Latest / KDAV Radio / Better Offline - The Truth About Data Centers w/ Paris Marx

Transcript

Better Offline - The Truth About Data Centers w/ Paris Marx · KDAV Radio

0:01
Speaker 1: Call zone media.
0:04
Speaker 2: You there, boy. What podcast is it? What podcast is it? Sir,
0:08
Speaker 2: it's Better Off Live.
0:17
Speaker 1: Better Off Live.
0:21
Speaker 2: That's right, folks. We're back. I'm Ed Zitron. This is
0:23
Speaker 2: Better Offline. Today, we're talking about a subject near and
0:26
Speaker 2: dear to my heart. That's, of course, those goddamn data centers.
0:29
Speaker 2: And I brought on an expert who has a wonderful
0:31
Speaker 2: book coming out, Paris Marks, host of Tech Won't Save
0:34
Speaker 2: Us and the author of the newly released Hyperscale, The
0:36
Speaker 2: Ambition and Excess of Big Tech's Data Empires. Paris, thank
0:39
Speaker 2: you for joining me.
0:40
Speaker 1: Absolutely. Great to be back on the show and to
0:42
Speaker 1: chat with you about all this stuff.
0:44
Speaker 3: All right.
0:44
Speaker 2: So walk me through the book because there've been a
0:47
Speaker 2: few people covering this, but what approach did you take
0:49
Speaker 2: to these things? Because I know we've had data centers
0:53
Speaker 2: for a while, they haven't been quite as offensive, but
0:56
Speaker 2: where do you start the story?
0:58
Speaker 1: Absolutely.
0:59
Speaker 3: So I really start by looking at one of the
1:02
Speaker 3: stories that really brought me into it. People might remember
1:05
Speaker 3: a few years ago, there was this case in the
1:07
Speaker 3: Dalles in Oregon where This was where Google built its
1:13
Speaker 3: first company-owned data centers. As it was building them out,
1:18
Speaker 3: as its search engine was growing, as it was launching
1:20
Speaker 3: new services, it realized that it needed more computation, more
1:24
Speaker 3: than just it could lease from other people. It was like, oh,
1:27
Speaker 3: we need to start building data centers. This is 15,
1:29
Speaker 3: 20 years ago. The DAOs is where.
1:33
Speaker 1: They end up going.
1:34
Speaker 3: And so by the time that it comes to 2021, 2022,
1:39
Speaker 3: Google is looking to expand, but they need a new
1:42
Speaker 3: water deal with the city, right?
1:44
Speaker 1: Right.
1:44
Speaker 2: And just water deal, what do you mean by that?
1:47
Speaker 3: Yeah, so a deal on how much water that they
1:49
Speaker 3: can actually consume, right? That they can take out of
1:52
Speaker 3: the public water system. And so people in the city
1:55
Speaker 3: are like, you know, okay, Google's been here a while. Like,
1:57
Speaker 3: you know, they're fine as like a, you know, corporate
2:00
Speaker 3: tenant or, you know, a corporation that's in the city.
2:02
Speaker 3: You know, they're not like amazing, like the industries that
2:05
Speaker 3: used to be there, the kind of industrial industries that
2:08
Speaker 3: left over time. But they're like, how much water is
2:11
Speaker 3: Google actually using? Like, you know, we think this is
2:13
Speaker 3: an important, you know, kind of figure for us to
2:16
Speaker 3: understand in order to talk about giving it more water.
2:20
Speaker 3: And surprisingly, Google will not even not share those figures
2:25
Speaker 3: But when a journalist at the Oregonian, you know, kind
2:28
Speaker 3: of statewide media.
2:31
Speaker 1: Yeah, exactly. Goes to ask about it.
2:34
Speaker 3: Google not only will not provide it, but actually gets
2:39
Speaker 3: the city council to sue the journalist to make sure
2:44
Speaker 3: that the information won't come out. And this case kind
2:47
Speaker 3: of proceeds for a whole year before the public relations
2:50
Speaker 3: backlash is essentially so bad. that Google relents and eventually
2:54
Speaker 3: shares that information and basically has to let the city
2:57
Speaker 3: know and the residents of the city that they are
3:00
Speaker 3: using 29% of the city's water, which for many people
3:05
Speaker 3: is staggering.
3:06
Speaker 1: Exactly, right?
3:07
Speaker 2: And this is the thing, though. I know that I
3:10
Speaker 2: don't want to go into it because these people are weird,
3:13
Speaker 2: but there was the whole pushback on Karen Howe because
3:15
Speaker 2: she made a mathematical error over water, but people are
3:17
Speaker 2: just pretending that water isn't involved in data centers now,
3:20
Speaker 2: which is bullshit.
3:21
Speaker 1: Oh, it's wild, right?
3:22
Speaker 3: And so there's a few different tacks that they take
3:24
Speaker 3: on this, right? And I feel like to me, when
3:26
Speaker 3: I see what they're doing, it's like the data center lobby,
3:29
Speaker 3: the data center supporters, whatever, they see that there are
3:33
Speaker 3: real challenges to public acceptance with these massive data centers
3:37
Speaker 3: that they're trying to build out across the United States
3:39
Speaker 3: and the world, right? And so they can see that
3:42
Speaker 3: they can't really make the argument on the kind of
3:46
Speaker 3: plane of electricity use and climate impacts because they're just
3:50
Speaker 3: using so many fossil fuels.
3:51
Speaker 1: It's like ridiculous.
3:53
Speaker 3: But then they see water and they're like, oh, I
3:54
Speaker 3: think that there's an opportunity here to muddy the public
3:57
Speaker 3: perception on what is going on, right? To say like, oh, look,
4:00
Speaker 3: golf courses and agriculture use so much more water, so
4:04
Speaker 3: we shouldn't be worried about the water that data centers
4:06
Speaker 3: are using. You know, when we're looking at kind of
4:09
Speaker 3: like abstract national level figures in the United States. But then,
4:13
Speaker 3: of course, when you zoom in on the local level,
4:16
Speaker 3: you can see that this can have some real major impacts,
4:18
Speaker 3: especially because there was some Bloomberg reporting, I think it
4:21
Speaker 3: was last year now, that there are a lot of
4:22
Speaker 3: these data centers basically being built in areas that are
4:26
Speaker 3: high water stress. They don't have a lot of water
4:28
Speaker 3: to provide to these other industries. And so I feel
4:32
Speaker 3: like they kind of try to muddy the waters, try
4:35
Speaker 3: to confuse people in that way. But then the thing
4:37
Speaker 3: that we have also seen shift in, I would say,
4:40
Speaker 3: the past six to eight months is the industry recognizing
4:43
Speaker 3: that this is a problem, Instead of basically planning data
4:47
Speaker 3: centers with evaporative cooling, which was the way that they
4:50
Speaker 3: traditionally did it in order to save money, so that's
4:52
Speaker 3: kind of like you have a constant draw of water
4:55
Speaker 3: into the data center. That water kind of flows through
4:58
Speaker 3: the data center and kind of takes the heat away
5:00
Speaker 3: from the computer hardware, the chips, the graphics cards, things
5:03
Speaker 3: like that, until it gets warmed to such a degree
5:06
Speaker 3: that that that water, for the most part, ends up
5:09
Speaker 3: evaporating away. And there's a bit of non-potable water that
5:12
Speaker 3: is ejected after that. But pretty much you need a
5:15
Speaker 3: lot of water coming into the data center to make
5:18
Speaker 3: that cooling system work. What they're doing now is they're
5:21
Speaker 3: switching to what's called closed-loop cooling, which I'm sure you've
5:24
Speaker 3: talked about on the show before. But it's essentially like, okay,
5:27
Speaker 3: we're going to take in a certain amount of water,
5:29
Speaker 3: we're going to keep that amount of water in the
5:30
Speaker 3: data center, it's going to flow through, and we're going
5:33
Speaker 3: to keep cooling it down as it goes through the
5:36
Speaker 3: data center. um obviously that's a system that so in
5:40
Speaker 3: some ways right you're obviously using less water by doing
5:42
Speaker 3: that um but you're using more electricity um you know
5:46
Speaker 3: because you need to cool the water down and so
5:48
Speaker 3: you know then there's the increased cost of the electricity
5:51
Speaker 3: you're probably burning more fossil fuels in order to do
5:53
Speaker 3: that given how we know these data centers are often
5:55
Speaker 3: powered there's also the water that goes into generating those
5:58
Speaker 3: fossil fuels but the other piece that they don't talk
6:01
Speaker 3: about and that we're seeing a growing focus on right now,
6:04
Speaker 3: but I feel like it's not fully broken out into
6:06
Speaker 3: the mainstream discussion on data centers, is that when you're
6:09
Speaker 3: doing this form of water-based cooling, where you need to
6:14
Speaker 3: keep cooling this water back down, there are also a
6:16
Speaker 3: lot of coolants, PFAS, chemicals involved in that process.
6:20
Speaker 1: So it's not just water that's going through the data center.
6:22
Speaker 2: So you're trading one shitty thing for another.
6:25
Speaker 1: Exactly.
6:25
Speaker 3: And we're already seeing concerns about those chemicals kind of
6:28
Speaker 3: leaking into water systems and things like that. So this is,
6:31
Speaker 3: I think, a story that's still developing and that we're
6:33
Speaker 3: going to see become part of the data center conversation
6:35
Speaker 3: a lot more. in kind of the coming six to
6:38
Speaker 3: 12 months is my, this.
6:39
Speaker 2: Is the thing. This is what pisses me off with
6:41
Speaker 2: these data center centrists. I call them the masses of
6:44
Speaker 2: the world. The guys who are like every day, like,
6:46
Speaker 2: I imagine they're once again, repeating the current how claims
6:51
Speaker 2: as if they sent us, they fixed all the problems
6:53
Speaker 2: with them as if every, by the way, there are
6:56
Speaker 2: still plenty of evaporative cooled.
6:59
Speaker 1: Oh, a hundred percent.
7:00
Speaker 2: Tons of them.
7:01
Speaker 3: Like they've not, cause they're only really like starting to
7:03
Speaker 3: switch over to closed loop now.
7:04
Speaker 1: Right? Like, And tons of the ones that are using
7:07
Speaker 1: it are still on evaporative.
7:08
Speaker 2: So with the book, what do you go through? Like,
7:11
Speaker 2: do you go through the history of data centers? Do you, like,
7:14
Speaker 2: where do you start the story though? Is it before
7:17
Speaker 2: the AI bubble, middle of it? Like, and where do
7:20
Speaker 2: you take the story? Yeah, it's kind of like right
7:23
Speaker 2: at the beginning of it, right? Because, you know, I
7:25
Speaker 2: was telling you about that story in Oregon, which is
7:27
Speaker 2: right around 2022, right? And of course, ChatGPT is November
7:31
Speaker 2: of 2022, it comes out. And so this whole story
7:33
Speaker 2: is kind of, playing out right before generative AI and
7:37
Speaker 2: kind of finalizes as generative AI is just starting to
7:40
Speaker 2: take off. And for me, it was an indication that like, okay,
7:44
Speaker 2: data centers were already an issue in certain communities, in
7:48
Speaker 2: areas where they were already building a lot, in areas
7:50
Speaker 2: that maybe didn't have the water or the electricity to
7:53
Speaker 2: provide to these massive infrastructures. But with generative AI, especially
7:57
Speaker 2: as we move into 23 and 24, we start to
7:59
Speaker 2: see that this is going to require a lot more
8:02
Speaker 2: data centers, which means a lot more communities are going
8:05
Speaker 2: to basically be facing up against what comes with having
8:08
Speaker 2: one of these hyperscale data centers located within your community
8:12
Speaker 2: or a cluster of data centers there. And that means
8:15
Speaker 2: that more communities are going to be facing, say, what
8:17
Speaker 2: the Dalles, Oregon faced, or what Santiago, Chile faced, or
8:21
Speaker 2: what Dublin and Ireland faced, just to pull out some.
8:23
Speaker 1: Of the examples.
8:25
Speaker 2: What did they face?
8:26
Speaker 1: Well, that's it, right?
8:27
Speaker 3: The Dalles, it was the concerns about water. In Santiago, Chile,
8:31
Speaker 3: it was also concerns about water. You know, say Google
8:33
Speaker 3: was planning this data center in Cerrios in like a
8:36
Speaker 3: working class neighborhood. And people there were really concerned about
8:40
Speaker 3: their continued access to running water because Google wanted to
8:43
Speaker 3: essentially take so much water from the system, right? It
8:46
Speaker 3: wouldn't provide assurances that it wouldn't actually reduce their access
8:50
Speaker 3: to water. And then over in Ireland, it's something else
8:52
Speaker 3: where it's not so much about water, though there has
8:55
Speaker 3: been a bit of a concern over there.
8:57
Speaker 1: It's been much more about electricity, right?
8:58
Speaker 3: To the point now, that more than 20% of the
9:01
Speaker 3: electricity in Ireland goes to data centers. And that has
9:04
Speaker 3: impacted grid reliability. It has pushed up electricity prices for people.
9:09
Speaker 3: Stories that we have been seeing play out in the
9:10
Speaker 3: United States in the past couple of years, but Ireland
9:13
Speaker 3: was experiencing earlier, right? And so for me, these were
9:16
Speaker 3: like indications of what can happen when there are a
9:19
Speaker 3: lot of data centers proposed for your region or area
9:23
Speaker 3: or community. And as generative AI was taking off, I
9:26
Speaker 3: was like, wait, so I can already see what's happening
9:28
Speaker 3: in these places. that's probably going to be coming to
9:31
Speaker 3: a lot more places as these companies are trying to
9:34
Speaker 3: build out so many more of these infrastructures, right?
9:37
Speaker 2: Did you go and visit any data centers if you
9:40
Speaker 2: kind of checked in on any of them?
9:42
Speaker 1: Absolutely.
9:43
Speaker 3: So, you know, I was outside of data centers in
9:45
Speaker 3: more places that I was inside, you know?
9:47
Speaker 2: Yeah, I mean, I didn't expect you to get inside. Exactly.
9:51
Speaker 2: I wouldn't let you inside either. Like, I mean that
9:53
Speaker 2: with love.
9:54
Speaker 3: Yeah, the major tech companies certainly didn't want me inside.
9:58
Speaker 3: I did ask a number of them. Microsoft was the
10:01
Speaker 3: only one that directly rejected me. The rest just didn't
10:04
Speaker 3: bother responding. But yeah, so I went to Northern Virginia,
10:09
Speaker 3: which is actually the site in the world that has
10:11
Speaker 3: the most data centers, basically stuffed into one region. And
10:15
Speaker 3: it was just wild driving through this part of the
10:18
Speaker 3: state that has been essentially dubbed data center alley. And
10:22
Speaker 3: just seeing like the degree of high voltage transmission lines
10:26
Speaker 3: and just the number of data centers. And also just
10:28
Speaker 3: like the thing that really stood out to me was
10:31
Speaker 3: how close they often are to residential communities. Like, you know,
10:35
Speaker 3: you could be kind of driving along like a freeway
10:38
Speaker 3: or something, or even just like, you know, kind of
10:41
Speaker 3: a more built up road, you know, with a number
10:43
Speaker 3: of lanes or whatnot. And on one side is this
10:45
Speaker 3: massive data center. And on the other side is this
10:47
Speaker 3: kind of you know, kind of suburban community. And what
10:51
Speaker 3: you often notice is that you drive into these communities
10:54
Speaker 3: and they have these kind of like walls of trial
10:56
Speaker 3: trees surrounding them so that they're not, you know, kind
11:00
Speaker 3: of seeing the data center that's right next door, though
11:02
Speaker 3: they're probably kind of experiencing some of the impacts that
11:05
Speaker 3: comes with that, you know, hearing the noise pollution, that
11:07
Speaker 3: kind of stuff. Yeah, it was just wild to see
11:10
Speaker 3: how close they were and also just how, you know,
11:14
Speaker 3: how large these data centers were as well, right? So
11:17
Speaker 3: I did that in Northern Virginia, in Ireland, and places
11:20
Speaker 3: like that. I actually went down to New Zealand because
11:23
Speaker 3: I knew some people down there who helped me get
11:25
Speaker 3: inside of a data center.
11:28
Speaker 1: I visited a couple down there.
11:31
Speaker 3: They certainly weren't the size of what a data center was.
11:35
Speaker 2: Were they even GPU compute? Were they just regular?
11:38
Speaker 3: There were some GPUs in them, but it was like
11:39
Speaker 3: a mix. So there were some GPU and water-based cooling,
11:44
Speaker 3: but then Most of it was your traditional hardware and
11:48
Speaker 3: air conditioning and that.
11:49
Speaker 1: Sort of a thing.
11:50
Speaker 3: I'm not even sure I would dub it a hyperscale
11:54
Speaker 3: data center. It was large, but it wasn't on the
11:57
Speaker 3: scale of what an Amazon, Microsoft, or Google or one
11:59
Speaker 3: of these major data.
12:00
Speaker 1: Center companies is building.
12:01
Speaker 3: But even then, the thing that struck me was, obviously
12:05
Speaker 3: I expected to go into the server room to see
12:09
Speaker 3: what it was like to have all these servers set up.
12:12
Speaker 3: Maybe people would be surprised to learn, or maybe not.
12:15
Speaker 3: that it's quite warm inside of those rooms. Obviously the
12:18
Speaker 3: computers are throwing off a lot of heat, but they
12:20
Speaker 3: don't cool the whole room because that would be really costly.
12:23
Speaker 3: They try to isolate the cooling to the server racks
12:26
Speaker 3: themselves so that you're not wasting all this cooling and
12:30
Speaker 3: paying all this additional money that you don't need to pay.
12:32
Speaker 3: But the thing that really stood out to me was
12:34
Speaker 3: the level of security. I needed to send my passport
12:38
Speaker 3: in advance so that they knew I was coming, and
12:40
Speaker 3: when I got there they needed to check my passport
12:42
Speaker 3: to make sure I was the same person. And then
12:44
Speaker 3: as we went through the data center to go through
12:46
Speaker 3: every single door, it needed to be like, you know,
12:49
Speaker 3: kind of the guy bringing me through needed to scan
12:51
Speaker 3: his card to unlock every single door.
12:55
Speaker 1: You know, like you could not get.
12:56
Speaker 2: Through a door without unlocking it first. And this was
12:57
Speaker 2: my experience of going into one like 15 years ago.
13:01
Speaker 2: I've only been inside one, yeah, I think it was
13:03
Speaker 2: about 15, maybe more actually years ago. And every few
13:07
Speaker 2: doors you had to, and any time you went from
13:10
Speaker 2: a room of service into another room, you had to
13:12
Speaker 2: beep through. Yeah. So it, the fact it must've got
13:16
Speaker 2: even crazier.
13:17
Speaker 1: I'm sure. I'm sure.
13:18
Speaker 3: And, and again, you know, this was a data center
13:20
Speaker 3: that was not, you know, built brand new. It was
13:23
Speaker 3: probably 10 to 15 years old though. Obviously it had
13:25
Speaker 3: been kind of updated over time and whatnot, uh, because
13:28
Speaker 3: a different company had taken it over from the one
13:29
Speaker 3: that built it initially.
13:31
Speaker 1: Um, but yeah, that was the thing.
13:33
Speaker 3: Like, obviously, you know, I expected it to be securitized,
13:37
Speaker 3: you know, I expected there to be a high degree
13:38
Speaker 3: of security. Um, But the fact that, like, every single
13:42
Speaker 3: door in the place was, you know, kind of.
13:46
Speaker 1: Locked, unless you unlocked it, was... Yeah, that one was
13:50
Speaker 1: a bit surprising to me. So.
14:04
Speaker 2: How did things shift upwards? I know that they are bigger,
14:07
Speaker 2: but did you notice anything shift when the real data
14:11
Speaker 2: center party started with regards to AI and the big
14:15
Speaker 2: boy GPUs? Because Just for the listeners, there is a
14:18
Speaker 2: big delineation between a data center for streaming media, which
14:24
Speaker 2: would be very CPU heavy, very storage heavy, a lot
14:28
Speaker 2: more cabling to an extent. No, actually, I'm sorry, the
14:31
Speaker 2: networking would actually be more in an AIGP, but nevertheless,
14:34
Speaker 2: they are materially different. Did you notice something different in
14:38
Speaker 2: the ramp up? Because you've been working on this a while.
14:40
Speaker 2: You could be working on this before most people were
14:42
Speaker 2: really obsessing over these things.
14:44
Speaker 1: Absolutely.
14:45
Speaker 3: And there were also people who were working on it
14:47
Speaker 3: long before me, right? Just to make sure that they're acknowledged.
14:50
Speaker 3: I wasn't suggesting that was the case, but just totally.
14:53
Speaker 2: You're the guest.
14:55
Speaker 3: Yeah, I appreciate it. And I wasn't suggesting that you
14:58
Speaker 3: weren't saying that. But yeah, to get into what you're saying,
15:00
Speaker 3: I do think, and you asked this earlier about what
15:03
Speaker 3: I get into in the book, I do think it's
15:05
Speaker 3: important for people to understand that data centers are not
15:08
Speaker 3: a new thing, right? In some form, they have been
15:10
Speaker 3: around since the beginning of computation, right? Obviously, there was
15:14
Speaker 3: a particular ramp up as the internet was privatized. We
15:18
Speaker 3: started to have all these digital services in the late
15:21
Speaker 3: 90s and into the early 2000s. Then in the book,
15:24
Speaker 3: I also really talk about the mid-2000s, especially with the
15:28
Speaker 3: creation of Amazon Web Services and then Microsoft Azure and
15:31
Speaker 3: Google Cloud afterward as an important turning point where you
15:35
Speaker 3: have much more of the movement toward the centralization of
15:38
Speaker 3: of computation by these major companies. I think that that
15:43
Speaker 3: really stands out to me as an important moment because
15:47
Speaker 3: without what happened then, without the beginning of the cloud,
15:52
Speaker 3: I would argue that you wouldn't really see generative AI
15:55
Speaker 3: in the form that we have it today without those
15:58
Speaker 3: companies beginning to have so much computation under their own
16:01
Speaker 3: roofs 15, 20 years ago. That was an important moment
16:07
Speaker 3: you know, an important moment where we had an important shift.
16:11
Speaker 3: But then if we kind of fast forward to the
16:13
Speaker 3: generative AI moment, we already had hyperscale data centers, you know,
16:18
Speaker 3: basically data centers that are a particular size that, you know,
16:22
Speaker 3: have tens of hundreds of thousands of servers within one roof.
16:27
Speaker 2: It's like 50 megawatts, I think, is the, I forget
16:30
Speaker 2: exactly what the turning point is to consider something hyperscale,
16:35
Speaker 2: but it's something like that, right?
16:37
Speaker 1: Exactly.
16:38
Speaker 3: For a while, there wasn't an industry-accepted definition, and I'm
16:42
Speaker 3: still not sure that there really is, but I feel
16:44
Speaker 3: like the 50 megawatts is one that has been gravitated around,
16:47
Speaker 3: especially in the past year, as the delineation between what's
16:50
Speaker 3: hyperscale and what isn't. And to be clear, 50 megawatts
16:55
Speaker 3: today is even on the lower side of what many of.
16:57
Speaker 2: These companies are building. It's a pathetic small one compared
17:00
Speaker 2: to the ones they're boasting about.
17:03
Speaker 1: Exactly.
17:04
Speaker 3: You're looking at like 100 and 150 on the lower
17:06
Speaker 3: end for some of these AI-focused ones. And then you're
17:09
Speaker 3: looking at clusters that are like a gigawatt, two gigawatts.
17:13
Speaker 3: Meta says that it eventually wants to build five gigawatts
17:16
Speaker 3: in one cluster in Louisiana.
17:17
Speaker 1: We'll see if that ever happens.
17:20
Speaker 2: It probably won't. Yeah, I don't think so either, right?
17:23
Speaker 1: But this is what they say they want to do.
17:26
Speaker 1: But I feel like the change that we started to
17:29
Speaker 1: see in 23, 24 was.
17:32
Speaker 3: One, an acceleration, right? So they're looking to build a
17:35
Speaker 3: lot more in a shorter period of time. There's also
17:38
Speaker 3: an increase of scale, right? So you already have hyperscale
17:41
Speaker 3: data centers, but they're getting larger and larger, and they're
17:44
Speaker 3: trying to cluster even more of them together in kind
17:47
Speaker 3: of a single region or a single space, if they
17:50
Speaker 3: can try to get away with it. And there's also
17:53
Speaker 3: the shift, as you were saying, from the more focus
17:57
Speaker 3: on CPUs and having hard drives in these data centers
18:00
Speaker 3: to using the graphics processing units from NVIDIA in particular,
18:04
Speaker 3: but obviously they're trying to use more different ones now
18:08
Speaker 3: if that's possible. And that's an important shift as well
18:11
Speaker 3: because the GPUs do run harder. They require more electricity,
18:18
Speaker 3: they throw off more heat, which means they need more
18:20
Speaker 3: cooling as well. And so even if you have a
18:23
Speaker 3: data center that's of similar size, and one is a
18:27
Speaker 3: more traditional data center and one is a more AI-focused
18:29
Speaker 3: data center, the AI-focused data center is going to have
18:33
Speaker 3: a bigger draw on electricity, on the water system, those
18:36
Speaker 3: sorts of things, than the one that would have been
18:39
Speaker 3: built before or the one that's built for more traditional
18:43
Speaker 3: use cases.
18:45
Speaker 1: These are all important developments.
18:47
Speaker 3: I feel like maybe the one other thing that really
18:49
Speaker 3: stands out for me when I think about it is
18:51
Speaker 3: just how aggressively the tech industry has really gone after
18:55
Speaker 3: fossil fuels in the past few years as well. If
18:58
Speaker 3: you go back to 2018, 2019, 2020, they were all
19:02
Speaker 3: starting to talk about how they cared about climate change
19:05
Speaker 3: and they were committed to net zero and reducing their
19:08
Speaker 3: impact and all this kind of stuff. And then as
19:10
Speaker 3: soon as generative AI comes along, all that gets thrown
19:14
Speaker 3: out the window. And now just the level of fossil
19:17
Speaker 3: fuels that they are building to power their data centers
19:21
Speaker 3: is just staggering. And at a level that makes really
19:25
Speaker 3: addressing climate change like seem virtually impossible.
19:29
Speaker 2: So and there has been pushback though and you've probably
19:33
Speaker 2: seen data center delays right did you happen to see
19:37
Speaker 2: any of the groundswell on the grant like have you
19:39
Speaker 2: seen any of the local opposition have you seen how people.
19:42
Speaker 3: Are pushing back on these things absolutely and so i
19:45
Speaker 3: met some people in some of the different countries that
19:48
Speaker 3: i went to who are doing this work, you know,
19:50
Speaker 3: in France, in Ireland, in places like that. And, you know,
19:54
Speaker 3: spoke to people in different parts of the U.S.
19:56
Speaker 2: As well, right?
19:57
Speaker 3: And I feel like this is one of the really
19:59
Speaker 3: hopeful stories for me. You know, we have really seen
20:02
Speaker 3: the tech industry push hard with its ambitions the past
20:06
Speaker 3: few years. You know, the push generated by AI into
20:08
Speaker 3: virtually every facet of our lives, to build out these
20:11
Speaker 3: massive data centers, to bring them to almost any community
20:14
Speaker 3: that they can possibly think to stick them in. But
20:19
Speaker 3: those communities have been really resilient in pushing back against it, right?
20:24
Speaker 3: We have seen across the United States hundreds of local
20:27
Speaker 3: moratoriums being passed on data center construction. Obviously, we have
20:31
Speaker 3: the first statewide moratorium in New York State right now,
20:35
Speaker 3: you know, people would quibble over, you know, whether it
20:37
Speaker 3: could have gone further, but, you know, at least it's
20:39
Speaker 3: kind of like a first step. You have the push
20:42
Speaker 3: on the federal level by people like Bernie Sanders to
20:45
Speaker 3: bring in a kind of federal moratorium on data center construction.
20:51
Speaker 3: And we see it in many other countries as well, right,
20:53
Speaker 3: where data center moratoriums have been passed. You know, just
20:56
Speaker 3: recently in Denmark, they basically put data centers at the
20:59
Speaker 3: bottom of their list for grid connections, so everything else
21:02
Speaker 3: has higher priority than a data center. In Canada, in Quebec,
21:06
Speaker 3: Hydro-Quebec just said data centers are going to pay double
21:09
Speaker 3: the rate of other electricity customers for their electricity. We
21:16
Speaker 3: do see this pushback, and I would argue that none
21:19
Speaker 3: of the moratoriums, very few of the government restrictions on
21:23
Speaker 3: data centers would be happening if it were not for
21:26
Speaker 3: people turning out in their communities and saying, we do
21:30
Speaker 3: not want this data center here. And I think that
21:33
Speaker 3: there are a number of reasons that they're doing that, right?
21:35
Speaker 3: I think, one, they have seen the stories of data
21:38
Speaker 3: centers for the past few years and what it can
21:40
Speaker 3: mean for their electricity prices, what it can mean for
21:44
Speaker 3: the local environment, what it means to live in the
21:46
Speaker 3: communities where they live. But then I also think that
21:49
Speaker 3: there are some people who are motivated by these bigger concerns.
21:53
Speaker 3: The concerns about what the tech industry has been doing,
21:56
Speaker 3: about how their products are affecting us, about the tech
21:59
Speaker 3: industry's alignment with Donald Trump and the far right. I
22:04
Speaker 3: think that there are a lot of reasons that people
22:06
Speaker 3: are getting motivated. But I just think generally there has
22:09
Speaker 3: been a growing frustration and anger with the tech industry
22:13
Speaker 3: And if you have a problem with, say, your social
22:16
Speaker 3: media app or something like that, it's hard to really,
22:19
Speaker 3: in a tangible way, make Mark Zuckerberg feel your anger
22:25
Speaker 3: at his decisions around how, say, the Facebook platform works
22:28
Speaker 3: or something like that. There have been a lot of
22:32
Speaker 3: moments where we've been angry at different apps or different companies,
22:35
Speaker 3: but it's hard to really do something tangible about it.
22:38
Speaker 3: And with the data centers, it feels like there's finally...
22:42
Speaker 3: a way to really stop something that they're doing because
22:46
Speaker 3: it is something tangible that would be coming to your
22:48
Speaker 3: community and you can turn out to your local council
22:51
Speaker 3: and you can get that project shut down in a
22:53
Speaker 3: way that you can't do with an app on your
22:55
Speaker 3: phone or something, you know?
22:56
Speaker 2: It's also a physical manifestation, like you said, of all
22:59
Speaker 2: the problems they're creating. And it also feels like a
23:02
Speaker 2: violation because the shit they've done to the internet is disgraceful.
23:07
Speaker 2: Like the ways they have twisted the internet up, they're...
23:11
Speaker 2: allegiance with the right wing their allegiance with the worst
23:14
Speaker 2: people alive and also they are now like no we
23:18
Speaker 2: need to build the biggest most power hungry ugliest fucking
23:21
Speaker 2: things you've ever seen just these giant shoebox looking motherfuckers
23:26
Speaker 2: that make endless noise and have tons of lights and
23:28
Speaker 2: all this shit and they're like no we need these
23:31
Speaker 2: and by the way in return we will get a
23:33
Speaker 2: tax break Yeah. And it's just, it's fucking ugly. I
23:37
Speaker 2: will say I've, I have seen, and if you were listening,
23:40
Speaker 2: there's a guy, I think in Wisconsin who in a
23:42
Speaker 2: data center meeting in like a town hall meeting call,
23:45
Speaker 2: like mentioned my name. If you're listening, mate, you're a legend.
23:49
Speaker 2: You're a legend of better off life. But there are
23:50
Speaker 2: people that listen to the show that go to these
23:52
Speaker 2: town hall meetings. You people are way more courageous than
23:54
Speaker 2: I am. You're doing you're like, whatever I do is
23:56
Speaker 2: nothing compared to the people actually going down there. These
23:58
Speaker 2: people are actual fucking heroes. and it's working as well.
24:03
Speaker 2: Like it's actually scaring the ever living shit out of
24:05
Speaker 2: these companies. I know because people in finance tell me
24:09
Speaker 2: and people in finance, in many cases don't have souls.
24:12
Speaker 2: So if they're saying it, it's because they're actually scared.
24:16
Speaker 1: Absolutely. Oh, I completely agree.
24:18
Speaker 3: And, you know, I think, and you would certainly be,
24:20
Speaker 3: be much more kind of on top of this than,
24:22
Speaker 3: than I am these days, because I've been just been
24:24
Speaker 3: so overwhelmed with, you know, everything that comes, it comes
24:27
Speaker 3: with releasing a book. Um, But, you know, obviously the
24:30
Speaker 3: recent stories around the growing questions about the financing of
24:34
Speaker 3: these data centers, you know, whether companies like Oracle or CoreWeave,
24:39
Speaker 3: you know, are actually going to be viable into the future,
24:42
Speaker 3: you know, whether all the money that they have expended
24:44
Speaker 3: on these data center plays and, you know, these kind
24:47
Speaker 3: of AI plays are actually going to pay off in
24:49
Speaker 3: the long run. You know, it has always been a question,
24:53
Speaker 3: you know, as you have been writing about for a
24:55
Speaker 3: long time, right? But it really feels that there's growing
25:00
Speaker 3: concern and questioning about the short-term viability of all of
25:05
Speaker 3: these investments and these speculative bets on generative AI and
25:10
Speaker 3: on the infrastructure that is supposed to be necessary in
25:14
Speaker 3: order to power generative AI and the vision of the
25:17
Speaker 3: future that these tech companies have.
25:21
Speaker 2: Within the book, when you were working on it, did
25:23
Speaker 2: you get any feedback from the industry? Did you hear
25:26
Speaker 2: back from any of the companies? I know Microsoft said no,
25:30
Speaker 2: but did you get any feedback from them? Were they
25:32
Speaker 2: trying to stop your line of inquiry anyway? Not really.
25:36
Speaker 3: And I wasn't very interested in really seeking their feedback anyway.
25:41
Speaker 2: No, I'm the same way. I don't need to hear
25:44
Speaker 2: from a bunch of liars.
25:46
Speaker 3: And we hear their perspective all the time, right? The
25:49
Speaker 3: media reports on their perspective constantly. If they have an
25:54
Speaker 3: opinion that you know, they want us to know on
25:56
Speaker 3: data centers, they're going to have said it somewhere anyway.
26:00
Speaker 3: And the focus of my book was really on, you know,
26:03
Speaker 3: what they're actually doing and also how that is affecting communities,
26:08
Speaker 3: you know, the wider public, you know, broadly the world
26:10
Speaker 3: that we live in, right? And so if it ever
26:13
Speaker 3: did feel necessary to put their perspective in there, I
26:16
Speaker 3: could grab it from somewhere else where they had said it.
26:18
Speaker 3: But in terms of actually going to them and be like,
26:21
Speaker 3: What's your response to the fact that you're burning all
26:23
Speaker 3: these fossil fuels? It just didn't seem important to essentially
26:29
Speaker 3: do that, right? Because we hear it anyway. And so
26:33
Speaker 3: why do I need that?
26:34
Speaker 2: Yeah, we just turn on CNBC, New York Times, TechCrunch,
26:37
Speaker 2: Business Inside, pretty much everywhere. I ask the question because
26:41
Speaker 2: I get asked it fairly often. Like, oh, do these
26:44
Speaker 2: companies ever say anything? I mean, no, because I'm calling
26:46
Speaker 2: them rat fucks and bastards. Also, if I needed them
26:51
Speaker 2: to say... For example, I worked on a story about...
26:56
Speaker 2: actually AI GPUs with the Guardian. And we got a
26:58
Speaker 2: quote from Microsoft because I had the numbers of how
27:01
Speaker 2: many GPUs they had. And they literally said, they're completely wrong.
27:04
Speaker 2: You're misreading them. Only for Bloomberg to report them two
27:07
Speaker 2: months later as completely what we said. It's just like,
27:10
Speaker 2: these people fucking lie. So why should we give them
27:12
Speaker 2: the time of day? And people will say, oh, it's
27:14
Speaker 2: not fair on them. Fuck that. It's not fair. They're
27:17
Speaker 2: not fair on us. They're not fair on you. They
27:20
Speaker 2: don't want to help.
27:21
Speaker 3: And it's also like, you know, if they had let
27:23
Speaker 3: me into the data centers, maybe there would have been,
27:26
Speaker 3: you know, an opportunity for a conversation there, right? But
27:30
Speaker 3: they were not interested in giving me kind of any
27:34
Speaker 3: access to see what they're doing. So why would I
27:36
Speaker 3: then turn to them and be like, oh, what's your perspective?
27:38
Speaker 3: And how are you going to try to mislead me
27:41
Speaker 3: on the things that you're very clearly doing? You know,
27:43
Speaker 3: I can quote their sustainability reports. I can quote what
27:47
Speaker 3: their executives have been saying or their comms people, you know,
27:51
Speaker 3: I often feel it's not valuable for me to go
27:53
Speaker 3: to them directly.
27:55
Speaker 4: Yeah.
27:56
Speaker 2: And here's another thing that just occurred to me as well.
27:58
Speaker 2: I'm going to guess you didn't find out how many
28:01
Speaker 2: gigawatts there were, because this is the biggest thing I've
28:03
Speaker 2: found with data centers that we quite literally don't know
28:06
Speaker 2: how many gigawatts there are, how many GPUs there are,
28:08
Speaker 2: how many are active. We don't know what it's like inside.
28:11
Speaker 2: We don't really know how many of them are cooled.
28:13
Speaker 2: It's just the whole thing, despite how prominent it is,
28:15
Speaker 2: is shrouded in secrecy.
28:17
Speaker 3: Oh, a hundred percent. And they work very hard to
28:20
Speaker 3: maintain that secrecy, as I'm sure I don't need to tell.
28:22
Speaker 1: You and your listeners.
28:24
Speaker 3: The amount of NDAs that they have people sign, the
28:28
Speaker 3: degree to which they try to even cover when they're
28:31
Speaker 3: entering a community and whether they're actually involved in a
28:34
Speaker 3: data center project until it's later down the line, they
28:38
Speaker 3: work really hard to try to keep these things secret,
28:42
Speaker 3: to try to keep them outside of the public conversation
28:45
Speaker 3: because they know that people are not going to be
28:48
Speaker 3: down with this and okay with this uh when they
28:50
Speaker 3: learn what is actually.
28:51
Speaker 2: Happening yeah and it's for me and this we talked
28:56
Speaker 2: about it before went on but amazon is business insider
28:59
Speaker 2: revealed that amazon is actually warehousing gpus fucking exactly what
29:03
Speaker 2: I said a month or two ago, but everyone said
29:05
Speaker 2: I was crazy. But nevertheless, it is kind of crazy
29:10
Speaker 2: to me that we don't even know how many are installed,
29:13
Speaker 2: that this is, well, hundreds of billions of dollars of financing,
29:17
Speaker 2: the economy resting on it to some extent. We don't
29:20
Speaker 2: know anything. And there's been no government attempts to demand
29:25
Speaker 2: any kind of clarity, it appears.
29:27
Speaker 3: Yeah.
29:28
Speaker 1: And of course, whose pocket is the government in? Unfortunately-
29:32
Speaker 1: it's often the AI companies, right?
29:33
Speaker 3: It's like you see Donald Trump come out and say
29:35
Speaker 3: that communities opposing data centers just want to be like,
29:39
Speaker 3: I think it was poor and backward.
29:41
Speaker 1: He said. Yes.
29:42
Speaker 2: It's the Chinese. Exactly.
29:45
Speaker 1: It's crazy, right?
29:46
Speaker 3: Especially when you see that so many of his supporters,
29:49
Speaker 3: like so many Republicans, are also opposed to these data
29:52
Speaker 3: center projects. So many of them are not buying into it.
29:56
Speaker 3: And it's kind of fascinating.
29:57
Speaker 1: Obviously, the U.S.
29:58
Speaker 3: Is heading into the midterm elections in just a couple
30:01
Speaker 3: weeks now. And you have, on the one hand, the
30:06
Speaker 3: Republican Party or the leadership of it that is very
30:10
Speaker 3: much in line with the tech industry and the AI industry.
30:14
Speaker 3: But so many of its supporters are not buying into
30:17
Speaker 3: that project. so many of the politicians on the lower
30:20
Speaker 3: levels are having to moderate their support for the industry
30:24
Speaker 3: and data centers because they recognize it is an electoral
30:28
Speaker 3: liability for them. They might actually lose their seats or
30:31
Speaker 3: their positions if they are as pro-data center or pro-AI
30:35
Speaker 3: as they have been in the past. And we've already
30:37
Speaker 3: seen this happen in certain states in the United States.
30:41
Speaker 3: So yeah, I don't know. It's kind of a fascinating
30:44
Speaker 3: moment because on the flip side of that then, you
30:46
Speaker 3: still have the AI industry trying to copy the playbook
30:48
Speaker 3: of the crypto industry two years ago and spending a
30:51
Speaker 3: ton of money in races to try to get the
30:54
Speaker 3: politicians that are more favorable to their kind of regulatory ambitions.
30:59
Speaker 3: And by that, I mean getting regulations that favor them
31:03
Speaker 3: so that they can say, oh, the industry is regulated,
31:06
Speaker 3: but it's regulated with regulations that they have written to
31:09
Speaker 3: benefit themselves, right?
31:24
Speaker 2: But the thing is, this is the thing that kind
31:26
Speaker 2: of pisses me off as well, is it's like people say, oh, tech,
31:30
Speaker 2: they just want special regulations. Like, what fucking regulation? We
31:35
Speaker 2: don't regulate tech. Precisely, right? We don't do it. And
31:39
Speaker 2: people think, oh, they want to rig the things. They've
31:42
Speaker 2: already rigged it. In their favor.
31:45
Speaker 1: Oh, I completely agree.
31:46
Speaker 3: And to me, if you look at this conversation that's
31:50
Speaker 3: been happening, that's really kind of gained more prominence the
31:53
Speaker 3: past couple months of the AI is going to destroy
31:56
Speaker 3: the world and is going to kill humanity in 10
31:58
Speaker 3: years or whatever, is very much an attempt to be like, okay,
32:02
Speaker 3: public anger is rising.
32:04
Speaker 1: There is pressure to actually.
32:05
Speaker 3: Regulate the things that we are doing and that we
32:09
Speaker 3: don't want to have regulated. And so we need to
32:11
Speaker 3: step into the conversation now and say, hey, there are
32:15
Speaker 3: problems with AI, but they're these future-focused fantastical problems, not
32:22
Speaker 3: the things that you're experiencing, whether it is deep fakes
32:25
Speaker 3: or the mental health spirals or the questions about cognitive
32:28
Speaker 3: capacity and learning from dependence on chatbots, or on and
32:32
Speaker 3: on and on, the impacts of data centers, whatever, I
32:35
Speaker 3: can give you a very long list. It's like, we
32:37
Speaker 3: do not want regulations on those things. that would actually
32:40
Speaker 3: affect how our products are operating right now, how our
32:43
Speaker 3: companies operate. Instead, we want lawmakers to be focused on
32:47
Speaker 3: the fantasies in 10 years, or maybe all the jobs
32:49
Speaker 3: are going or whatever, because those are things they're just
32:54
Speaker 3: making up anyway. But if they can make sure that, say,
32:57
Speaker 3: Bernie Sanders or someone, when they're talking about AI, is
33:00
Speaker 3: talking about the made-up narratives that they're putting out there
33:03
Speaker 3: about AI, that the industry is putting out there, then
33:05
Speaker 3: that benefits them immensely, because then the regulation is based
33:09
Speaker 3: on fantasy and not on reality and kind of, you know,
33:14
Speaker 3: actual harms that people are feeling right now.
33:16
Speaker 2: And that's kind of what pissed me off about Bernie Sanders.
33:18
Speaker 2: I like Bernie. I like what Bernie is pushed for,
33:21
Speaker 2: but I feel like he's just boofed this entire AI thing.
33:25
Speaker 2: Ever since he did that weird conversation with Claude, he's like, Claude,
33:28
Speaker 2: are you alive? And it's just, but it is exactly
33:32
Speaker 2: what you said. It's this thing of getting- This imaginary
33:36
Speaker 2: creature that doesn't exist and won't exist regulated so that
33:39
Speaker 2: they can keep doing whatever they want. right now. There
33:43
Speaker 2: are obvious regulations they could make right now. My favorite
33:46
Speaker 2: one to choose is notifications. I think if we regulate
33:50
Speaker 2: them like emissions, I think it could do genuinely good
33:54
Speaker 2: to hundreds of millions, if not billions of people. It
33:56
Speaker 2: sounds small, but making it so the growth hacking things
34:00
Speaker 2: are harder to do, that would be an obvious one.
34:02
Speaker 2: But it's like, no, we need to make sure that
34:04
Speaker 2: when AI makes a profit- The profit share might happen
34:07
Speaker 2: one time if you're this person. And it's just like
34:10
Speaker 2: fucking useless, worse than useless.
34:14
Speaker 1: Yeah, I completely agree.
34:15
Speaker 3: And, you know, I think I hate to, you know,
34:19
Speaker 3: kind of make the age commenter.
34:21
Speaker 1: I don't know. I don't really. No, you can't go
34:23
Speaker 1: for it.
34:24
Speaker 3: I think Bernie Sanders is like a very well-meaning person,
34:27
Speaker 3: but I think he's a bit in over his head
34:28
Speaker 3: on this and really doesn't understand what's going on with
34:31
Speaker 3: the AI industry. And, you know, that's not to say
34:35
Speaker 3: that I don't think older people can understand technology. I
34:37
Speaker 3: think they definitely can. And I think part of it
34:39
Speaker 3: is probably like his staffers are probably feeding him this
34:42
Speaker 3: stuff as well, right? Which means that he doesn't have
34:44
Speaker 3: the right people around him to help him understand what
34:46
Speaker 3: is actually going on with the technology and what the
34:49
Speaker 3: right approach would be. You know, I was completely behind
34:52
Speaker 3: the effort to, you know, bring forward a nationwide moratorium
34:57
Speaker 3: on data centers that he was advocating end of last year.
35:00
Speaker 3: But then when it came to actually looking at the
35:02
Speaker 3: regulations on AI that he was moving forward. It really
35:05
Speaker 3: felt that he was just kind of getting caught up
35:07
Speaker 3: in the narratives that came from industry that were designed
35:10
Speaker 3: to be very deceptive.
35:12
Speaker 1: I would say, I.
35:13
Speaker 3: Think that Alexandria Ocasio-Cortez has been better on this and
35:17
Speaker 3: has been more focused on deepfakes and mental health and
35:20
Speaker 3: these sorts of questions. but I feel like her efforts
35:23
Speaker 3: have not been getting the attention and the traction that
35:26
Speaker 3: Bernie Sanders have because Bernie's reflect the goals and intentions
35:31
Speaker 3: of the industry and the sensationalized claims that they put
35:35
Speaker 3: out there, whereas hers have been focused much more on
35:37
Speaker 3: very tangible things that are actually hurting people and that
35:40
Speaker 3: obviously don't get the same degree of attention.
35:44
Speaker 2: The other bloke on the bill, I forget who it was,
35:47
Speaker 2: These PR people came at me. They tried to go
35:50
Speaker 2: through contacts of mine to try and get on. And
35:51
Speaker 2: just to be clear, if you were listening, mate, I
35:53
Speaker 2: don't give a fuck about moderating super intelligence. Stop pretending
35:56
Speaker 2: that that's the problem. No one gives a fuck about,
35:59
Speaker 2: if you're regulating the scary computer because Jacob Coxon said something,
36:02
Speaker 2: you've already fucking lost. Yeah, exactly. It's a classic capitalism
36:06
Speaker 2: trick as well. It's obfuscating the actual problem so that
36:10
Speaker 2: people can feel good about themselves because you can be like, oh,
36:14
Speaker 2: when the scary computer wakes up, this will happen and
36:17
Speaker 2: we will already be ready we get to feel good
36:19
Speaker 2: about something without doing anything without restraining capital even slightly
36:24
Speaker 2: and it's just it's very frustrating because by the time
36:28
Speaker 2: this all unwinds we are also going to need regulations
36:31
Speaker 2: around what to do with unbuilt data centers yeah and
36:36
Speaker 2: we don't we are not going to be prepared for
36:37
Speaker 2: that we're going to be prepared for the mystical jack
36:40
Speaker 2: clark computer machine that wakes up and fucks your granny
36:43
Speaker 2: or whatever and it's it's just it frustrates me as well,
36:46
Speaker 2: because you know this as well as anyone, the problems
36:49
Speaker 2: have been obvious for years. All of them, like the environmental,
36:53
Speaker 2: the social, the economic, they've been there. They've pretty much
36:57
Speaker 2: been the same, but maybe smaller a few years ago.
37:00
Speaker 2: And we have not tried to stop any of them.
37:04
Speaker 2: Like we've just, we focused entirely on an imaginary one.
37:07
Speaker 1: Exactly.
37:08
Speaker 3: Like if you're paying attention, you knew this stuff, right? And,
37:10
Speaker 3: you know, I would say if you're a lawmaker and
37:12
Speaker 3: you have an office and staff, Though staff could have
37:15
Speaker 3: been paying attention to it as well if this is
37:17
Speaker 3: an issue that you as the lawmaker do want to.
37:20
Speaker 1: Be active on and doing things on.
37:24
Speaker 3: The other thing that is frustrating, of course, obviously someone
37:28
Speaker 3: like Bernie Sanders being on the left, very opposed to
37:31
Speaker 3: the far right, has been very vocal about the power
37:34
Speaker 3: that Trump has, the way that the tech industry has
37:37
Speaker 3: been working with the far right. And then you see
37:40
Speaker 3: him on stage with Steve Bannon because AI is going
37:43
Speaker 3: to kill the world or something or kill humanity or something.
37:45
Speaker 3: It's like, what are you doing? And what kind of
37:49
Speaker 3: alliances are you forming? And how are you giving, you know,
37:54
Speaker 3: basically credibility to someone who a few years ago you
37:59
Speaker 3: would have been like, you know, you can never touch
38:02
Speaker 3: that person.
38:02
Speaker 1: You can never work with that person.
38:04
Speaker 3: And there's some really dangerous, I think, alliances being built
38:08
Speaker 3: in that moment as well which is quite distinct, I
38:11
Speaker 3: would say. You know, when we're seeing like a Bernie
38:13
Speaker 3: Sanders and Steve Bannon on stage together, I think that's.
38:15
Speaker 2: A big problem. Yeah, that turned my fucking stomach. Oh my, 100%. Like,
38:21
Speaker 2: especially given the, I would say questionable previous thing people
38:26
Speaker 2: are called Bernie bros, but it's like, what the fuck
38:28
Speaker 2: do you say to that when you're on stage with
38:30
Speaker 2: fucking Steve? What do you say to a woman to
38:33
Speaker 2: like try and defend Bernie Sanders where you're like, he's
38:35
Speaker 2: standing on the stage with fucking Steve fucking Bannon. And
38:38
Speaker 2: it's because of this wibbly-wobbliness of AI. And honestly, I
38:43
Speaker 2: also think it's an attempt to paraphrase hyper-normalization to make
38:47
Speaker 2: a complex situation simple. Oh, if we only prepare for
38:52
Speaker 2: a future that sounds roughly like the ones we've seen
38:56
Speaker 2: in the movies, we'll be fine. Rather than dealing with...
39:02
Speaker 2: rather than dealing with the actual obvious problems, which is
39:05
Speaker 2: we have unrestrained capital that is able to manipulate the
39:10
Speaker 2: debt markets, and we have a captured tech and business
39:13
Speaker 2: media who are actively marketing for the capital classes.
39:18
Speaker 1: Exactly. 100%. I completely agree.
39:19
Speaker 3: And just to finish the point I was making there,
39:22
Speaker 3: it's like there's a real problem.
39:24
Speaker 1: No, not at all.
39:25
Speaker 3: There's a real problem with Bernie Sanders and Steve Bannon
39:28
Speaker 3: on stage together. I don't see that problem at all
39:31
Speaker 3: when you're looking at, you know, kind of local organizers
39:34
Speaker 3: who are Democrats, who are Republicans, who don't have a
39:37
Speaker 3: fucking clue what their political ideology is, and they're all
39:39
Speaker 3: coming together to kind of face off against these data
39:44
Speaker 3: centers together to take them on and to have, like,
39:48
Speaker 3: maybe they come at it from very different political persuasions.
39:51
Speaker 3: They have very different reasons for opposing these data centers.
39:54
Speaker 3: But at the end of the day, they are people
39:56
Speaker 3: who are motivated to protect their community, to push back
39:59
Speaker 3: on the power of the tech industry. And maybe it's
40:01
Speaker 3: actually going to lead them to have some interesting conversations
40:04
Speaker 3: that they wouldn't have usually had because they wouldn't be
40:06
Speaker 3: talking to one another in this way right and so
40:08
Speaker 3: i think that and.
40:09
Speaker 2: That is actually bipartisan that is actually two people potentially
40:12
Speaker 2: different political creeds who share a common community bernie sanders
40:16
Speaker 2: and steve bannon have maybe two or three things in
40:19
Speaker 2: common like old white but not looking great like i
40:23
Speaker 2: don't fucking know and it's just frustrating as well because
40:26
Speaker 2: especially again being on the left myself and like watching
40:30
Speaker 2: bernie and being like wow did like this guy believes
40:32
Speaker 2: in things i do wow like my father worked at
40:34
Speaker 2: the nhs like i like i really believe in in
40:37
Speaker 2: these socialist causes. And then this fucking guy gets on
40:40
Speaker 2: stage with Steve Bannon to basically protect the interests of
40:44
Speaker 2: capital by claiming you're regulating them.
40:46
Speaker 1: Yeah.
40:47
Speaker 2: And it just makes me wonder how much the rest
40:50
Speaker 2: of it meant. I don't know. Maybe that's a little bit,
40:52
Speaker 2: taking it a little bit far, but it's when your
40:54
Speaker 2: inherent beliefs are so, or even just your knowledge is
40:58
Speaker 2: so weak that you're willing to align with an actual fascist,
41:01
Speaker 2: a guy who has called multiple times for actual real
41:04
Speaker 2: life fascism, just like, Let's be clear about who Steve
41:07
Speaker 2: fucking Bannon is. And you are on stage with him
41:11
Speaker 2: because he also agrees that a thing that isn't happening
41:15
Speaker 2: might happen and it's bad, even though he is obviously
41:18
Speaker 2: doesn't give a shit about anything and just wants attention. Jesus,
41:23
Speaker 2: it turns my stomach. It's the same criticism I have
41:27
Speaker 2: for the media. It's these people who claim, oh, we're
41:30
Speaker 2: here to find the truth and report on the truth,
41:32
Speaker 2: and then they just repeat whatever the fuck Altman or
41:34
Speaker 2: Amaday says. It just genuinely pisses me off.
41:38
Speaker 1: I completely agree with you.
41:39
Speaker 3: And I- have not been quiet about my frustrations with,
41:44
Speaker 3: with the media and how it reports on any of
41:45
Speaker 3: this either.
41:46
Speaker 1: Right. You know, we're very, we're very.
41:48
Speaker 3: Aligned on that, on this question because it's like, you know,
41:51
Speaker 3: for me, I see it as a real, you know,
41:53
Speaker 3: if you think about what journalism is supposed to do, right.
41:55
Speaker 3: It's really supposed to help to have an informed public, right.
41:59
Speaker 3: You know, they know what's going on, but also you
42:01
Speaker 3: would hope, you know, if you think about the more
42:04
Speaker 3: kind of idealistic visions of what journalism is, it's also
42:08
Speaker 3: to hold power to account. And unfortunately what we see,
42:12
Speaker 3: is a journalism that all too often defends power, but
42:16
Speaker 3: also repeats power's narratives such that the public is misinformed
42:20
Speaker 3: about what is actually happening. The public is led to
42:23
Speaker 3: believe that these chatbots are actually thinking and developing independent
42:28
Speaker 3: consciousness and all this kind of stuff when we know
42:32
Speaker 3: very tangibly that that is not occurring. Repeating the other
42:36
Speaker 3: narratives that come from these executives and these companies is
42:40
Speaker 3: that are designed to deceive and mislead. And so then
42:44
Speaker 3: the public, when they have their conversations about AI, um, or,
42:48
Speaker 3: you know, about any of this stuff that's going on
42:49
Speaker 3: with the tech industry, it is kind of shaped and
42:52
Speaker 3: influenced by, um, this deception that has been echoed and proliferated, um,
42:58
Speaker 3: by the media, uh, instead of them kind of stepping
43:01
Speaker 3: in and saying like, this is what the CEOs are saying.
43:04
Speaker 3: This is what the companies are saying. Um, This is
43:07
Speaker 3: what experts are saying, though, and that does not really
43:10
Speaker 3: align with what is going on there. And certainly there's
43:12
Speaker 3: some of that sometimes, but as me and you and
43:15
Speaker 3: many of your listeners will know, the overwhelming amount of
43:20
Speaker 3: writing and reporting that happens on tech is just echoing
43:22
Speaker 3: what the companies themselves say, which is not helping.
43:26
Speaker 2: And what's crazy is, kind of like Bernie, I think
43:28
Speaker 2: a lot of these writers think they're doing a good job.
43:31
Speaker 2: I think they think that they are Because in a
43:34
Speaker 2: story that largely parrots what the rich say, they have
43:37
Speaker 2: a single paragraph that says, there are questions about the
43:39
Speaker 2: economic viability of this.
43:41
Speaker 1: And I just want to say- Yeah, like paragraph 38
43:43
Speaker 1: in the story or something.
43:44
Speaker 4: Exactly.
43:44
Speaker 2: And if you're a journalist listening to this, I am
43:47
Speaker 2: talking about you. Just if you're wondering if I'm talking
43:50
Speaker 2: about you, yes. Sorry, it's just, I'm just going to
43:53
Speaker 2: skip right to it. If you think I'm critiquing your work,
43:55
Speaker 2: I probably am. If you have any concern about this,
43:58
Speaker 2: I am talking about you. Fucking do your fucking job.
44:01
Speaker 2: I'm fucking sick of reading your bullshit. I'm fucking sick.
44:03
Speaker 2: Like I'm actually, this Amazon data center GPU thing, the GPUs,
44:08
Speaker 2: it has really pissed me off because like Jensen Huang
44:12
Speaker 2: basically made everything more expensive. to sell these GPUs in advance.
44:16
Speaker 2: They build these data centers. And for years, the journalists
44:19
Speaker 2: just gone, wow, they're building the data centers. Wow, they're
44:22
Speaker 2: so big. Wow, these things are going to print money. Wow,
44:24
Speaker 2: Anthropic has five gigawatts of capacity. Even though they don't,
44:28
Speaker 2: there's no proof of that. Let's just keep repeating it.
44:31
Speaker 2: And it's why these companies keep being able to do
44:33
Speaker 2: what they do. It's not just a regulatory problem. It's
44:36
Speaker 2: an accountability problem led by journalism.
44:39
Speaker 1: Absolutely.
44:40
Speaker 3: And the other piece, and one of the things I
44:42
Speaker 3: really try to get to in the book is like, okay,
44:45
Speaker 3: we're building out all this computation, or at least supposedly
44:47
Speaker 3: building it out, planning to build it out, whatever. The
44:49
Speaker 3: companies are saying that this is all required and that
44:52
Speaker 3: they're deploying all this capital to build out these data centers.
44:55
Speaker 3: But what is actually the consequence of that? We know
44:59
Speaker 3: that there's all these resources that we have that are
45:02
Speaker 3: being deployed now for data centers, whether those are capital resources,
45:06
Speaker 3: but also energy, fossil fuels, water, mineral resources, if you're
45:12
Speaker 3: thinking about the production of the chips and everything that
45:14
Speaker 3: goes into that, right? We're kind of dedicating so much
45:17
Speaker 3: capacity to building out these data centers. And we're also
45:20
Speaker 3: accepting that there are a lot of harms, or at
45:23
Speaker 3: least the industry is hoping that we accept, that we're
45:26
Speaker 3: just going to burn a ton more fossil fuels and
45:29
Speaker 3: contribute a bunch of emissions to the climate. And we're
45:32
Speaker 3: going to make a ton of communities be more worried
45:35
Speaker 3: about whether they're actually going to have access to water
45:37
Speaker 3: or even threaten that access to water. Not to mention
45:41
Speaker 3: all the things in the supply chain and the impacts
45:45
Speaker 3: on where the production of the chips happen and where
45:47
Speaker 3: the mining of the minerals happens. There are all these
45:50
Speaker 3: impacts that we're just supposed to go along with because
45:54
Speaker 3: generative AI is.
45:54
Speaker 1: The next big thing.
45:55
Speaker 3: But then, what are we actually getting from generative AI?
45:59
Speaker 3: A few years ago, the companies were saying it was
46:01
Speaker 3: going to be this next industrial revolution. There was going
46:03
Speaker 3: to be so much efficiency and productivity that was going
46:06
Speaker 3: to come from it. you know, that it was going
46:08
Speaker 3: to improve access to like key services like healthcare and education.
46:11
Speaker 3: As far as I'm concerned, and as far as I've seen,
46:15
Speaker 3: we've not seen much of that actually come to be.
46:19
Speaker 3: You know, we've not seen many, if any, of these
46:22
Speaker 3: kind of positive benefits be realized. But the amount of
46:25
Speaker 3: kind of social, community, environmental harms that have proliferated because
46:29
Speaker 3: of these technologies are The list is so long, and
46:33
Speaker 3: I mentioned some of them earlier. I could go on,
46:36
Speaker 3: but I'm sure that your listeners know what they are.
46:39
Speaker 3: And so it's like, okay, we're deploying all these resources
46:41
Speaker 3: for data centers. We're kind of pissing off all these
46:45
Speaker 3: communities in order to build them. But what are we
46:47
Speaker 3: actually even getting for all that? And I just don't
46:50
Speaker 3: see how any of it is justified. And to add
46:53
Speaker 3: to that, to pick up on our conversation that we
46:55
Speaker 3: were just having, how is it that our lawmakers and
46:58
Speaker 3: governments are just going along with all this are not
47:01
Speaker 3: questioning whether any of it makes sense just because there's
47:03
Speaker 3: so much money wrapped up in it. It's like, it
47:05
Speaker 3: just makes you kind of disillusioned even further with the
47:09
Speaker 3: whole system.
47:10
Speaker 4: Yeah.
47:10
Speaker 2: It's just, I mean, it's because we're in the scourge
47:13
Speaker 2: of neoliberalism. It's just, we are still within a world
47:17
Speaker 2: that believes in growth at all costs. And also I
47:20
Speaker 2: won't be honest. I think a lot of journalists, I
47:23
Speaker 2: think a lot of lawmakers, I think that they want
47:27
Speaker 2: the tech companies to win. I think it's that simple.
47:31
Speaker 2: I think that they believe it's a good thing. I
47:33
Speaker 2: think we still have people in this kind of cutesy
47:37
Speaker 2: 2011 mindset that these are little startups versus basically the
47:41
Speaker 2: blitzpins of our economy in a very brittle way. And
47:44
Speaker 2: it fucking pisses me off. I don't have an eloquent
47:47
Speaker 2: way of saying it. It's just frustrating. But I think
47:51
Speaker 2: we could... The thing that I'm trying to come around to,
47:54
Speaker 2: because it's been in the back of my mind for
47:56
Speaker 2: a while, is that... The financial apocalypse to follow will
47:59
Speaker 2: be a good thing because I think there's going to
48:01
Speaker 2: be a division between people based on how they reacted
48:04
Speaker 2: to AI. And just to be clear, the people that
48:07
Speaker 2: use LLMs normally, the people like, oh, this is kind
48:09
Speaker 2: of useful. I don't think they're– it's fine. Like, whatever. Oh, 100%.
48:13
Speaker 3: And there's a lot of people who use generative AI
48:16
Speaker 3: tools who also don't like what generative AI is and
48:19
Speaker 3: the impact that it has, right?
48:21
Speaker 2: But there are also the rest of them. And there's
48:24
Speaker 2: so many people who acted profoundly weirdly over this, who
48:27
Speaker 2: are also waved away, hand wave, the whole economic side,
48:31
Speaker 2: the social problems, socioeconomic problems, the deep fakes. They were
48:36
Speaker 2: just like, oh, you know, but we're getting something in return.
48:38
Speaker 2: I don't know how we are going to talk to
48:41
Speaker 2: those people in the future. I don't know how we
48:44
Speaker 2: are going. It's going to, it's like we've permanently put
48:46
Speaker 2: on the, they live glasses, just these people. I don't
48:50
Speaker 2: know how we are meant to have a conversation with,
48:53
Speaker 2: with someone who was super pro-AI, who was super aggressively
48:57
Speaker 2: defensive of AI, and they're everywhere.
49:03
Speaker 3: I want to believe and hope that what you're saying
49:07
Speaker 3: is accurate, but then I also think back to the
49:11
Speaker 3: crypto moment, and so many of the people who were
49:13
Speaker 3: boosting crypto just moved into boosting AI and kept the
49:17
Speaker 3: grift going. Obviously, we can think about many people, but
49:24
Speaker 3: someone like Kevin Roos comes to mind, obviously. Formerly of
49:28
Speaker 3: the New York Times, I can't remember exactly what he's.
49:30
Speaker 2: At these days. He's got to NPR, I think, now.
49:33
Speaker 1: Okay, interesting. But he was.
49:36
Speaker 3: He was telling us how crypto and NFTs were the
49:38
Speaker 3: next big thing and then it all collapsed and he
49:41
Speaker 3: didn't really seem to have to apologize for that and
49:44
Speaker 3: just kind of pivoted into telling us that AI was
49:46
Speaker 3: the next big thing and we were all going to
49:48
Speaker 3: have to use that and blah, blah, blah, right?
49:50
Speaker 2: There is one big difference though. And that's with crypto,
49:54
Speaker 2: it never actually got much social presence. Like crypto never,
50:01
Speaker 2: people weren't really and aren't really transacting with crypto anymore.
50:05
Speaker 2: there wasn't a massive build out of, there was a
50:08
Speaker 2: build out, don't get me wrong, it's where CoreWeave came from,
50:11
Speaker 2: many others, I think of Cypher Mining as well, like
50:14
Speaker 2: came from crypto, but we didn't have a massive one.
50:16
Speaker 2: And we also, while we had some media pressure, we
50:20
Speaker 2: certainly didn't have everyone screaming, this is the next big
50:24
Speaker 2: thing in the way we do today. And on top
50:26
Speaker 2: of it, people didn't change their entire lives, like regular folk,
50:31
Speaker 2: like people, not even talking Bruce here, who gives a shit, he's,
50:35
Speaker 2: He'll have fun in a year watching this burn. But
50:39
Speaker 2: he is... I'm talking more... There are people who are
50:44
Speaker 2: relatively normal who something about AI and defending the AI
50:48
Speaker 2: industry just uncorked them. And there are a lot of them.
50:52
Speaker 2: And they're everywhere. There are journalists, policymakers, regular people we know.
50:57
Speaker 2: And these people have alienated themselves in their friend bases
51:00
Speaker 2: because they... just became weird AI freaks. Like, I don't
51:05
Speaker 2: know how else to put it. And the people who
51:06
Speaker 2: were defending the financials were like, oh, I'm taking detailed
51:09
Speaker 2: notes on those who would say OpenAI would go bankrupt.
51:12
Speaker 2: Could be talking about anyone. And it's just, I don't
51:16
Speaker 2: think that it's going to be easy to put that
51:19
Speaker 2: toothpaste back in the tube because of the level of
51:23
Speaker 2: crazy they went. Like, the things that people have been
51:25
Speaker 2: saying about AI economics alone are fucking crazy. No, I
51:30
Speaker 2: I'm definitely on.
51:31
Speaker 1: Board with you.
51:32
Speaker 3: I definitely hope that's the way it plays out. It
51:34
Speaker 3: just seems to feel like any time one of these
51:37
Speaker 3: things happen, if you're on board with what the industry
51:40
Speaker 3: says the next big thing was supposed to be, they
51:42
Speaker 3: will make room for you to have a resurrection on
51:45
Speaker 3: the other side of it. But hopefully because of the
51:48
Speaker 3: scale of what we see with AI and data centers
51:52
Speaker 3: and the scale of what the collapse will ultimately have
51:54
Speaker 3: to be when that comes, and hopefully we're starting to
51:58
Speaker 3: see the beginning of it now, is yeah hopefully that
52:02
Speaker 3: means that that it is different this time and that
52:05
Speaker 3: there will have to be a proper reckoning.
52:08
Speaker 2: All right as we wrap up here i just want
52:09
Speaker 2: to be clear there will be a reckoning i've been
52:11
Speaker 2: taking detailed notes and every journalist has massively backed this
52:14
Speaker 2: i have i will not be kind i will not
52:17
Speaker 2: be sweet and i will not let up until everyone
52:20
Speaker 2: has apologized or left the industry you have been listening
52:23
Speaker 2: to better offline paris thank you so much for joining
52:25
Speaker 2: me i'll have a link to your book and everything
52:27
Speaker 2: the podcast as well in the notes thanks for joining
52:30
Speaker 2: man Absolutely.
52:31
Speaker 1: Thanks so much. And looking forward to having you back
52:33
Speaker 1: on Tech Won't Save Us soon.
52:34
Speaker 2: Looking forward to it. Thanks for listening, everyone. Thank you
52:45
Speaker 2: for listening to Better Offline. The editor and composer of
52:48
Speaker 2: the Better Offline theme song is Matt Osowski. You can
52:51
Speaker 2: check out more of his music and audio projects at mattosowski.com.
52:55
Speaker 2: M-A-T-T-O-S-O-W-S-K-I dot com. You can email me at ez at
53:01
Speaker 2: betteroffline.com or visit betteroffline.com to find more podcast links and,
53:05
Speaker 2: of course, my newsletter. I also really recommend you go
53:08
Speaker 2: to chat.wheresyoured.at to visit the Discord and go to r
53:12
Speaker 2: slash betteroffline to check out our Reddit. Thank you so
53:16
Speaker 2: much for listening.
53:17
Speaker 4: Better Offline is a production of Cool Zone Media. For
53:20
Speaker 4: more from Cool Zone Media, visit our website, coolzonemedia.com, or
53:24
Speaker 4: check us out on the iHeartRadio app, Apple Podcasts, or
53:27
Speaker 4: wherever you get your podcasts.