank you so much for joining us today.
Oh thank you.
Coming up a look at Elon Musk's history of government contracts. Stay with us. If you're even mildly tuned into the news, you can't escape the mention of the Department of Government Efficiency aka DOGE. DOGE is also a cryptocurrency issued by Musk, but it's not actually a department of government like Defense
or Treasury. It's an entity within the United States Digital Service, which has now been renamed by Executive order to United States DOGE Service, which means the head of the US DOGE Service, Elon Musk, does not need to be vetted or approved by Congress, which meant Elon had nothing to stop him from getting straight to work, and so he
and Doge have been very, very, very busy. In the twenty six days since President Trump's inauguration, Doge has been tearing through the federal government officially in the name of reform and efficiency. Government staffers across agencies are being asked to resign, programs are being put on hold, Budgets are getting slashed, sometimes with the assistance of AI, and this is of course nothing new for the man at the
center of it all, Elon Musk. In fact, it mirrors his managerial style at other companies slash first, fix later, and he's shown time and again that he'd rather cut too much and add back than be too gentle at the first pass. As of this taping, those claims have identified and cut over one billion dollars in government spending the goal two trillion dollars, which means that while it's already been a chaotic few weeks, there's no sign of
anything slowing down. This got me interested in how Musk and his fellow Silicon Valley titans first intersected with government spending and what their relationship to government efficiency really is, Because, of course, efficiency is in the eye of the beholder.
To prepare for this segment, I had an interesting conversation with David Eves, who's a professor of Digital Government at University College London, and he said to me, quote, what's disappointing is the lack of imagination for a bunch of people who benefited from various forms of industrial policy to have this simplistic shrink government is always attitude Tesla SpaceX. These organizations exist because people chose to invest. Like they saw,
they had a vision and they chose to invest. Now we'll have more of David's take later, but first let's establish and baseline points. There is, indeed lots of spending from Washington. The national debt is at about thirty two trillion dollars right now, and surely, as in any complex entity, whether a government or a massive corporation or a university,
not every dollar is getting much bang for its well buck. Still, as much as there is interesting cutting costs and eliminating debt from leaders in Congress, whether calling for cuts in defense or the National Endowment for the Arts, many elected officials and civil servants are strongly opposed to. This particular style of cost cutting is very Silicon Valley very moved fast and break things again. It's a tactic that seems to have worked for the richest man alive in his
work in the private sector, specifically in Silicon Valley. But what I find so interesting is the fact that Elon Musk and his companies have actually benefited more than a little from government spending. Some might even call it inefficiency. According to The New York Times, Musk had around one hundred contracts through seventeen different federal agencies last year alone, and within the last decade, musks companies, SpaceX and Tesla
have received at least eighteen billion dollars in government contracts. Now, tens of billions of dollars might sound like a pittance next to Elon's estimated net worth, which fluctuates, but last time I checked was around three hundred and eighty billion dollars. But those government contracts have driven up the share price his companies and personally enriched him, and the timing and the nature of the government funding given to SpaceX and
Tesla has at times been crucial to their success. Let's start with Tesla. Back in twenty ten, the company was just starting out and had only sold a couple of thousand cars two years before. They'd nearly gone bankrupt, but then the Department of Energy stepped in and gave Tesla a four hundred and sixty five million dollar loan, which the company used to develop a new car model. Just three years after this loan, in twenty fourteen, with the model s on the market, the carmaker had sold just
over thirty one thousand cars. With that success, they were able to pay back their loan to the Department of Energy early. But some have pointed out that a private sector investor would have demanded more than just a repayment of the loan. They would have demanded a stake in Tesla itself, which today will be worth many, many multiples of what it was then. So why doesn't the taxpayer
see that benefit when they're taking the risk. There's this concept in investing called the efficient frontier, which basically tells investors how much risk is acceptable in proportion to a potential reward. Per my favorite source investor Pedia quote, portfolios that lie below the efficient frontier are suboptimal because they do not provide enough return for the level of risk.
So if you're a US taxpayer, it kind of sucks that you take on all the risk of bailing out a company but get barely any return if it survives. Remember the Auto bailout anyway. Since then, Tesla has been awarded contracts for things like building solar panels to the government, providing tactical vehicles for US embassies, money for things the
government deemed necessary. Then there's also the ev tax credit, which existed during the early years of Tesla until twenty nineteen, and then was reinstated during the Biden administration seventy five hundred dollars to Americans who purchased an electric vehicle. No doubt, this encouraged sales and allowed Tesla to set a slightly higher price than the market may have allowed before the
tax credit. As for SpaceX, well, Musk's rocket manufacturer has gotten a lot of discretionary government funding, in particular from NASA and the Department of Defense. Musk's contracts with those agencies total up to about fifteen billion dollars in the last decade, for things like designing space landing systems and government use of Starlink, which is a satellite network that
helps people in remote areas connect to the Internet. David Eves, the professor, pointed out to me that a major turning point for SpaceX was the Russian incursion into Ukraine in twenty fourteen. Up until that point, the US had relied on Russian made rocket engines to launch stuff into space, but after the conflict in Ukraine, the US government deemed this a threat to national security and created the National Security Space Launch Program, which gave contracts to US companies
to develop rocket technology. SpaceX won one of these contracts. As David pointed out, this was not a government decision to maximize efficiency. The Russian rockets were cheap and plentiful, but it arguably benefited the United States and certainly benefited SpaceX. More money came must sway during the war in Ukraine
through Starlink, the satellite network. By the fall of twenty twenty two, Ukraine had received more than twenty thousand Starlink terminals, which aided in restoring military and civilian communications in the country. Some of these terminals were donated by Starlink, but some were purchased by USAID with government funds. That's USA, the
agency whose operations have essentially been halted by DOGE. In twenty twenty one, SpaceX was awarded a multi billion dollar contract to help land the first woman and first person of color on the Moon as part of NASA's Artemis programme. While Mask's DOGE is cutting DEI efforts across federal agencies, SpaceX is still building a spacecraft for long term exploration of the Moon, and for SpaceX, the contracts keep on rolling.
Last year, the company was selected to build a vehicle that would take the International Space Station safely out of orbit in twenty thirty due to its outdated technology and capabilities. But in many ways, this is a story far bigger than just Elon. As David Eves told me, quote, one cannot underestimate the importance of the US federal government in
the creation of Silicon Valley. For example, you wouldn't have the smartphone without the touchscreen, which was developed by researchers at the University of Delaware with funding from the CIA and the National Science Foundation, not to mention the countless innovations through DARPER, the Research and Development Agency of the
Department of Defense. I mean, I think we all know this, but we wouldn't even have the Internet without DARPER, which funded the development of upernet, the Advanced Research Projects Agency network. Back in the sixties, and this network was declared operational in nineteen seventy one, and soon there was an early form of email through advances in remote login and file transfer.
We wouldn't have GPS or autonomous vehicles without Darper or digital virtual assistance like Siri, which were originally meant to provide assistance to soldiers in the field. Back in the sixties, one million dollars was provided to start building Arpernet, which with inflation, would be over nine million dollars today. The wealth creation that money has spurred, particularly in Silicon Valley,
has been completely insane. It just bears remembering many of the inventions that are at the heart of so much tech we use today were funded because someone in the federal government said, here, have some money, this is important, figure it out. And those priorities, of course change and evolve from administration to administration. But one of the frustrating things about watching Doge hack away agency budgets is that there's very little transparency surrounding the tactics or motivations of
Elon Musk and the Department of Government Efficiency. Most of all, I'm actually craving a definition of the word efficiency. A final thought from David quote. The government's main role is an aggregator of capital so it can make investments. And if the public is robbed of its confidence in the public sector's ability to make investments that generate public good outcomes, then we've been robbed of a critical tool to address
the key problems of our time. Food for thought. But that's all we have time for this week on tech Stuff. For tech Stuff, I'm os Vloshin. This episode was produced by Eliza Dennis, Victoria de Mingez, and Lizzie Jacobs. It was executive produced by me Kara Price and Kate Osborne for Kaleidoscope and Katrina Norvel b iheartpodt but he'd Fraser is our engineer and Kyle Murdoch mixed this episode and
also wrote the theme song. Join us next Wednesday for tex Stuff The Story, when we'll share an in depth conversation with Harney Freed, the man the world turns to when they aren't quite sure is this image real or not? Please rate, review, and reach out to us at tech Stuff Podcast at gmail dot com. We love hearing from you.