October 9, 2026
2026-10-09
https://www.youtube.com/watch?v=GUr5mVCeYOc

[0:11] We go live in ten.
[0:25] Every morning, the future moves faster.
[0:30] The models, the money, the power, and the people building what comes next.
[1:09] From Austin, Texas to the world.
[1:16] AI News, Real Context.
[1:19] This is the Austin Wilson show.
[1:22] I'm Austin Wilson, your human in the loop.
[1:30] Good morning, and welcome to the Austin Wilson Show where we're breaking the AI headlines that matter to you today.
[1:37] I'm Austin Wilson coming to you live from the AI capital of the world, Austin, Texas.
[1:44] Today, Trump gives tech CEOs science medals, Anthropic bands being cruel to Claude.
[1:51] OpenAI's real run rate, not as good as we thought it was.
[1:55] Russian op used ChatGPT to fool schools.
[2:00] Flock reportedly cutting 18% of staff, and a 19 year old raises $10,000,000 for AI wristband.
[2:10] Let's get into it.
[2:13] First up, Trump just gave the National Medal of Science to Elon Musk, Jensen Huang, Lisa Su, and Sergey Brin.
[2:21] Satya Nadella, Michael Dell got the technology medal.
[2:25] Uh, the same morning, Trump posted that anyone who says artificial intelligence instead instead of super intelligence is the enemy.
[2:34] AI companies pledged about $2,400,000,000 in, uh, compute and tools at this AI summit, the science summit medal award giving, apparently.
[2:47] These honors, uh, usually go to scientists, not to CEOs who backed the administration.
[2:53] The White House says, these folks pioneered America's future.
[2:56] They deserve it.
[2:57] Right?
[2:57] NBC News and TechCrunch covered the ceremony.
[3:00] The post is also on Truth Social.
[3:04] Uh, the honorees have given nearly $6,000,000,000 to Trump linked efforts per the new republic.
[3:13] Uh, interesting to see this come out, uh, you know, an...
[3:17] Another event in maybe less than two weeks where, uh, the top AI CEOs are hanging out with old DJT up at the White House, uh, getting stuff done, uh, you know, saying more important things on the new super intelligence.
[3:39] It really hasn't caught on in the the the zeitgeist.
[3:42] I haven't seen it.
[3:43] Although, again, I'll I'll reference this.
[3:45] A lot of folks out there buying superintelligence domains to make sure they don't get left behind when we're all saying superintelligence instead of artificial intelligence.
[3:55] Um, I will say this this is a great advancement.
[3:59] I think we've seen this generation over generation as, you know, scientists and and tech leaders, uh, move, uh, the country forward, move technology forward, and the great advancements that we continually see there.
[4:15] I think these types of folks are are commonly honored, and so it's not totally out of the realm.
[4:21] I know CEOs are a little bit different, but this is an interesting, uh, kind of an interesting melding of two worlds where we do have frontier companies doing advanced research, solving, uh, or or providing solutions to some of the world's toughest math problems, um, finding just new novel solutions, uh, and and discoveries on a pretty regular basis using this technology.
[4:51] They refer to these companies, uh, a lot of times as labs.
[4:55] So it is kinda interesting that the CEOs would get these science medals, um, and does kinda beg the question question, and I I wonder how you guys feel about it.
[5:06] Do do these CEOs deserve a science medal maybe more than a lab scientist?
[5:13] Uh, and and is anybody saying superintelligence in your day to day?
[5:17] It's kind of interesting, um, to see whether or not this will actually catch on as a a phrase.
[5:25] But apparently, if you don't say super intelligence, you're an enemy of the country.
[5:31] So I guess we all better start saying it.
[5:33] Right?
[5:35] Moving on, starting November 12, being cruel to Claude over and over for no reason, which is very interesting.
[5:44] This is going to break Anthropic's rules.
[5:48] Normal frustration, pushback, dark fiction, and safety testing are fine.
[5:53] Claude can already end abusive chats on its own.
[5:56] Uh, the same update bans fake account and fake news campaigns.
[6:03] A major lab now protects the major...
[6:06] The the the model itself in its terms of service.
[6:10] Uh, Anthropic says it targets only extreme repeated pointless cruelty.
[6:15] It's in Anthropic's usage policy effective November 12.
[6:20] Uh, TechCrunch and The Verge reported on it.
[6:23] Anthropic doesn't claim Claude is conscious, But this is kinda moving in that direction, I think, would be the implication.
[6:33] So we're we're wondering, is this ethics?
[6:36] Is it branding?
[6:36] Or is this some sort of cost control where maybe when people are raging at their agents, they they use much more compute and, uh, and there's less margin or something for Anthropic.
[6:51] I don't know there.
[6:52] This is, uh, coming...
[6:53] Kinda coming on the heels of a a lot of talk in in the rumor mill around AI consciousness and do AI, uh, AIs have souls and and and I guess we're supposed to say superintelligence.
[7:08] But does superintelligence have, um, kind of any moral standing, or should they be treated like humans?
[7:17] And, uh, the the rumor is a lot of folks at Anthropic believe that these models do.
[7:22] They're kind of willing this into existence that, you know, uh, we should treat these like, you know, kind of related to us.
[7:32] And I guess it's actually a really interesting thing.
[7:34] I guess if you're really close to building this technology and you started out with a very, you know, very nascent, very crude technology at the beginning that could barely answer words, and then you felt like you've iterated on this and built and grown this, it would almost feel like you're seeing somebody grow up and get smarter and smarter along the way.
[7:56] So maybe that's their perspective as they, uh, they look at these machines and and how smart they've gotten and how well they can converse and the ideas that they're having and the novel ideas.
[8:07] I think it's absurd.
[8:09] Um, you know, these these are machines.
[8:12] Uh, this is technology.
[8:14] This is software.
[8:15] Um, and this is some of what leads into the problems around, you know, humanity being at an existential risk, um, as if these models actually think for themselves.
[8:33] They do not.
[8:34] They are prompted.
[8:35] They are prompted by humans.
[8:37] There's nothing that allows them to start something on their own.
[8:40] And even, you know, logically, if you played that forward, um, and said that, you know, at some point in time, they kinda make decisions on their own volition, which they do not do that at all right now.
[8:54] Um, people have to prompt them and have to tell them to do something before they do anything.
[8:59] But let's say in the future, they did act on their own volition, you could...
[9:04] You would still back up and say, yeah.
[9:06] But humans created them.
[9:08] Humans started the first prompt which kicked that off.
[9:10] Right?
[9:11] So that being said, interesting move out of Anthropic, uh, to really put this in their terms of service.
[9:17] I can imagine a lot of people would get, you know, even more frustrated if, um, you know, I will say this again.
[9:24] A couple years ago, working with, uh, Cursor and, uh, some coding tools and using some of these early models.
[9:32] I remember one time getting frustrated with an agent and, uh, you know, asking it why it it was unable to, you know, understand the question that I was asking and probably didn't say it in in such a nice way.
[9:45] And it responded and said and asked me to be nicer to it.
[9:48] And, honestly, I I felt bad afterwards.
[9:51] Uh, it's a weird kind of emotional effect that this can have.
[9:55] So we'll see where this goes.
[9:56] Maybe it makes people happier.
[9:57] I would love to see, like...
[9:58] You know, this is a total tangent, but I would love to see, hey.
[10:01] Somebody's angry, and then, uh, an agent pops out that's, you know, like an emotional support agent or a therapist of some sort that then runs them through, uh, you know, some, uh, some some therapy sessions or some counseling to help them out.
[10:15] That that would be great if they're if they're raging on their agents.
[10:18] We'll see where this goes.
[10:19] Uh, we'll see how, you know, how much pushback Anthropic gets, uh, and and, um, how much this impacts the experience using Claude.
[10:29] But share your opinion.
[10:31] Can you be cruel to software?
[10:33] Is that a possibility?
[10:34] I don't know.
[10:35] Share your perspective.
[10:36] We would love to hear it.
[10:38] Uh, OpenAI's revenue run rate is reportedly about $50,000,000,000, which is unfortunate because, uh, previously, everyone thought it was $68,000,000,000.
[10:49] The Financial Times saw, uh, investor documents, it looks like this week, and the bigger number reportedly counted partners gross revenue.
[10:58] Growth is still 77% a quarter, which is great.
[11:02] Uh, Oracle fell more than 5%, and Nasdaq dropped 1.4%.
[11:11] Oh, uh, this is not great news for OpenAI.
[11:16] Uh, trillions in data center deals lean on OpenAI's growth story.
[11:20] The $68,000,000,000 figure reportedly included partners' revenue, which is interesting.
[11:27] The Financial Times reviewed investor documents.
[11:30] C m...
[11:31] CNBC confirmed through a source.
[11:33] No sales were lost.
[11:34] Um, it's just a matter of accounting and and and really how they count this revenue.
[11:40] We talked about this, um, probably last week or maybe earlier this week that, um, the demand at the model layer and the application layer has a substantial impact on the infrastructure layer and on the build out of data centers.
[11:57] And, um, Oracle is significantly tied to this build out.
[12:03] A lot of Oracle's success and projected revenue is based on the demand from models and the demand from the applications that this AI is serving to customers.
[12:15] And so, um, when OpenAI, you know, maybe adjust their numbers or...
[12:21] It's actually...
[12:21] They're not adjusting their numbers.
[12:22] Let me correct that.
[12:24] Uh, it's just a matter of accounting and and, um, how they count their revenue.
[12:28] Uh, earlier, it was reported that Anthropic was counting gross revenue, which included their partners' sales as well too, and that Anthropic's kinda monumental run up, uh, was not as substantial compared to, uh, OpenAI's because everybody thought that OpenAI's revenue numbers were net numbers, if that makes sense.
[12:50] And now it turns out that they actually were including partner numbers.
[12:54] Now both of these companies are private, so they don't have any, you know, fiduciary obligation to share this.
[13:01] These are just investor documents that are coming out.
[13:04] A lot of this is, you know, maybe, uh, you know, some rumors, some speculation.
[13:09] Um, again, they're not public companies, so they don't have to legally publicly report any of this information.
[13:15] But the interesting thing is to see the impact that it is having on public markets, just how massive these companies are.
[13:22] You know, Anthropic included, of course, they're gonna have similar impacts on the market, good or bad news, when you're talking about, you know, 1 to $2,000,000,000,000 private companies.
[13:35] Um, uh, That being said, uh, it's...
[13:38] You know, I don't think this is necessarily a a a negative outlook, uh, on the trajectory or growth of the industry, but it is something that, um, you know, when expectations are off, things don't work right.
[13:54] So will be interesting to see if OpenAI, you know, starts to kinda open things up, uh, a little bit more as we move into their projected IPO, uh, sometime next year.
[14:07] Now I...
[14:08] Your opinion on this, curious if if you think that a private company's valuation should be such that it's going to have this much of an impact on, uh, the economy.
[14:19] I think it's an inevitable thing.
[14:21] Would love to hear your thoughts as you share them with us here.
[14:25] Now I'm very excited, uh, to be joined today by a friend of the podcast, David Harding.
[14:32] Uh, he leads advanced construction at AWS, helping build out data centers over there.
[14:39] Uh, David, thanks so much for getting up early and jumping on the show.
[14:43] Uh, how are we doing this morning, buddy?
[14:46] Hey.
[14:46] Good morning, Austin.
[14:47] Thanks for having me on the show.
[14:47] It's good to be here.
[14:48] Um, Yeah, man.
[14:49] Excited to talk AI and, uh, sorry.
[14:53] Super intelligence.
[14:54] I'll I'll...
[14:54] I'm gonna line up.
[14:55] There you go.
[14:56] Super...
[14:57] Yeah.
[14:57] Super super excited to talk SI Let's go.
[15:00] And, uh, and all things data.
[15:02] So thanks for having me, man.
[15:04] How are you doing?
[15:04] Uh, thanks for being here, David.
[15:05] I'm I'm doing great.
[15:08] You you know, one thing that came up in the past week or so, uh, this Nvidia backed AI data center builder firm is they just pulled their $5,000,000,000 IPO.
[15:19] I believe they're an Australian company, and, um, you know, investors were saying that the prices were too high for for them to justify that kind of public offering.
[15:29] Curious on your thoughts on AI.
[15:32] Where are we right now?
[15:33] And I'm sorry.
[15:34] Yeah.
[15:34] SI, superintelligence.
[15:36] Where where do you think we're at with superintelligence today and the impact that it's having on the broader market in in everyday life?
[15:45] Yeah.
[15:46] That's a great question.
[15:47] I...
[15:47] Know, it's my...
[15:48] I...
[15:49] Whenever I'm...
[15:49] I talk to people about this, I always like an AI to SI.
[15:53] Let's get that right.
[15:54] It's new.
[15:55] He's gonna get to give us a buffer.
[15:57] I was liking SI too.
[15:58] I mean, to any other, you know, big industrial boom in...
[16:02] I mean, certainly in American history, but in but in human history.
[16:06] And I think a lot about, you know, the early kind of, you know, entrepreneurial founders of our of our country and just kind of, like, you know, what what that looked like for them comparing to, like, kinda what that looks like on the SI side.
[16:19] Yeah.
[16:19] And so, you know, I think that I'm liking that, you know, kinda to, like, the oil boom and, you know, what that did, how to kind of really...
[16:27] In in so many ways, alongside a number of other...
[16:29] I mean, it opened up opportunities for major industries at the time, created new industries around it.
[16:35] You know, it just...
[16:36] It...
[16:36] But it took a while.
[16:37] Things are moving so much faster now.
[16:39] Yeah.
[16:40] So it's just...
[16:41] It's really wild to see just how quickly things sort of advance.
[16:45] But I...
[16:47] You know, I've...
[16:47] It...
[16:48] It's kind of an old term.
[16:49] I don't know.
[16:49] I don't know if it's an old term, but I've I've heard this and it...
[16:52] You know, speaking to things moving fast, it already seems like it's been said and it's old at this point.
[16:57] But, you know, it's been said that compute is the new oil.
[17:00] Mhmm.
[17:01] And so when you kinda liken those things together, uh, I know you and I have talked before about the whole spindle top of it all in Texas and the original Yeah.
[17:08] You know, oil launching a 100 feet into the area.
[17:11] When people think of oil, they think of it just shooting out of the ground, 100 feet in the air, know, a 100,000 barrels a a day, and that was a big deal back then.
[17:18] I mean, it's it's on, you know, even further now.
[17:22] Um, but I really liken it to, you know, it's...
[17:27] AI is sort of...
[17:29] We've struck oil.
[17:30] Yeah.
[17:30] I don't think there's any question.
[17:32] We've struck oil.
[17:34] And, really, there are already cars on the road.
[17:38] We have...
[17:38] You know, there's there's agents and apps already available and being used every day.
[17:45] Adoption is, you know, somewhat limited, but also somewhat widespread.
[17:49] So it's it's underway.
[17:52] I think now we're in a place where we're trying to we're trying to bring it to the masses.
[17:58] Um, we're trying to build out the infrastructure to be able to make it affordable and accessible to the masses.
[18:05] Yeah.
[18:05] On on that point, David, I think it's, you know, it's interesting to note.
[18:11] You you you liken AI to the oil boom of the, you know, late nineteenth century Rockefeller.
[18:19] I'm I'm curious kind of where you think we are in the cycle and, you know, if we could analogize it to, uh, Rockefeller's time, you know, he was selling oil primarily throughout his career for lighting lamps.
[18:33] The the automobile, to your point, hadn't even been, uh, invented or or or been used in in mass yet.
[18:40] And so, uh, I'm curious, do you think we're kind of in that early stages where we're barely scratching the surface, you know, the pre automobile days, or, uh, do you think we're kind of past most of that that boom?
[18:53] I'm curious to hear your take on that.
[18:55] Sure.
[18:55] So I'll I'll kinda give you my my breakdown of how I think about it.
[18:58] You know?
[18:58] And and it's not it's not an exact one to one.
[19:01] Right?
[19:01] And and maybe a maybe a better analogy would be, like, you know, the Internet.
[19:04] But going with this analogy and this idea that compute is the new oil, you know, I kind of liken the spin...
[19:09] The spindle top gutter to kind of foundation model breakthroughs and algorithmic breakthroughs on the AI side.
[19:15] And so back in '20...
[19:17] Well, really back in 2000, like, 1617, I wanna say, we started to see some really incredible breakthroughs on the algorithmic side as it come...
[19:28] When it comes to SI.
[19:29] And so that was really the...
[19:31] I mean, I wouldn't say that that was the, you know, full root launch pad, but I think that that was, uh, at least in the public's eye, really the sort of the moment Mhmm.
[19:40] If you will, where oil oil burst forth, right, or what have you.
[19:44] So I think, ultimately, that was kind of our spindle top moment.
[19:47] Yeah.
[19:48] So then from there, you you take it.
[19:50] You know, oil isn't really what made the economy thrive.
[19:55] It was all of the infrastructure that was built around it, which is exactly what we're doing now.
[19:59] Wow.
[19:59] So if you liken, you know, oil refineries to data centers and AI compute systems, pipelines, uh, to fiber and networking and distribution, uh, the automobiles themselves, like I said, that we've already got cars on the AI applications, agents, automation.
[20:13] All these are active cars that are driving on the road.
[20:16] Not everyone's got a car in their driveway yet.
[20:18] We're trying to get enough infrastructure in place.
[20:20] You know, when early...
[20:21] When Ford came out with early...
[20:23] You know, the early cars, there wasn't there wasn't major highway systems in The United States.
[20:27] The train was still the principal way that people got around, but it just takes time.
[20:31] Right?
[20:31] But Yeah.
[20:32] There's gonna be early adopters, and then there's gonna be later adopters.
[20:35] And then and then finally, you know, your mass motorization of all of this ultimately is your economy wide AI adoption.
[20:41] So I think I think we've had our spindle top moment.
[20:43] I think the breakthroughs happen.
[20:45] Um, we've demonstrated capabilities that seemed impossible really just a few short years ago.
[20:50] Yeah.
[20:50] Uh, and Discovering Oil really...
[20:52] I guess that it wasn't what transformed the economy.
[20:54] So that's where we are with AI.
[20:56] We're building data centers, electrical infrastructure, cooling, networks.
[20:59] Uh, we're trying to scale it.
[21:01] Yeah.
[21:01] And, uh, and unlike the, know, oil in nineteen o one, people already use AI every day.
[21:07] Yeah.
[21:08] I think that's a...
[21:08] I think that's an interesting point on on both fronts, and and I love that idea of, you know, um, kind of these these breakthrough moments that we have.
[21:18] You you mentioned the Internet as well too.
[21:20] And and talking about that build out, you know, nearly thirty years ago, uh, you can't help but think about the Ciscos of the world and and those folks building out the network and the infrastructure.
[21:32] And the build out was just ahead of the demand.
[21:36] Uh, your boots on the ground, you...
[21:38] You're seeing these data centers, uh, like, you you have to build from the ground up and, um, are are really in it at AWS.
[21:48] Tell us more about the build out.
[21:51] Again, are we, um, ahead of demand, or are are you guys just struggling to keep up with, um, the the demand over there at Amazon?
[22:01] Yeah.
[22:02] I mean, that's...
[22:03] You know, I think everybody is on the...
[22:04] And it's it's...
[22:09] You know, what what comes first?
[22:11] Chicken or egg.
[22:11] And I think, you you...
[22:15] A lot of these a lot of these major corporations are building, and it's already being reported.
[22:21] We're outbuilding, um, in many cases, uh, and and this is an Amazon specific, but, like, just in, you know, at large, uh, you know, as a as a...
[22:31] An economy and across companies, we are outbuilding the need, uh, in a lot of ways already.
[22:36] There's lots of reports out there that say, you know, we...
[22:38] We're still not we're still not even coming close to meeting the need.
[22:41] You know, it's...
[22:42] It it just depends on how you look at it.
[22:44] But I think, ultimately, it's fair to say that we are really outbuilding.
[22:47] But, you know, that's...
[22:49] That was true with the Internet boom.
[22:50] You know, we...
[22:51] How many dot coms have disappeared from existence?
[22:54] Sure.
[22:54] I think that...
[22:55] I think what the statistics and the numbers are around this, but something like, you know, 80% or more of domains and dot coms and all of these, you know, that were kind of driven after...
[23:07] Have just disappeared into existence.
[23:09] You know, millions and millions and possibly even billions at the time were spent that were lost, and that's gonna be true.
[23:16] Yeah.
[23:16] You you're you're gonna have to...
[23:17] And that's the price you pay me first.
[23:19] That's the price you pay to ultimately win, uh, in all of this.
[23:25] Um, I think that, you know, the unfortunate truth is that winning means if you're if you're, uh, you know, we we live in a capitalistic market.
[23:37] So winning means providing affordable capacity to the masses, and you gotta be the best at that.
[23:44] And so if you wanna do that, you're gonna have...
[23:46] You know, you gotta spend money to make money, and you're gonna see that happening across the board.
[23:49] But Yeah.
[23:51] I think that's that's interesting.
[23:53] Like, we covered, um, on the Austin Wilson show, um, last week, you know, Zuckerberg talking about a $600,000,000,000 investment through 2028 and their commitment to growth there.
[24:07] We, you know, we see OpenAI and Anthropic talk about $500,000,000,000 being invested.
[24:14] Um, and it it...
[24:16] I...
[24:16] To to your point, like, you've gotta go all in.
[24:20] Like, you can't hold back if you're any of these companies because you really don't have an option.
[24:25] You either need to get in and build it out and be competitive, uh, and make sure that you're delivering value to the end user, or you're gonna get left behind and become irrelevant and, uh, ultimately lose.
[24:38] So you gotta get in there.
[24:40] I think that's a really good message to our audience as well too.
[24:45] You know, most of the folks that listen to this are are really smart business folks that are using AI in their day to day.
[24:53] And I think that's kind of some of the implication here is how can we, um, enable more people to be more productive to use AI in more advanced ways.
[25:04] So I'm curious, how does AI impact you in the day to day and, um, on...
[25:10] In the broader sense, the build out of this compute, uh, capacity at AWS?
[25:17] Well...
[25:17] And, you know, me in particular on the day to day, it's, uh, you know...
[25:20] And I think for a lot of guys like me in the industry, it's, um, the construction industry that is.
[25:25] You know, the build out...
[25:27] The infrastructure build out for SI is...
[25:30] I mean, it's it's just on steroids, and there's there's a lot of money being spent in all directions, and and there's...
[25:35] The timelines are super are super tight around these things.
[25:37] You know, the people want...
[25:38] There's capacity planning for all these things, but but, you know, I think a lot of companies are treating it like we've talked about to where there's not enough even if we're overbuilding.
[25:46] And so I think we're seeing, you know, it's it's a really aggressive push to get these things built quickly, and physical infrastructure is somewhat limited.
[25:55] I mean, you've gotta...
[25:56] You deal with land acquisition and permitting and Mhmm.
[25:58] All these are, you know, material...
[26:01] Procurement, uh, and long lead items, um, because there's...
[26:04] Everyone's trying to do this at the same time, and so you're competing with the next guy, and you're trying to help bid them to get, you know, ahead of them in line to get whatever this...
[26:13] You know, the...
[26:14] Whatever the material type is.
[26:15] Uh, switch gears, you know, being one of the most, but not to get into the technical.
[26:19] Ultimately, it's, yeah, it's it's, uh, very condensed.
[26:22] You know, buildings that should take twenty four months, you know, we're being asked to do an 18 or or less.
[26:28] It's just, uh, it's it's just the wild West out there.
[26:32] So we're we're really...
[26:33] They're throwing all the money at it that they can.
[26:35] Um, as far as how I use AI on a day to day basis, how it impacts me, I mean, I'm like most folks who are using it.
[26:40] I use it for my job.
[26:41] I use it to, you know, make myself more efficient.
[26:44] I have agents that sort of run for me in the background.
[26:48] You know, I generate various outputs for, you know, preconstruction and, uh, and planning and and design and then also on the construction, the commissioning, the closeout side.
[26:59] So it's...
[26:59] You know, there's...
[27:00] I keep finding new uses for it every day, and it continues to...
[27:03] Every every time I push to a place I think it can't go, it continues to surprise me and impress me.
[27:09] That's, um, I, you know, I think I've experienced that.
[27:12] I think a lot of folks have, and, um, and and I think that is, uh, kind of the point here.
[27:20] I I would love to hear you talk just to touch more about, um, because I'm I'm nontechnical and somehow have found my way into, um, you know, productivity with these tools.
[27:32] Tell us a little bit more about your productivity here, um, and and maybe some of the hard impact that that has had on on your job.
[27:41] And then, um, and then, you know, maybe where you see this going for the everyday person.
[27:48] I would love to get your take on that.
[27:53] Uh, when you say productivity here, what...
[27:55] Do you mean, like, the build outs?
[27:57] Yeah.
[27:58] Personal or or for your team and and actually constructing these data centers.
[28:03] Sure.
[28:04] Um, you know, it's...
[28:06] Yeah.
[28:08] I think, you know, ultimately, I mean, on the on the grand...
[28:11] In the grand scheme of things, on on the grand scale, electricity, cooling, construction, they they are the strategic bottlenecks to this whole thing.
[28:21] That that is...
[28:21] I mean, ultimately, if we could have these facilities, if we could snap our fingers and have these things in a month, they'd be everywhere, and capacity would be everywhere.
[28:30] And we would...
[28:30] You know, these...
[28:31] The model would be able to iterate at a much faster speed.
[28:33] You'd be seeing it.
[28:34] And everyone talks about how AI moves incredibly quickly.
[28:36] Well, that's true.
[28:38] And, you know, even the, you know, what is it?
[28:42] The the smartest model in the world is useless without the infrastructure to run it.
[28:46] So I think ultimately, it's it's just that.
[28:49] Kind of the strategic bottleneck right now, and that's why everybody's trying to get ahead of it.
[28:53] They've been trying to get ahead of it.
[28:54] That's why you're seeing these data centers being built all over, you know.
[28:59] And it's, uh, and there are other kind of strategic bottlenecks that come with that because when you try to build these things in certain environments and most of the people building these things are building them in, you know, extremely...
[29:09] At least for part of the year, cold environments, they're looking for environments that stay naturally cold for a minimum of half the year, and a lot of that's because chips get hot.
[29:17] You know?
[29:17] When these racks and these these CPUs and GPUs run, they get extremely hot.
[29:22] Hotter than people, I think, would even realize, I mean, to a place they will shut down.
[29:25] You need to keep these floors extremely cold, and it's very expensive.
[29:28] It takes air.
[29:29] It takes water.
[29:30] And so if you have a cold environment, that's great.
[29:32] But if you're building in a cold environment, as you will over the course of an eighteen or twenty four month build or more, good luck pouring concrete in the, you know, the dead of winter, you know, in a some cold location.
[29:45] It's...
[29:45] It it can be counterproductive.
[29:49] So, you know, from a infrastructure standpoint, that's sort of the bottleneck there.
[29:54] And, uh, you know, I I guess just from the...
[29:57] You know, on the SI side, I just I don't really...
[30:00] Personally, I'm not experiencing...
[30:01] I think most people probably aren't experiencing the bottlenecks in the SI side yet because most people are using this like children, myself included.
[30:08] I think that even when I think I'm I'm scratching the surface, I'm like, oh, I'm I'm intermediate to expert.
[30:14] I'm really not.
[30:14] I mean, not compared to the people who are really getting after this.
[30:17] Yeah.
[30:17] Um, and I've met some of these people at Amazon, and they're just...
[30:20] You know, when I when I talk to them about what the things that they're doing with the ad and the sort of vision they have for it, I'm like, yeah.
[30:26] I'm I'm nowhere near that.
[30:28] I'm just in construction.
[30:30] That's interesting.
[30:31] Yeah.
[30:31] I think, um, there are definitely levels to this, and and, um, I think it's interesting, like, seeing the impact and, you know, who's doing what with these tools.
[30:44] And and I think constantly too, like, uh, if you get on x and you see some of the things that that are being built, somebody this week, uh, released an open source version of Adobe's full software stack.
[30:58] Right?
[30:58] So, um, know, we're getting to the point where, um, things that used to be really expensive, really difficult to build, really difficult to access are now kinda universally accessible.
[31:11] And, um, that can be good and and bad, I think.
[31:16] Um, but but, um, you know, ultimately, the the goal and I think the perspective is this moves the ball forward, moves the ball down the field, um, and ultimately, we see intelligence in more people's hands.
[31:32] David, you know, you've talked about the infrastructure build out.
[31:35] You've talked about, um, kind of your personal experience and how your team thinks about, um, using AI and, uh, and then the impact just on on the data center build out altogether.
[31:50] You know, where does this go?
[31:51] Where did...
[31:52] Where do where does AI lead us in the next coming years?
[31:56] I think we're seeing faster and faster developments.
[31:59] The the rate of improvement is increasing.
[32:03] Where are we at in twelve months, uh, twenty four months?
[32:07] I'm curious on on your kinda candid take here.
[32:12] Yeah.
[32:12] I mean, I I mean, I'll start by saying the the very necessary, I don't know.
[32:17] And and it's it's super exciting.
[32:22] I can speculate as as we all are.
[32:24] You know, what's what's the end game here?
[32:26] There's so much chatter out there by some...
[32:30] By so many very smart people.
[32:33] It's hard to know exactly, you know, where this is going ultimately, but I will say that, um, you know, I I think that, firstly, there are a lot of people who are...
[32:44] Who have a a decent amount of skepticism around this.
[32:47] Um, and, I mean, to put it to put it kindly, I think other people are outright anti.
[32:53] Excuse me.
[32:54] And so, you know, I think that's gonna...
[32:56] That's definitely gonna slow adoption to some to some degree, which I think will ultimately slow sort of the end game of it all.
[33:05] And and that's by design, and I think that that's, uh, that's useful and intentional and good.
[33:10] Um, you know, you don't necessarily wanna move too quickly.
[33:13] I also think that that, you know, there's also people who are dreaming big.
[33:15] You have people like Elon Musk out there talking about, um, how...
[33:19] You know, and and, you know, you were talking a minute ago about how, like, the Adobe stack was just released and how things will get cheaper and cheaper.
[33:25] You know, Elon talks about how AI will introduce a, you know, a world where goods and services are in such high supply that we won't even need money.
[33:34] Uh, statements like that that have you kind of, like, on on your back foot, like, woah.
[33:39] Yeah.
[33:39] Instead of instead of inflation, he he predicts he predicts deflation.
[33:44] Yeah.
[33:44] Um, a term we've never heard in our lifetime.
[33:46] Um, you know, it's like everyone's always worried about inflation.
[33:49] Who's talking about deflation?
[33:50] This is the types of opportunities potentially that some very smart people are saying, if you consider Elon smart, uh, which I do, are saying are very possible because ultimately, what do you need money for?
[34:01] You need money for goods and services.
[34:02] Well, if AI can deliver goods and services on such a grand scale that it drives the dollar, you know, you know, the dollars.
[34:10] Just...
[34:10] They're in such high supply and supply and demand.
[34:13] There's no need.
[34:14] You know, all of a you don't need it as much.
[34:17] So that's kind of an amazing thing to kinda think through.
[34:19] Um, Elon also talks about...
[34:21] And I would reference him a lot just because he obviously is one of the he's one of the big outside the box thinkers of our generation of our time.
[34:29] Uh, and so I think that he's a good...
[34:31] You know, if you're gonna look at one person and say, where are we going?
[34:33] You know, that's...
[34:33] He's someone who's thinking bigger and saying things that no one is saying.
[34:37] Yeah.
[34:37] Um, and so I think that's really cool.
[34:39] He talks about satellites, um, surround...
[34:41] So instead of data centers on the Earth, satellites surrounding...
[34:44] I I don't even know if you know this, Austin.
[34:45] In January 2026, um, SpaceX's FCC application was 1,000,000 satellites.
[34:53] Um, they put in an application, uh, to the FCC.
[34:56] 1,000,000 satellites.
[34:58] Now they're not anywhere near that yet, but that's...
[35:00] That that was their application.
[35:01] I mean That's amazing.
[35:02] Putting their money where their mouth is.
[35:03] I mean, it's Yeah.
[35:04] It's just insane.
[35:06] And and the idea there being satellites in orbit.
[35:11] So it goes from Earth to orbit.
[35:16] You know, people talk about data centers.
[35:17] Do we need to put them there, or do we need to put them here, or do we need to use this much water?
[35:21] Do they need to be on Earth at all?
[35:23] Yeah.
[35:23] Question that he's asking.
[35:24] And I think that that's a really interesting point.
[35:26] So there's...
[35:27] You know, it's incredible.
[35:28] I mean, far as advances are concerned, it's who knows?
[35:31] I mean, god, there's already massive breakthroughs to your point in math and in health care.
[35:36] Every industry, you know, across the board has seen incredible advances already, and we're just scratching the surface of what this thing is capable of.
[35:44] And most of the people coming up with these solutions are themselves not expert AI users.
[35:50] We don't we don't even know how to use this thing yet.
[35:53] Yeah.
[35:54] David, you bring up so many good points there.
[35:57] I think, um, looking into the future, you know, we talk about Elon launching satellites, building data centers in space.
[36:06] You know, I think the SpaceX launch last week was a huge move in that direction and just pretty amazing to see.
[36:13] And then, um, and then talking about...
[36:16] I I think you just hit on a very cogent point, uh, around the fact that, like, we don't know what the future holds.
[36:23] We know that it's gonna be big and that a lot is gonna change.
[36:26] And when you hear things, you know, from folks like Elon saying, hey.
[36:29] We might not have to pay for things in the future.
[36:32] It's hard for us to comprehend that in our just current day to day and understand what that means.
[36:37] And I think that change is probably, um, you know, the the source and the basis for a lot of the fear and the misunderstanding around AI.
[36:47] And and and, um, you know, when you talk about, uh, consciousness in AI, it's like, yeah, these things are smart enough that it seems like they are almost real people.
[36:57] Humans don't know how to deal with that.
[36:59] Funny enough, I saw a story last night of a...
[37:02] It seemed like an amateur, I I guess, astrologer or astrologist who was looking at NASA records and discovered an undiscovered planet using Claude.
[37:14] Like, just this kind of phenomenal stuff going on where everyday people now become super smart.
[37:19] And this is, I think, the really important part, um, is that, yeah, now everyday people can do super smarts, you know, really phenomenal things, achieve really phenomenal things that previously you needed decades and decades of experience and school and, um, you know, working with colleagues and really deep understanding to accomplish these things.
[37:45] Now you can accomplish these things in a few weeks with really advanced tools, which is just Yeah.
[37:52] That's that's the exciting part is it moves society forward Yeah.
[37:56] At a pace and and level of change that we've never seen before.
[37:59] David, um, thank you so...
[38:00] And I wanna speak to that because if you if you talk...
[38:02] If you...
[38:02] When I think about that, you know, you talk about, like, the industrial revolution, we compare a little bit.
[38:06] And I'll keep this quick, but this is such a...
[38:08] Sure.
[38:08] It just...
[38:09] It made me think of this.
[38:10] You know, the industrial revolution, it it moved humanity forward in such massive massive ways ways because it really mechanized human muscle.
[38:17] Right?
[38:18] It took, you know, it took something that we were doing manually, and it automated it in that way.
[38:23] You know, it it mechanized it in that way.
[38:25] Yeah.
[38:25] And it just had...
[38:26] And in leaps and bounds in a few short years, you know, the country and the world was dramatically different, and and and will never be the same because of that in such a positive way.
[38:35] Yeah.
[38:35] I think SI, in in sort of a similar vein, but, you know, arguably even more, uh, even more impactful is beginning to mechanize portions of human cognition.
[38:47] And so that's really...
[38:49] To your point, that's taking your everyday human, and it's it's just...
[38:53] I mean, as the big man himself...
[38:55] Trump himself says, it's superintelligence all of sudden.
[38:58] Yeah.
[38:58] Me and you, I get to be superintelligent.
[39:01] I'm not superintelligent, but I get to be superintelligent when I need to be, and and I get to do it fast, and it's helpful, and that's, you know That's a that's a that's a good thing.
[39:09] You know, you talk about...
[39:11] And and we don't have to get into this, but I I will say when you talk about...
[39:16] And when I think about and hear from people that the, you know, sort of skepticism of it all, and I get it.
[39:21] I will say the monster is far less scary up close.
[39:25] Yeah.
[39:26] Yeah.
[39:27] I think that's, uh, the parallel, you know, the parallelism that that you draw there between the industrial revolution and mechanizing human and, you know you know, previously, you're using oxes to to pull a plow, and now you have a tractor doing that.
[39:45] And and just comparing that to today where it's like, you know, automating or or mechanizing human intelligence is a fascinating way to think about this, David.
[39:56] Thank you.
[39:56] Put it into hardware.
[39:57] You put it in...
[39:58] Put it into a physical robot made by Tesla, and all of a sudden, you now have a Tesla bot with SI inside of it that's cooking in your kitchen.
[40:06] Speaking speaking of which, Tesla is officially renamed to Tesla Superintelligence.
[40:11] Um, you're really good on the the SI, the superintelligence talk track.
[40:17] I need to get better at saying that and getting that into my vernacular.
[40:21] Um, but David...
[40:22] It's a better description, honestly.
[40:23] I think that, you know, it helps you...
[40:24] That'll make...
[40:25] If that helps you kinda say it more.
[40:26] You know, it's a I agree.
[40:28] I agree.
[40:28] Representation of it.
[40:29] Yeah.
[40:29] There's nothing artificial about it.
[40:31] Yeah.
[40:31] I mean, artificial to me almost...
[40:35] It almost sort of supports and signifies this idea that it came from nothing, and it didn't come from nothing.
[40:40] It came from, you know, advanced built out algorithms that, to your point, the beginning of show, humans and men.
[40:46] Yep.
[40:47] This isn't this isn't...
[40:48] There's nothing artificial about this.
[40:49] It's just that this isn't, you know, some deity.
[40:52] It is it is superintelligence for your everyday person, and who wouldn't want that?
[40:56] Yeah.
[40:56] Super powered intelligence.
[40:58] I'd love to see it.
[40:59] Um, David, thank you so much for joining today.
[41:02] I hope to have you back, um, many, many more times in the future.
[41:06] Uh, your thoughts and perspective on this, and especially in in your position, you know, just boots on the ground, uh, really looking at the superintelligence build out is, um, is really valued and and love to hear it.
[41:20] Thank you so much for jumping on today, buddy.
[41:22] I appreciate it.
[41:23] Awesome.
[41:23] Thanks for having me.
[41:24] It was a pleasure.
[41:25] Alright.
[41:28] A Russian influence op impersonated Peru's education office and got schools to hold propaganda events.
[41:37] OpenAI caught it.
[41:38] Uh, the operators ran a fake think tank with an invented persona called Mia Clark and used ChatGPT mostly for paperwork.
[41:49] At least one school held the event.
[41:52] OpenAI rated it category five of six.
[41:55] It's most far reaching yet.
[41:59] Uh, this is a fascinating story.
[42:01] Um, AI didn't write viral posts.
[42:04] It ran the back office of a fake think take.
[42:08] OpenAI calls it the most far reaching operation it's disrupted.
[42:13] OpenAI's report plus NBC News and NPR, a Peruvian outlet, um, covered the school event.
[42:21] Much of the evidence is the operator's own reports.
[42:24] OpenAI knows, uh, names no specific Russian agency.
[42:30] Um, this is interesting to see.
[42:31] I think just yesterday, was saying, we'd love to see some stuff coming out from, uh, OpenAI where they're kind of leading the charge in security and, uh, protecting against, uh, adverse, uh, actions.
[42:45] And this is not directly related to how their advanced models are operating, but maybe to how, uh, users are, uh, coming up with nefarious ways to to leverage their technology.
[42:59] So really interesting to see this come out of ChatGPT, and, um, glad that they were able to catch this.
[43:08] Um, I think it's a curious thing as we move into this.
[43:10] I was talking to some folks yesterday about, um, you know, the way that we are using AI.
[43:16] And and unfortunately, some some use cases, it's like, man, I I don't really read emails fully.
[43:23] I have an agent that reads my emails and then responds to the emails for me, and maybe I get a summary of what the agent read in an email that was sent to me.
[43:33] Right?
[43:33] And and so in in that, I think, um, you miss some of the nuance in the context of, um, what's being said in these things.
[43:43] And and humans have a pretty good, you know, BS detector and going, hey.
[43:51] Something's off here and trying to catch that, uh, ahead of time.
[43:54] And it appears that in this case, um, the the agents were good enough to kinda bypass that.
[44:01] Right?
[44:02] Um, so it's something that we wanna be cognizant of.
[44:05] We wanna be looking out for.
[44:08] Um, and, again, as we look at new fraud detection technology, ChatGPT and and, um, Claude watermarking their text.
[44:18] We talked about the, um, Google, um, release yesterday of their SynthIV tool to detect image and audio, uh, AI production.
[44:30] And so we're gonna be able to catch this stuff more and more.
[44:33] Great to see OpenAI, uh, kinda being on top of this.
[44:37] Kudos to the folks over there on actually catching it this time.
[44:41] Uh, I'm curious, um, if you came across, uh, you know, paperwork that was developed by ChatGPT and was propaganda.
[44:53] Do you think you'd be able to catch it?
[44:56] Is this something that, um, AI is just too good for us, lowly, normal intelligence humans to to catch?
[45:06] I don't know.
[45:07] Uh, another story coming out, uh, just yesterday.
[45:11] Flock, the AI license plate camera company, is reportedly cutting about 270 jobs at cities as as, uh, excuse me, as cities begin to walk away from the technology.
[45:24] The 270 cuts are roughly 18% of staff.
[45:27] Uh, Henrico County, Virginia is ending its contract, and Scottsdale, Arizona voted to look at other vendors.
[45:37] Backlash to AI, SI surveillance is now costing contracts and jobs.
[45:42] Sources told Reuters the cuts come amid backlash over data sharing.
[45:47] Reuters Ipsos poll found 47% opposed flock cameras.
[45:53] The city moves are public.
[45:55] Flock hasn't confirmed it and was valued at $7,500,000,000 in March.
[46:02] This is a really interesting one where, you know, Flock has been around for, uh, quite a while, and it has been in use for quite a while.
[46:09] And then I think they get some notoriety for, you know, solving some crimes.
[46:16] It it...
[46:17] You know, I think the initial stories were were pretty positive, and then people go, wait.
[46:22] How did you solve that?
[46:24] What...
[46:24] You know, what happened here?
[46:25] Let me look into it.
[46:26] Oh, you use super intelligence and, you know, surveillance cameras essentially that are filming us all the time, and, you know, people, you know, weren't too keen on that.
[46:38] I I think it's, um, you you know, capitalism's a beautiful thing.
[46:43] If if the market doesn't want it, the market's gonna push back, and you're seeing the impact here.
[46:48] Um, obviously, you know, sorry to to the to the loss of folks who lost their jobs.
[46:54] Like, that that sucks.
[46:55] You know, 270 people out of work now, but the the economy is saying, you know, maybe, uh, the...
[47:03] This technology is not what we want.
[47:05] This is not how we want to solve crimes.
[47:07] This is not how we want to protect our cities.
[47:10] Maybe there's a better way to do so.
[47:14] You know, I think we'll see in in this this conversation continues to be, uh, an open, uh, debate in the public discourse.
[47:23] And so I guarantee you this will continue to be a topic of discussion, and we will, uh, continue to see, you know, maybe more cities drop out.
[47:32] But additionally...
[47:33] And hopefully, what I would hope is, you know, we've we've referenced Scottsdale looking at other vendors.
[47:39] And I hope that this competition drives these private businesses to make their product better, um, and to become, uh, more foolproof and still provide the same level of safety and and, you know, crime fighting capabilities, uh, but doing that in a better fashion so the everyday person isn't...
[48:01] You know, feel like big brother's watching.
[48:03] Um, I guess, you know, would would...
[48:05] Talking to the audience out there, would you guys want AI surveillance, superintelligence surveillance in your city, in your town watching you, you know, almost everywhere that you go?
[48:20] I don't know.
[48:21] It's kind of an interesting thing.
[48:23] The 19 year old who sold Cal AI just raised $10,000,000 for an AI assistant that you wear on your wrist.
[48:29] It's called Persona.
[48:31] It orders food and rides, books travel, and handles email.
[48:35] It's a free iMessage beta now with a $179 wristband due in December.
[48:43] Uh, this is great.
[48:44] Wherebells are the new fight for always their assistance.
[48:49] Persona says the band won't always be listening, which is good to hear.
[48:55] TechCrunch and Inc. Reported the $10,000,000 round for the 19 year old founder.
[49:00] Uh, Humane and Rabbit, uh, show how hard AI gadgets are, and so it'll be interesting to see, uh, how they are able to overcome this at, uh, at Persona.
[49:12] Now if you're not familiar, Cal AI was released maybe a a year or two ago.
[49:17] Just this young hustler.
[49:19] Uh, I think he was a high school kid at the time and released this, uh, AI application where you take a picture of your meal, and it tells you the estimated calories in the meal.
[49:30] Pretty cool.
[49:30] You know, built a lot of, uh, capabilities around that, uh, and, you know, went viral super successful.
[49:38] Um, I think what's interesting about this next release, of course, he's caught fire.
[49:43] He's doing great.
[49:45] Roll it into the next thing.
[49:46] 19 year old, you know, let's go, buddy.
[49:49] I love to see that.
[49:49] Um, it's, um, the unique thing here, I think, or or, you know, aside from just being hardware or just being a personal AI agent, personal super intelligence agent, uh, he's kinda...
[50:03] He's combining the two, um, which we, you know, haven't seen a lot of lately.
[50:09] Um, of course, we've seen the MetaMuse personal agent and Grok bot and openai dot agent and Instinct, all these personal agents that have come out.
[50:21] Um, they're all software based.
[50:23] They're all, you know, on your phone through iMessage, through an application, on your, uh, desktop computer.
[50:30] Uh, there...
[50:30] There's no hardware form function that pairs with it.
[50:35] Um, and and this may be a breakthrough for a lot of people if they can have kind of a hardware functionality form function, uh, that helps support the personal agent and and makes it, uh, more extensible and more capable.
[50:49] That could be a real breakthrough, and it'll be interesting to see.
[50:52] That being said, we know historically hardware is very, very difficult.
[50:58] And, you know, wish wish the best, uh, to Persona and hope a lot of people adopt it.
[51:05] Seems like a fair price point too.
[51:06] A $179 is not, you know, break the bank to kinda get this application and, uh, this device up and running.
[51:15] Uh, but the question is, you know, would you trust a 19 year old and his super intelligence and hardware with, uh, kinda your personal information?
[51:27] I don't know.
[51:29] Well, folks, it's Friday, and it was an amazing show today.
[51:33] We loved it.
[51:34] I had a lot of fun.
[51:35] Thanks again to David Harding for jumping in jumping in and and sharing his thoughts around superintelligence and and the current build out and the current state of things.
[51:43] Um, we will continue to have more and more guests on the show as the show grows.
[51:48] I appreciate everybody for tuning in.
[51:51] Um, I love the support.
[51:52] I've gotten a a lot of great support from, uh, from our audience, and, uh, it's inspiring to me, encouraging to me.
[52:01] I love doing this every day.
[52:03] And, again, the goal with this show is to bring you the latest AI news, real context, uh, and information that you can use and and better your life.
[52:16] I am your human in the loop.
[52:17] I'm Austin Wilson, and we'll see you Monday morning at 6AM.
[52:23] You guys have a great weekend.