This week, we’re joined by tech critic Paris Marx to discuss Data Vampires, his latest Tech Won’t Save us podcast series. We chat about how data centers suck up vast amounts of power, water and other resources, and why the AI boom is exacerbating those issues. Also, Devindra and Ben dive into a few news stories, including the DOJ inching closer towards a Google antitrust breakup; Nintendo's adorable motion sensing alarm clock, Alarmo; and why Google's Deepmind AI head won the Nobel Prize for chemistry.
Listen below or subscribe on your podcast app of choice. If you've got suggestions or topics you'd like covered on the show, be sure to email us or drop a note in the comments! And be sure to check out our other podcast, Engadget News!
Subscribe!
Topics
Interview with Tech Won’t Save Us host Paris Marx on his new series, Data Vampires – 2:09
U.S. regulators continue to float the possibility of breaking Google up in antitrust ruling – 25:54
Nintendo announces new hardware…Alarmo, a motion sensing alarm clock – 39:33
Apple Intelligence likely arrives October 28 – 42:27
343 Industries rebrands as Halo Studios and shows off Unreal Engine 5 demo – 44:46
Pop culture picks – 50:36
Livestream
Credits
Hosts: Devindra Hardawar
Guest: Paris Marx
Producer: Ben Ellman
Music: Dale North and Terrence O'Brien
Transcript
(Produced together with Descript's AI transcription.)
Devindra: What's up, Internet? Welcome back to the Engadget Podcast. I'm senior editor Devindra Hardawar. This week I'm joined by podcast producer Ben Ellman. Hey, Ben.
Ben: Hello. Let's talk about Nintendo. And also Google. Google more important.
Devindra: And also all sorts of things. Let's talk about data centers. We've got special guest on Paris Marx, the author, podcast host, and tech critic.
He has a new series at his podcast, Tech Won't Save Us, about data center vampires. So, um You know, we will talk a bit about that. But first folks, if you're enjoying the show, please subscribe to us on iTunes or your podcatcher of choice, leave us a review on iTunes and drop us an email at podcastinggadget.
com. You can also join us Thursday mornings around 10 45 AM Eastern on our YouTube channel for our live stream. This week we did a fun Q and a, which is actually not in this not in the recorded episode at all. So if you want to join us for that fun or go check that out, go take a look at our YouTube channel.
All right. So. I sat down with Paris Marx, who I think has been doing great work over at Tech Won't Save Us, which is a much more, it's a critical look at the tech industry, and Paris has the time and energy to really focus on what the industry is doing wrong. His most recent series, Data Vampires, is pretty much all about data centers and the, The impacts they have on our environment, the resources they use when it comes to power, which is obscene.
They're requiring more and more power from our grid, which is already kind of a mess. You've probably seen the news. We've talked about this too, I think, about Microsoft re upping Three Mile Island, like turning it back on just to power AI data centers. A lot of other companies are thinking about this too.
Water is a big thing. There was a story a couple of years ago about Google essentially hiding the metric crap ton of water they were using from a town in Oregon. And because they didn't want people to know like how much it took to cool those data centers down and things like that. So anyway, Paris and I.
Had a really good chat about this series. So, take a listen, and I'm sure you'll learn a thing or two about data centers and cloud computing. Paris Marx, thank you so much for joining us on the Engadget podcast.
Paris: Absolutely, great to join you.
Devindra: Can you tell us briefly, what are you trying to cover with Data Vampires, and why you're specifically focusing on data centers right now?
Paris: Yeah, it's a really good question, right? And I feel like data centers have gotten more in the public's consciousness through the generative AI moment, but also to a certain degree crypto as well, right? Remember when we were talking about how much energy use crypto was having and, you know, the impacts of these major miners as they were setting up in places around the world and the concerns about them keeping like fossil fuel energy online or even reviving fossil fuel plants.
We've seen a lot of those similar concerns with generative AI. But the thing that really stood out to me is that in certain places where a lot of data centers have been being built for some time, places like Northern Virginia or Ireland, for example, we were seeing these concerns in the communities for some years now, pre pandemic and kind of well before, right?
But what we've seen in the past few years is that as the number of these, especially hyperscale data centers that these major cloud companies like Amazon, Microsoft and Google have been building around the world, have You know, accelerated. What we've seen is not only those issues in, say, Northern Virginia and Ireland get more acute, but that in more and more communities around the world where these things are being built, we're seeing similar concerns and similar opposition.
And so it felt like something to really want to tap in and pay attention to. So the series, you know, looks at why. We're building all these, you know, hyper scale data centers. Looks at some of that community opposition in different parts of the world. You know, the growing kind of climate impacts of something like this and the broader potential harms of generative AI and the types of things that these data centers are powering.
And then, of course, looks at the broader ideology behind all this, that these tech billionaires are trying to push.
Devindra: This is actually really good timing Paris, because I forget if it was during our live stream or a recent podcast episode our listeners were asking the questions about AWS, you know, where did AWS come from kind of, how did we get here?
And I had to like, just pull back from what I remember from reporting over the last few years, but I'm very glad you guys covered that in the first episode because I feel like that sets the stage for. Kind of where we are, right? Like AWS, an offshoot of Amazon trying to figure out its own infrastructure, but basically coming up with the idea that they could rent servers time and server space to two other companies rather than those companies building their own server infrastructure.
Can you talk a bit about that and like how that plays out? basically helped get us here.
Paris: Yeah, definitely. Cause it's such a key moment, right? Not just for what I'm talking about with the series, but for so much of how, you know, digital technology and the internet have developed in the years since, right?
Because so many of these you know, online services and things are built on the cloud now and the cloud really comes out of Amazon web services. So, you know, you go back to the early two thousands and you're starting to have these ideas percolate within Amazon itself, the company, right? Because they're trying to make their processes more efficient, you know, because they're, they are an e commerce company, you know, generally a quite low margin business.
And so they want to do things like as efficiently as possible is, you know, how the story is told. Right. And so, you know, at a certain point these particular people at the company, Chris Pinkham is one of them starts to develop this proposal for something that would, you know, basically create this web service for Amazon itself, you know, by so that all these different teams working on these different projects wouldn't have to spin up their own servers or, you know, figure out their own kind of web services and, and whatnot.
And then what happens then is they say, okay, this, you know, Isn't just something that would be useful inside of Amazon, but it's something that we can then sell to other companies. And I think even in that moment, so this is around you know, 2004, 2005, they're, they're really working on spinning this up.
And Chris Pinkham gets permission to go back to South Africa, where he's from to put a team together to work on this. Cause he wanted to go, you know, back to the country where he came from. And Amazon wanted to keep him. So they said, you go back there, you work on this, you know, you kind of figure it out.
Right. And, and then. You know, they kind of come up with this proposal. I think, you know, they have an idea that this is probably something that's going to be useful. But then there's this interview I found with Jeff Bezos in 2008 where, oh, Malik, you know, the, the tech journalist was asking him about, you know, Whether he was expecting like cloud startups to, to, you know, be built on AWS.
And even at that time, he was like, the venture capitalists are all pushing them to, and we'll serve them if, you know, that's going to be helpful. So even then it's you know, it's still quite nascent. You know, if you think back now, like it's so hard to, to, you know, think of a time, like pre smartphone, like that's even you know, the iPhone is just kind of like getting launched, you know, this is like early days for the transformations that are going to come with like Web 2. 0 and the mobile internet and all this sort of stuff. So it's really like the smartphone and cloud computing that really set the stage for what happens through the 2010s. I think,
Devindra: yeah, this is the birth of cloud computing as we know it.
Because yeah, before companies, if they, they wanted data storage or other, other sort of processes, they had to set up their own servers, which they did. But. I feel like for the likes of like maybe Dell or HP, there's less of a reason for them to do that when they could just get space from Amazon, especially if like usage is is like volatile, like you don't know how much people are going to need.
So you don't want to build out too much hardware. I want to say I've been reporting on startups since 2009, 2010, and like clearly the cloud. Computing element is a big reason why so many of these startups were able to become a thing, right? They didn't have to worry about infrastructure.
They could just have an idea. Instagram didn't need to build a ton to data centers to store photos. They could just get some Amazon time, just a couple of dudes to build a photo filter app and then get bought for a billion dollars. So it all kind of led to that. And Amazon is not the only one we've talked about Microsoft and Azure and everything.
And Azure is doing gangbusters for Microsoft. Like when I write up their earnings, it really is just, yeah, Azure money just keeps coming in and it doesn't look like it's slowing down anytime soon. It's really funny to draw that line, I guess, because I also remember Paris, like when I was doing it work in college, right.
That was like, Oh, one to Oh five. I remember like our email service was an exchange server on site. That is how people used to do computing. Then 05 hit and Gmail came about, right? And then, oh, all of a sudden, viable web email became a thing. And I don't think I've configured an email client since then. So we also shifted a lot of our computing to the cloud just as users.
I guess it makes sense that the companies did that too.
Paris: Yeah, absolutely. Like it was so much more convenient to do that. Especially as these companies made it the convenient thing to do right to try to incentivize that to happen. But even with the companies, as you're saying you know, I talked to Dwayne Monroe, who's a cloud technologist, who's been doing this for 20 years, over 20 years, you know, for the for the series, and he was kind of like giving me these different examples of like, why different companies move to the cloud and things like that.
And in some cases, he was like, you know, the company itself. Was resistant to moving to the cloud. You know, this is like kind of back in the early days, say late 2000s, early 2010s because they didn't want to be dependent on you know, Amazon's infrastructure or one of these major tech companies infrastructure,
Devindra: but
Paris: they but they also didn't want to spend the money on like the capital expenses.
to build out the server infrastructure that they actually needed. You know, as the demand for their website and stuff was growing, right? Like he talked about a book business in particular that was seeing like orders be lost at, at peak times because you know, there was so much demand on, on the servers there.
And so the people at the company itself, like spun up this cloud solution, taking advantage of Amazon web services and then presented it to like management later and was like, look, this works. This is solving our problem. Either we do this or you give us the money for the servers and management was just like, okay, I guess we're going to, you know, use this solution.
And it's one way that these companies got onto it. But then of course the other way was that Amazon and Microsoft and Google all told these companies that if you come onto the cloud, it will be cheaper. You know, you'll save money by not having. You know, so much of your own server infrastructure.
And, you know, that was partly the case. But it has certainly become more expensive over the years as they've sold them more you know, AI tools and all that kind of stuff that you get by being on one of these cloud services. But, you know, I think for a lot of these big companies, there's still plenty of reason to be on you know, one of these cloud providers because of the benefits that it provides and because then they don't need to keep up with their own you know, Microsoft Mechanics infrastructural servers and all the costs and concerns that come along with that.
And they can focus on the things that are much more core to their businesses.
Devindra: It's kind of, it's kind of like, it's a very smart, smart business strategy, right? Like you're telling people, we'll make you, we'll do this cheaper than if you did on your own, but also it makes you dependent on them. And then it's a closed ecosystem.
So you continue to use other products that they have. And that's kind of Microsoft's whole deal with Azure too. Like they're tying co pilot and everything into all of that. So it is kind of a vicious circle of I dunno, of just money and a commitment to these specific companies.
Paris: I think that's an important point you make though, right?
Because earlier you were, you were talking about how, you know, in the early days, all these startups were founded on the cloud and how you know, without the cloud being there, without Amazon web services and Google cloud and Amazon Azure. You know, being these options for these companies, it would have been much more difficult for this kind of startup boom in the post recession times to have really happened.
Right. As we were seeing all this excitement in like the early 2010s about all these companies coming out of the tech industry and whatnot. And so that's one element of that. And then you fast forward to the past few years and Without that massive centralized computational infrastructure that Amazon, Microsoft and Google have built up, it would have been very difficult to see this kind of generative AI boom and generative AI hype that we have, you know, been experiencing for the past year and a half or so or almost two years now, really because, you know, These massive models that use that require so much data and so much computation to train and to use are basically not possible without the centralized infrastructures that these companies have built
Autentifică-te pentru a adăuga comentarii
