00:01:32.040We have a great feature interview today with Alam Bokhari,
00:01:35.420talking about AI and data centers and basically what we should do about our
00:01:40.860worries that maybe things will head to a dystopian future.
00:01:44.620Alam's got some thoughtful answers to that.
00:01:46.560But first, I want to show you about a 10-minute video we made last night after I was done this show when I received a phone call from our truck driver, Mark, that our billboard truck had been swarmed by cops downtown because we had a message on against men colonizing women's sports because the WNBA was in town.
00:02:09.280that's the women's professional basketball league. And our truck was down there and the cops gave
00:02:16.200Mark two tickets for $615 each. And they made up something about needing a permit or something.
00:02:23.940We've been driving that truck for years. I checked the law. There is no requirement for a billboard
00:02:28.680truck to have a special permit. They made that up and they threatened and tried to intimidate
00:02:33.680our driver. It took me about 40 minutes to get down there, but I did eventually get down there.
00:02:38.040I also talked to one of the cops on the phone.
00:02:40.320So I want to show you 10 minutes of what happened last night.
00:02:45.880And I think we're in, I think we're in a dangerous place.
00:02:49.220This is the second time the police have accosted our truck when we had transgender messages on it.
00:02:55.640You might remember the first time it was when the truck was parked and a cop said he was doing a hate crime investigation.
00:15:55.760And if there is, I know our guest today will help explain it to us.
00:15:59.720I'm talking about our old friend, Alan Bukhari, who we first met when he was a senior tech
00:16:04.320editor at Breitbart.com some 10 years ago. And now he's the managing director of the Foundation
00:16:09.200for Freedom Online. And he's the author of a very important book from a few years back called
00:16:14.540Deleted Big Tech's Battle to Erase the Trump Movement and Subvert Democracy. I'm going to ask
00:16:21.620him if some of those concerns are still alive today. Joining us now via Zoom is our friend
00:16:26.160Alam. Alam, even the fact that we're talking, it's great to see you again, by the way,
00:16:29.280Even the fact that we're talking through Zoom and that that's completely normal, that's part of what a data center is.
00:16:36.160Just every single thing in our life uses data, whether it's a Netflix show, a Facebook page, everything.
00:16:43.380But suddenly that's being multiplied by a thousand or a billion because the amount of computer power needed for this AI future is so much larger.
00:17:55.820And I think the tech industry is, on the censorship front, the tech industry has realized that they were effectively being bullied by, you know, the federal, you know, federal officials and the advertising industry and all of these interest groups that wanted to shut down free speech on the Internet.
00:18:13.440They have realized that they've woken up to it.
00:18:15.500And a lot of the tech backlash is actually coming from global progressives and the left who are upset that people like Elon Musk are now allies with the Trump administration, that they're now on the political right.
00:18:29.740This is a very big change that only happened in the last few years.
00:18:32.940They're upset at people like Mark Zuckerberg because they relaxed censorship on their platforms.
00:18:37.080And so we have to be cognizant that some of the hostility towards tech right now is coming from sort of these jilted progressives, these globalists, these global censors who feel that the tech industry has betrayed them, that they're no longer playing along.
00:18:52.220and you know some of the uh activism i think against data centers is driven by that you see
00:18:58.640like old school progressive activists uh i believe erin brokovich uh very old school
00:19:04.180progressive activist from the 1990s has a map of data centers around the country and is encouraging
00:19:08.800people to lobby against them so a lot of it's coming from that so on the other hand
00:19:14.060uh on the right even there are people who are concerned that tech overlords are trying to
00:19:19.880monopolize these very powerful technologies to keep all the benefits for themselves and uh and
00:19:27.440i think there's some justification there you see you see companies like anthropic for example that
00:19:33.260do seem set on you know uh on crowding out their competitors so i think there are there are
00:19:38.980legitimate concerns alongside some perhaps partisan political ones you know there's so many things i
00:19:45.200want to dig into one of the things we've talked about on the show before is how the tech broke
00:19:49.860as they're sometimes called, or the oligarchs, a critical term for them, they can be very inspiring.
00:19:56.900They're all geniuses, of course. They all have a degree of ruthlessness and charisma to succeed
00:20:05.780in that level. You've got to. And there's a lot to admire about all of them. I'm a fan of Elon Musk
00:20:12.760for largely political reasons, but I also admire his sheer determination to do things that he was
00:20:18.720told was impossible. I mean, admiring his rocketry. But not everyone, when they see an ultra successful
00:20:28.060person, has admiration and respect. Some people have skepticism. Some people say, well, what's the
00:20:34.700dark side? What's the trick? What's the other side? Well, there's just envy or there's worry.
00:20:39.540Because these are extremely powerful people, unlimited amounts of money. And some of them,
00:20:46.800Well, I think of Kevin O'Leary, the TV bad guy who's a proponent for data centers.
00:20:54.220I mean, a lot of people hate this list of people I've just mentioned.
00:20:59.320I mean, I don't know if it's even possible to sell AI when people are so deeply afraid of it.
00:21:07.460How would you make the case for artificial intelligence to someone who's going to say they're going to put me out of a job?
00:21:14.180I don't even know what I'm going to do.
00:21:16.060I don't know how I'm going to live. And even if there is some ultra wealth to go around, I'll be purposeless because the challenge of a workday, I think people need that sort of thing. How would you make the case for AI? Sorry, I'm throwing so many things at you all at once. It's just such a big issue.
00:21:35.300Yeah, well, I'd say two things to that.
00:21:37.100First of all, the people who are worried about AI being dangerous are not entirely wrong.
00:21:42.820On the extreme end, you hear the Doomer wing of AI discourse saying that AIs are like nuclear weapons
00:21:52.320because you can use them to generate bioweapons, you can use them to mount nationwide cyber attacks.
00:21:59.620And they're not entirely wrong about AI's capabilities.
00:22:02.720This is something the AI companies say themselves.
00:22:05.460But I will say, if that's what you believe, do you really want adversaries like China
00:25:33.700I will say it's a good thing that the rest of the AI industry
00:25:36.680and Anthropik are sort of at odds at the moment.
00:25:40.780And I think most of the AI industry is actually is thinking in clearer terms about this, about there needs to be a more open marketplace.
00:25:48.660But Anthropik is sort of like a black sheep at the moment.
00:25:51.780I actually wrote an article at the American Conservative right now on this very topic.
00:25:55.880They're the only company that's opposing what's called open weight models, where the weights of the algorithm, so how the AI algorithm determines certain things, inputs and outputs are open to everyone so that you can essentially run AI models on your own devices.
00:26:15.800You don't have to outsource to just one company.
00:26:21.800Anthropic is lobbying heavily against that.
00:26:41.220I think this is absolutely key because this will determine whether these technologies will be open to all or whether it'll just be a small group of tech elites, perhaps working hand in glove with nation states that control this.
00:26:55.060I think this is a critical debate that people need to be paying attention to.
00:26:58.540Because I don't think you're going to stop AI, but you can make sure it's open to everyone.
00:27:02.900And I think Anthropik's position on this is really quite troubling.
00:27:05.600Hmm. I remember when Donald Trump appointed David Sachs to be sort of an AI crypto advisor. Maybe it was one or the other, but I think it was both, actually. And I was excited by that because, you know, he had a history as a Democrat, but he came over to Trump's way of thinking on a number of things. And he seems to be freedom oriented.
00:27:27.280I actually chatted with him briefly at the World Economic Forum.
00:32:52.440In fact, some blue states have even been trying to pass laws to that effect, saying that AIs can't be used for what they call algorithmic discrimination.
00:33:02.560That's a real phrase that was used in, I believe, in a Colorado AI bill.
00:33:07.520And I think that bill said algorithmic discrimination can only be permitted when it's used to redress historical imbalances.
00:33:15.420What they actually mean is it's allowed, but only if it's discriminating against white males.
00:33:20.260So, you know, you've got to be careful about how the word safety gets expanded.
00:33:24.340So I think a Trump, back to your question, a Trumpian AI policy will have a very narrow definition of safety that's focused just on the cybersecurity element and the national security element and doesn't get into this expansive civil rights territory that, you know, where Democrats and these left-wing machine learning fairness academics want to push it.
00:33:46.500And I think the existence of Grok, the fact that it may not be the leading AI model, but the fact that it's there, that the fact that Musk says it wants to be truth maximizing, that he doesn't want to have, you know, these sort of anti-racist civil rights guardrails in it is very, very, very critical.
00:34:02.500It's important that there's not just one AI company like Anthropoc says there's going to be, but there are multiple ones that doesn't get control from the top down by very ideological forces that try and control every technology from the top down.
00:34:17.420I saw a discussion amongst various tech leaders that the Biden administration's approach to AI and to a lot of the new front lines of technology was taken as threatening and anti-entrepreneurial and basically just saying the word no.
00:34:39.680and i imagine that had kamala harris won those policies probably would have been continued
00:34:47.860and and i just have to guess here that the ai industry would not have been as well funded or
00:34:54.560developed as it as it is now and i'm trying to imagine an alternative situation where america
00:35:02.480was not the tech leader here, but China.
00:35:06.280And again, that's a similar thought process of imagine if Elon Musk for some way, God
00:37:16.680if Democrats win at the next election.
00:37:18.780And this is a very dangerous thing to be doing to someone who, like you said, is so critical for American leadership in in space launches.
00:37:26.600But I will say that on the on the on the on the A.I. point, one of the things that Anthropica is trying to do, I mentioned their opposition to open weights.
00:37:34.920They're not completely opposed to open weights. They're actually quite focused on this one thing called distillation.
00:37:41.460distillation means the use of one large language model to train another. And this is how AI
00:37:49.940companies compete with each other. They use each other's models to train on each other.
00:37:54.660But it's also like China has used distillation. They've used that to build up their own models
00:38:00.580as well. So Anthropic is arguing that distillation should be restricted on that basis. I'm not so
00:38:07.940sure that's a good idea because like i said all ai companies use distillation to trade on each
00:38:13.260other's models so if you cut that off from american companies you're not going to be able to stop the
00:38:17.980chinese ones from doing it number one um but uh you're also you're preventing american companies
00:38:23.620from from uh learning from each other as well and i i think that's another thing that is not so much
00:38:29.340a uh not so much of relevance to the china issue but it is it is relevant to a company that perhaps
00:38:36.200wants to undermine its competitors in the u.s you know almost a generation ago when a lot of
00:38:42.320blue-collar jobs were going away they were moving to mexico or being shut down i think of coal and
00:38:48.920steel i think of pennsylvania and there was a snide remark that was making the rounds in social
00:38:54.880media learn to code as in let's say you're a 60 year old man who's worked in a auto factory your
00:39:02.100whole life and you're out of a job suddenly and some kid in college said well just learn to code
00:39:07.560i mean it's it's impossible that that would happen and that such a late bloomer would be
00:39:13.160able to compete with the new kids well now and by the way that was flipped around on a lot of
00:39:18.160i mean that phrase learn to code i there's a truth within it but it's also sneering
00:39:25.680well just become an automaton but these days learning to code which basically means learn to
00:39:32.900to program computers and and other tech i don't even know if that's good advice if taken at face
00:39:40.560value because that's sort of the thing i think that ai is going to eliminate first so it went
00:39:48.160from a sneering get in line and be a cog in the machine to well machine cogs don't need people
00:39:56.300anymore if you had advice for an 18 year old so when just finishing high school on which path to
00:40:04.280go in life to be successful and not to be made redundant um which i think a lot of coding will
00:40:11.700be. What would your advice to a normal human be? I would get really good at deploying AI agents at
00:40:19.980the moment. I think that is going to be the future. It's interesting because I think the
00:40:24.920jobs least at risk of displacement are actually the manual jobs now. It's the people, it's more
00:40:30.100white collar professionals and coders, as you said, who are under threat from AI. It's a mix
00:40:36.460of people right so you mentioned earlier that you know it's hard not to be sympathetic for a cab
00:40:41.980driver who might be displaced by an autonomous car but at the same time there are some industries
00:40:46.800that would actually really benefit from disruption and we saw how social media disrupted the media
00:40:52.440industry for example that was great uh you know rebel users here because of it and we were and
00:40:58.400you know we have such a wide variety of news sources now compared to when there wasn't any
00:41:03.440digital online media uh ai might disrupt the movie making industry in the same way it might
00:41:09.440disrupt the video games industry in the same way uh it might disrupt the music industry in the same
00:41:14.400way you know just you know in the case of movies it would just like it's it's already right really
00:41:18.760lowering the cost of special effects and i think that would have a really beneficial
00:41:22.460effect on the industry by taking power away from these big hollywood studios that uh that that are
00:41:30.320the only ones currently capable of big budget productions.
00:41:33.620I think there's some very good disruption
00:41:51.040What are you gonna have AI go through them for you?
00:41:55.380And at the Foundation for Freedom Online,
00:41:58.340We're constantly looking through the labyrinth of federal grants that have gone out to, you know, anti-disinformation programs.
00:42:06.160And it's a lot easier to get all that information together, which is deliberately spread, by the way, out of across dozens of agencies and, you know, hundreds of NGOs.
00:42:16.700It's a lot easier to pass all that information when you have an AI.
00:42:20.540And that makes it harder for governments to conceal what they're doing from public view.
00:42:26.800So there are many benefits to AI disruption as well as some potentially negative consequences, you know, like you said, for the cab driver.
00:42:34.820But even on the cab driver, I'll make one final point.
00:42:38.320This is another area where you have to make a distinction between the fork in the road.
00:42:42.560Are we going to have one company controlling everything or are we going to have a more decentralized way of doing things?
00:42:48.580I know Tesla's plan is to allow people to buy autonomous Tesla cars and then have them autonomously drive around at night as autonomous taxis.
00:43:01.700So that would be individuals in charge of the autonomous taxis rather than a single company.
00:43:07.640Right. Yeah. I mean, you can already see those little food delivery robots in some cities.
00:43:13.140A lot of people instinctively hate those. There's so many videos on social media of people seeing these little food delivery robots and fighting with them. I mean, it's a one-way fight. They're just little boxes with wheels. But I understand it. That must have been how Ned Ludd felt a few hundred years ago.
00:43:33.180Well, let's just make one quick point there. That might be the best industry to disrupt food delivery, because especially in Europe, the people who are doing the low cost food delivery services tend to be illegal immigrants. So I would be very happy if that particular industry got disrupted.
00:43:53.820Well, yeah. I mean, really, so many of the motivations for mass migration, I think of Germany and the massive Turkish immigration, a lot of that was just for low-cost labor. In Canada, we have something called temporary foreign workers. It's just low-cost labor. And in Canada, too, Uber, DoorDash, it's all migrants.
00:44:16.360And, I mean, it would remove the national demand for that.
00:44:31.440I mean, it could be five minutes from now to have the first digital-only movie star.
00:44:37.900Like, I mean, AI fakes are so good now.
00:44:41.440Like just in the last year, they've got, they're so hard to tell that I don't, I mean, I bet you there's a hundred people right now making serious movies from a real script or maybe from a script generated by AI.
00:44:56.080Maybe, maybe AI watches a thousand movies.
00:45:00.500I mean, I can imagine making a prompt, watch the most successful 1000 science fiction movies or war movies or historical movies.
00:45:09.880watch it learn from them learn what what the most exciting parts were and now recreate a story in
00:45:18.820that genre and now like like you could imagine writing these prompts in ai pretty soon it would
00:45:24.860be able to do it i mean we could all be movie stars we could all be movie moguls um i don't
00:45:31.160know what's going to happen in that industry but i think that's probably underway don't you think
00:45:34.420absolutely people are always already producing whole animated shows using ai and they don't
00:45:40.260seem that different from hand-drawn animation uh they they look frankly frankly indistinguishable
00:45:45.700uh you know ai generated movie stars maybe a little bit further away but certainly
00:45:50.260certainly uh certainly special effects i think the cost of that will come will come way down
00:45:55.840uh and couldn't couldn't come soon enough in my opinion i think uh holly it would be great to
00:46:02.120take power away from the hollywood studios you know when i went to the uh meta announced it's
00:46:08.580building a 13 billion dollar um data center in alberta canada and they're building their own
00:46:15.900four billion dollar power plant right next to it which is quite an interesting combination that's
00:46:21.100a very large investment the first in canada that size and i don't think it'll be the last um and
00:46:26.840it's in a real industrial area in northern alberta near a bunch of oil refineries so there's there's
00:46:31.840no residential areas nearby there's no tourism nearby it really is in a place where you would
00:46:36.920expect to find a heavy oil upgrader or something so there's so there really was no local objection
00:46:42.980and it was approved from conception to shovel in the ground in two years which is remarkably fast
00:46:48.740for canada for something of that size i went there and i and i talked to alberta's minister of
00:46:55.080tech and innovation. And I put to him, I said, you know, what would you say to an ordinary person
00:47:01.960who's worried? Because I think a lot of the opposition to these data centers is opposition
00:47:06.520to AI. And his answer was, well, I would tell them the story of my staffer who became 50 times
00:47:15.180more efficient. Here's a quick clip of him telling the story of his staffer who became so much more
00:47:20.380efficient and useful using AI. Take a quick peek. What do you say to people who are uncomfortable
00:47:26.800or unfamiliar with AI, which is, I think most of us, taking it out on the nearest physical thing.
00:47:34.280So they can't punch back at Mark Zuckerberg in person, but maybe they can oppose this data
00:47:41.100center as their way of saying, I'm against AI. What would you say? Because I heard a lot of that,
00:47:46.180actually. What would you say to those people? Well, I would say, look, I understand that concern.
00:47:53.020I understand fear of the unknown. And when we talk about AI, a lot of folks haven't firsthand
00:47:58.860experienced how AI can make their life better, how it can make them more productive, how it can help
00:48:03.100them to earn more money, how it can take them to a more productive state. And so if you haven't
00:48:09.460experienced that firsthand, I can understand some of the hesitation. And when I talk to folks like
00:48:14.760that, I like to tell them a story about something that we've experienced inside the government of
00:48:18.260Alberta. Our goal is to become the most innovative jurisdiction in North America, the most AI native
00:48:23.620public service. And so we've invested heavily into training our personnel to say, we want to
00:48:28.840equip you with the tools and the techniques and the training to be able to use AI. And so one of
00:48:33.960my employees, his name is Harris. He's not a senior executive. He's a business analyst who
00:48:40.460specializes in contracts and he specializes in our enterprise resource planning system.
00:48:44.760And he finished our training, and then he went back to his day job with us, and he was looking at his to-do list, Ezra, and the biggest thing on his to-do list was he needed to port 1,500 contracts into the ERP system.
00:48:59.460Now, these contracts were 300, 400 pages long, highly technical, and he was going to have to map it line by line, manually, and then input that data into the system.
00:49:09.400This was going to take him 15 months of full-time work.
00:49:13.280Now, this was going to be mind-numbing, tedious work.