October 19, 2026
2026-10-19
https://www.youtube.com/watch?v=f_Ay0fhvm6Q

[0:04] We go live in ten.
[0:18] Every morning, the future moves faster.
[0:23] The models, the money, the power, and the people building what comes next.
[1:02] From Austin, Texas Texas to the world.
[1:09] AI News, Real Context.
[1:12] This is the Austin Wilson Show.
[1:14] I'm Austin Wilson, your human in the loop.
[1:23] Good morning, and welcome to the Austin Wilson Show where we're breaking the AI headlines that matter to you.
[1:29] Today.
[1:30] I'm Austin Wilson coming to you live from the AI capital of the world, Austin, Texas.
[1:35] We're testing out a new format for the show today.
[1:38] I hope you guys enjoy a little bit, uh, later of a take.
[1:42] Today, we've got a full docket.
[1:44] We're talking about a handful of different subjects that really matter to you guys in the AI space.
[1:50] We're looking at Anthropic's, uh, $48,000,000,000 fund, and, uh, Elon Musk's commitment to, uh, match or or or at least, uh, get into the sharing game, if you will.
[2:03] Uh, AI cameras on Dallas garbage trucks, chip stocks, longest losing streak since 2008, SoftBank eyes robot security guards.
[2:14] Uh, Claude is cracking down on a decades old NSA mystery and the NBA's new AI that hunts shady bits.
[2:24] Let's get into it.
[2:28] First up, we are talking about, uh, what is happening with Anthropic and their recent announcement.
[2:36] Uh, Anthropic is giving 5% of their company to a new fund for the public.
[2:42] This is worth about $48,000,000,000, uh, according to their last valuation.
[2:47] When they go public, this could be upwards of a $100,000,000,000 depending upon the valuation of the company when they when they do go public.
[2:55] Elon Musk replied on x, I recommend giving the shares directly to the public, which is interesting.
[3:01] And then overnight, he made his own pledge.
[3:03] He said, next year, I will give one of my, uh, SpaceX shares to every kid in America.
[3:10] This will be the first of many.
[3:13] Wow.
[3:14] What an exciting thing to see coming out of, uh, the richest man in the world.
[3:17] The fund could pay people directly, uh, as AI starts to impact and change how jobs are created and and people are are working.
[3:28] Anthropic has made pledges before.
[3:30] It says about 20% of the company now goes to the public good.
[3:34] Anthropic's own report says beginners are now nine percent of its tech hires, down from 27%.
[3:43] But Anthropic's stockholders still have to approve the pledge.
[3:46] The $48,000,000,000 is an estimate, and Musk gave no exact details.
[3:54] This is a very exciting thing to see coming out of Anthropic and a really interesting, uh, way for them to address, uh, how AI is impacting workers and, uh, something that, you know, amazing to see Elon jump onto it right away.
[4:09] This is getting a lot of buzz on x as well.
[4:12] And just seeing this as a a a good faith measure and a goodwill measure for these large AI, very, very wealthy AI companies to start giving back to the community and give back to people.
[4:26] Uh, and I think regardless of whether or not people are impacted from job loss or, uh, you know, incoming folks or or or people just starting out their career if they're being impacted.
[4:39] I I I think that's...
[4:40] You know, regardless of that, this is a great look for Anthropic.
[4:43] This is obviously a great look for for Elon Musk to agree to produce these funds that that would give money directly to, uh, American kids and and people impacted by these, uh, you know, maybe shortages in in workforce.
[5:00] That being said, what I would love to see is giving this money...
[5:05] You know, obviously, helping folks out that are impacted by job loss or or or or impacted natively a...
[5:12] By AI.
[5:13] This is an amazing step forward, but I would love to see this going towards education and enablement of, uh, folks that maybe are displaced to enable them to use AI tools and to become really effective and to increase the capabilities of our workforce rather than just subsidizing a loss and maybe a temporarily subsidization, uh, subsidization as well too.
[5:35] So it'd it'd be interesting.
[5:38] You know, obviously, we're gonna follow this along, and I...
[5:40] And who knows where this is gonna go and how this is actually paid out and and who gets the money.
[5:47] There is like maybe kind of a dystopian way to look at this as well too is where, uh, Anthropic has said some, uh, pretty maybe contentious things over the...
[6:00] You know, really over the years, but the past six months have have been pretty intense.
[6:06] Some of the stuff that they said, one of one of which was, uh, Dario, the CEO of Anthropic said something to the effect that Anthropic would be the only company in the world that essentially they would, uh, you know, be big enough and intelligent enough to run everything through Anthropic.
[6:23] There would be no other companies.
[6:24] So then you're kind of looking at, oh, is this, you know, just, uh, you know, Anthropic deciding who gets money, when, where, and what.
[6:34] And, you know, that...
[6:35] That's all speculation, of course, but, uh, interesting to see where that could go if you played that tape forward a little bit.
[6:43] But that being said, this is a great first step.
[6:45] I think a really good way that answer...
[6:48] Anthropic, but tech companies and AI companies in general should be looking at ways that they can give back.
[6:54] You know?
[6:55] Um, I think this is, uh, net positive.
[6:58] And then to see everybody kinda jumping on to this trend really, really quickly is awesome, and and hopefully, see them follow through with their their words and and not just, um, you know, make this a headline, but actually implement this.
[7:16] Uh, moving on, Dallas garbage trucks now have AI cameras.
[7:20] They flag homes for code violations, but the fund that helps owners fix them is frozen.
[7:26] NBC five found the cameras flagged over 20,000 possible violations.
[7:32] Every paid fine sends $36 to that fund, and it used to give low income owners up to $20,000 for repairs, but those grants were paused in 2024.
[7:44] Uh, AI lets, uh, Dallas check whole neighborhoods at once.
[7:49] Owners who can't pay are stuck.
[7:52] The city says it may restart this fund.
[7:55] Um, it helped about 20 to 30 families, uh, a little over a year ago, but the city staff, um, reviewed these flagged incidences first.
[8:04] Uh, from April to mid September, they sent about 3,700 notices and more than 80 citations.
[8:11] Still, the city won't say how much money is sitting in the fund.
[8:16] So let's let's break this down.
[8:17] It it...
[8:18] What's going on here is, um, the the city of Dallas uses, uh, smart cameras, AI on their garbage trucks to see who is violating city code and to be able to track that and then go after their folks.
[8:34] You know, really, uh, I mean, smart and innovative way to keep track of code violations.
[8:41] Um, and then they use the the fees paid by citizens to go into a fund that then that fund would help lower income citizens pay for repairs that they wouldn't be able to afford.
[8:53] The problem here is that the the fund is paused.
[8:57] So they're collecting the money, but not kind of redistributing that to lower income folks, and then using AI to kind of supercharge what's going on here.
[9:07] I, you know, I think if we zoom back, it's...
[9:10] Zoom out a little bit.
[9:11] It's interesting to see how municipalities and and cities are are going to start using AI.
[9:18] And and then just generally speaking, using AI in the physical world, uh, is really interesting to to watch and and to see how people get creative with the different use cases.
[9:30] That being said, we need to connect all the dots.
[9:33] Right?
[9:33] You you can't use this super intelligence to make things much more efficient and faster and then not, you know, back up the, uh, the promises that you made to the citizenry.
[9:45] I think that's the biggest issue here.
[9:48] Again though, I think that, you know, initially, there's going to be kind of some hiccups in implementation like this, but we'll see faster and faster implementation and more efficiency, which means, right, like, more surplus in dollars on on the back end.
[10:03] You know, they're missing tiny little violations here and there, uh, and have for, you know, as long as these, you know, this code has been in place.
[10:13] They've missed these little violations where maybe they're collecting, you know, $20.30, $50 from homeowners for for these violations.
[10:20] Number one, they get to clean up their city.
[10:23] Right?
[10:23] You know, the the the code is in place for a specific reason, and they they wanna have nice homes and nice neighborhoods.
[10:29] And so, the...
[10:31] These folks will be incentivized to, you know, clean up whatever the violation is, but then the city is able to collect on, you know, these little fines that that will, you know, add up over time, and then use that fund to go help people who can't afford to to fix their homes up.
[10:47] I think this is, you know, ultimately, if AI can make this more efficient and and increase the surplus in the city's coffers, and and then the city can help out those folks.
[10:58] It's it's a net positive, but we're gonna see these gaps as you initially implement new technology.
[11:04] I, you know, I guess we've seen this kind of with flock and and a big pushback there.
[11:10] AI cameras on trash trucks.
[11:13] Smart city or creepy city?
[11:17] You be the judge.
[11:18] Moving on.
[11:20] We are looking at chip stocks.
[11:22] Uh, they fell eight days in a row.
[11:24] That's their longest losing streak since 2008.
[11:27] Uh, on Friday, the chip index fell almost 4%.
[11:32] Every one of its 30 stocks went down.
[11:34] This morning, chips are bouncing back up almost 3%, uh, but they're still down about 13% since October 6.
[11:45] Curious to see where this takes us.
[11:47] Um, if you own an S and P 500 fund, then you own a big slice of these chip stocks.
[11:53] Nvidia CEO, Jensen Huang, told Bloomberg, all of this is going to pass.
[11:58] Two worries drove the slide.
[12:00] AI companies are borrowing heavily, and inflation came in hotter than expected.
[12:06] But zoom out, a big chip stock fund is still up about two thirds this year, and, of course, Wang sells the chips.
[12:14] Yes.
[12:15] So, of course, Jensen's going to jump in and say, hey.
[12:17] Hey.
[12:17] Hey.
[12:18] Yeah.
[12:18] Everything's cool.
[12:19] No need to worry here.
[12:21] But the the...
[12:21] There is a key point here.
[12:23] It's...
[12:23] This is physical infrastructure.
[12:25] This is, you know, this is the new age of manufacturing.
[12:30] This is not a software.
[12:32] This is not a...
[12:33] It's some sort of digital product or or service.
[12:36] These are physical goods, and a lot of that takes financing to go build out.
[12:41] And so when we see inflation increase, we see the ten year treasury note in...
[12:47] Increase.
[12:48] That that means that the cost of borrowing increases and adds, uh, additional risk to the the chip build out here.
[12:57] They...
[12:58] To go along with that, we know that a lot of circular financing is going on here where the chip companies, for instance, Nvidia is financing some of these data center build outs or helping provide funding for these data center build outs who are are are then, you know, helping to subsidize AI startups, who are then subsidizing their customer.
[13:25] So we see kind of...
[13:26] And and generally speaking, this is a general idea.
[13:29] Obviously, it's not, you know, specific, but, uh, when you see kind of this circular financing, it's it's going around where, uh, if one of these dominoes falls, does that create a big issue for the whole ecosystem?
[13:44] I think the good news here for the AI industry as a whole is we have so much demand for this.
[13:53] Unlike, you know, we talk a lot about the .com boom and bust in in, you know, late nineties, early two thousand.
[14:01] The...
[14:01] Back then, what happened was we were building out a lot of physical infrastructure.
[14:07] The demand wasn't there yet.
[14:09] People hadn't signed up for the Internet.
[14:11] They weren't paying.
[14:12] The customers weren't there.
[14:13] They weren't using it.
[14:14] They were just beginning to adopt the new technology, and and we didn't have paying customers to back up all the debt that we took on to build out the physical infrastructure there.
[14:27] That being said, this time we do.
[14:30] We have the demand.
[14:31] The demand is is substantial.
[14:34] Talking about, you know, demand for AI chips and data centers going out into, you know, 2029, 2030 potentially.
[14:47] That is substantial.
[14:49] And then we're also seeing the the other portion of this is these chips and the financing of these chips, you know, that we thought that the useful life of them was was, you know, a certain time period.
[15:02] Right?
[15:03] Um, whether that be four, five, six, seven years, whatever it is.
[15:07] Right?
[15:07] Um, we're seeing that the data centers and the AI developers are able to use these chips much longer.
[15:17] Right?
[15:17] So some of the older older generation GPUs from Nvidia are actually still useful and people are still using those, which which means the the life of these chips is is, uh, much longer than we thought, which means that you're able to...
[15:34] You know, that financing, you have a longer term financing period anyways.
[15:38] So I I agree with Jensen.
[15:40] I think this is, um, I think this is a blip.
[15:44] I don't think this is, you know, the beginning of end by any means.
[15:47] Uh, I think this is an opportunity.
[15:51] You know, obviously, this is gonna put pressure on some people.
[15:54] And and at some point, I think we are within this cycle, we are going to see consolidation and a a lot of moves, and there are gonna be winners and losers.
[16:07] But my question to you guys, are you buying the dip or are you getting out?
[16:13] Tell me in the comments.
[16:17] Alright.
[16:17] SoftBank is reportedly thinking about buying America's largest security guard company.
[16:23] The deal would be worth about $18,000,000,000.
[16:26] The company is Allied Universal.
[16:28] It has about 770,000 workers.
[16:32] Sources say SoftBank would eventually use robots for some guard tasks, Robot guards that look like human humans would would take longer.
[16:42] Okay.
[16:43] That's interesting.
[16:45] Getting into this hundreds of thousands of people do this security guard job, and and SoftBank is betting robots can do part of it.
[16:55] Um, Bloomberg says SoftBank founder, Matsyoshi Son, wants the deal to put robots to work.
[17:01] SoftBank already agreed to buy ABB's robot business for about $5,400,000,000, but it's, uh, early and the sources are unnamed.
[17:11] SoftBank won't comment.
[17:12] Nobody has said how many jobs robots would take.
[17:16] Alright.
[17:17] Uh, let's let's jump into this.
[17:19] This is, uh, kind of a different any iteration of a trend that we've seen, uh, kind of in the private equity space, the m and a space.
[17:26] A lot of folks are, uh, looking at this play or the...
[17:31] This the...
[17:31] This playbook where you can go in and buy traditionally, um, you know, labor intensive businesses that require a lot of humans and replace that with AI in some form or fashion, increase your margin, and you've got a good business.
[17:46] Right?
[17:46] And so we've...
[17:48] We're we're seeing that pretty actively happen on on probably, like, smaller businesses.
[17:56] I've heard a lot around, like, firms where, hey, we could use, you know, the latest AI models that are really good at accounting, could offload some of the work of the accountants.
[18:09] We don't need as many, you know, to hire as many accountants or they can service more customers and, um, you know, it it...
[18:19] It's just much more efficient.
[18:20] Right?
[18:20] So now we're seeing that in kind of the physical space or purportedly here.
[18:24] If they were to...
[18:25] If SoftBank were to acquire Allied Universal, uh, and then replace humans with robots, that's pretty interesting.
[18:33] This, you know, and this is a big play and the big bet that everybody in the robot space has is that, you know, we hear a lot of use cases around like cleaning the house or, you know, a personal chef or kind of these personal robots.
[18:45] But think about that scale, 770,000 employees, uh, that are security guards effectively or to to some extent or supporting security guards.
[18:55] If you could use robots in place, you know, obviously, you...
[19:00] There's a safety element there.
[19:01] Now, you know, if there is a dangerous situation, you've got a robot in place versus a human, which is favorable, but also, um, you've got, um, you know, there's this question here coming back to our initial story from Anthropic.
[19:17] Like, if we're gonna displace all these jobs, what do we do with these workers and and how do we help them out?
[19:24] This also kind of, you know, can't help but think about Robocop or any of these, you know, Terminator or or these kind of movies where you've got kind of these security, uh, robots, which is an interesting thing to think about.
[19:43] I don't know how I feel about it.
[19:45] I I think ultimately, I...
[19:46] It would be, you know, I would prefer to have a robot guard, but it it it gets a little bit dicey when you talk about protection and that kind of stuff.
[19:59] If they're smart enough and we can trust them, then then great.
[20:04] I would love to see that.
[20:05] But again, we gotta figure out the job displacement issue in that.
[20:09] Question to you, would you feel safer with a robot guard or a human one?
[20:18] The NSA put mystery numbers into a common security standard in the year 2000, and nobody could explain them.
[20:26] Well, Claude cracked them in under thirty minutes.
[20:29] Many thought the numbers hit a secret phrase.
[20:31] The the truth was actually much more boring than that.
[20:34] They came from computer clock times in 1997 and 1998.
[20:38] It's just an old random number generator that that did the rest of the work here.
[20:44] People feared for a long time, you know, twenty five, twenty six years here that the NSA hid a backdoor into these numbers for security, uh, that it now looks like is much less likely.
[20:58] Anthropic says this isn't a hack, and it doesn't weaken the security standard.
[21:02] Uh, cryptographer, Filippo Valzorda, ran the $36,864 bounty.
[21:10] Uh, he redid the work himself and got the same numbers.
[21:14] The money goes to charity, but Anthropic used an internal model that is not available to the general public, and one of the numbers is still a mystery.
[21:23] What does this mean for us?
[21:24] Basically, it's really interesting with OpenAI coming out with their their math solutions or proposed math solutions a few weeks ago.
[21:33] There was a a curious missing link in or a missing, uh, area of mathematics in everything that they released, and that was cryptography.
[21:44] Right?
[21:45] So...
[21:45] And cryptography is what we use to secure many of our digital assets.
[21:50] You think about, uh, crypto.
[21:52] Right?
[21:52] We we know what what, uh, crypto coins are, you know, Bitcoin and the like.
[21:58] That's all based on cryptography.
[21:59] Right?
[22:00] And makes it, uh, really difficult or or nearly impossible for anybody to, um, break in or or or steal these tokens or infiltrate a system in any way because of how advanced this mathematics is around cryptography.
[22:18] Right?
[22:20] One of these standards was presented by the NSA in the the year 2000.
[22:26] It has become a common standard.
[22:27] Right?
[22:28] That being said, everybody feared that there was an underlying backdoor so that the NSA could essentially get into any system that they wanted to because they had the secret key, and they could break in.
[22:42] They they kind of built that in, and nobody would know if they got in.
[22:45] That apparently is false now.
[22:48] They were able to uncover what the source of this, uh, cryptographic key was, and they've uncovered that that it was just a simple random number generator using computer clock times, which is a completely normal way to build out cryptography.
[23:06] That being said, the the cryptographic portion of the the...
[23:11] This key and the the mechanism in which it's used is still works, still great.
[23:16] We love it.
[23:17] It's perfect.
[23:18] It's gonna be, uh, perfectly safe to continue to use, but this big mystery has been uncovered, which is fun to see and fun to see, you know, Claude uncovering it in thirty minutes after, you know, another one of these where, oh, people have only been trying to figure this out for two and a half decades and Claude figured it out in thirty minutes, which I think that's probably the bigger implication here is how smart and capable the the super intelligence models have become and to see where this is gonna take us moving forward.
[23:52] Just humankind.
[23:53] Like, if the smartest people in the world weren't able to break this in twenty five, twenty six years, AI could break in thirty minutes.
[24:02] The implication on what we can use AI to to solve for moving forward is just phenomenal and and a really exciting and great thing for humanity.
[24:14] Love to see it.
[24:15] My question to you guys, are you impressed or creeped out that AI cracks secrets this fast and is this smart?
[24:28] The NBA just told sportsbooks to stop offering some popular bets, and it built an AI to spot suspicious betting.
[24:38] Bets on who scores first are gone.
[24:41] So our bets on the opening tip, and so our bets that players who barely play will come up short.
[24:47] The AI watches injury reports, social media, and betting patterns.
[24:53] This is interesting to see.
[24:54] If you bet on the NBA, your options just shrank, and the league is watching closer.
[24:59] The league says the tool will help it investigate unusual bets.
[25:04] This follows gambling cases tied to Terry Rozier, Chauncey Billups, and Jontay Porter.
[25:10] But the league and sportsbooks have used tools like this for years, and the NBA still partners with, uh, sportsbooks pretty closely.
[25:17] We...
[25:18] You know, in any kind of gambling, everybody's trying to find their edge and find their angle and get in there and what's the the sure bet.
[25:26] And we've seen many a controversy in sports betting over the years where people did not play by the rules, so to speak.
[25:35] And this is a really great move by the NBA in partnership with sports books to make this...
[25:42] You know, maybe remove ways for people to cheat.
[25:46] And some of those bets are interesting to to see the ones that they wanted to remove.
[25:54] You know, I've done my fair share of of wagering, and it's impossible.
[25:58] Right?
[25:58] Like, who...
[25:58] Who's who's gonna pick who gets the ball first or, uh, who scores first is a really interesting bet.
[26:06] And and I'm curious to see if this was...
[26:10] If they were able to surface some sort of insider information on this and and the model surface some insider knowledge, or if they just realized that those those bets were way too easy to to cheat at.
[26:25] I don't know.
[26:26] The other great takeaway from this again too, I think, is seeing the application.
[26:33] The NBA is a big big numbers game.
[26:36] Right?
[26:36] Like, any any type of sports, you know, big numbers.
[26:39] We we need to know the the stats on each of the players and the teams, and, you know, looking at TV viewership and attendance records and, you know, home team versus away team and all kinds of interesting statistics play into this, and this is, you know, data science at its core.
[26:57] And so nothing new to the NBA to to be really keyed in on the statistics and and the data.
[27:06] That being said, really cool to see an application here where they're using AI to understand, hey, maybe where is their inefficiency embedding and how can we partner...
[27:20] If we are going to do gambling, how do we partner with the sports books to make sure that they're they're creating fair bets and, um, and, you know, probably more efficiency and just a better experience for the fans and the the users as a whole.
[27:38] So really cool to see that out of the NBA.
[27:42] And I'm curious other use cases that that they're using this for.
[27:47] I, you know, I was thinking about this the other day and just we have seen this in the NFL quite a bit where it's like the, you know, they they go for it on fourth and three on the 46 yard line, the opponent's 46 yard line, whatever.
[27:59] And it's like, well, the numbers say that at this time of the game, we should do that.
[28:02] Well, what they're doing is they're running AI models and they're understanding, hey, historically, this is the right play to to make.
[28:11] And just thinking about you could get way more detailed with that and and go very, very specific in terms of the gameplay.
[28:18] And I'm sure with the NBA, it's a little bit different because you've got open play and they're, you know, kind of running consistently, but using that to optimize, you know, maybe when you sit players versus how many minutes they're playing, uh, looking at injury reports or who's hot or who's not, whatever it may be.
[28:37] It'll be interesting to see.
[28:40] Quick aside, my brother-in-law does work with a company that they look at baseball players, and they can predict when baseball players are going to get injured or or be out on on injured reserve to a pretty accurate degree, which is informs the managers of when they should sit players, how, you know, how much rest that they need, how to rotate the the bullpen, and all that kind of stuff.
[29:10] So really interesting to see this technology coming in and this intelligence coming into sports.
[29:15] When it comes down to to betting, you know, do you guys think we should ban risky bets or just let fed...
[29:23] Let the fans bet what they want?
[29:26] Ban the risky bets or or let the fans roll the dice?
[29:32] Well, it has been another great show.
[29:35] I think this is what we would call a double occasion, a first and last time that we're going to do a show this late.
[29:42] It's way too late.
[29:44] I like the the early shows better, but it just worked out with my schedule today.
[29:49] Really, you know, great stories this morning.
[29:53] I think with what's happening with Anthropic and their giving pledge and seeing Elon jump into that.
[30:01] And I'm sure a lot of other AI companies are going to participate and and match.
[30:05] And let's hope that we actually see some of those promises fulfilled and really, you you know, help people that may be impacted.
[30:14] Impacted.
[30:14] And again, let's see some more enablement.
[30:16] Let's see some training and get people in there learning how to use these tools to their benefit.
[30:24] And, uh, you know, all the way through to sports betting and how AI is having an impact in very interesting use cases from NBA sports betting to garbage trucks with super intelligent cameras on them and, uh, that and much more.
[30:41] We are here every weekday morning for AI news, real context.
[30:45] I'm your human in the loop.
[30:47] I'm Austin Wilson.
[30:48] If you're watching this as a clip, go ahead and follow and like and subscribe.
[30:55] We are at taws daily or taws.live.
[31:01] That's t a w s, the Austin Wilson show live.
[31:05] Every morning, I break down AI news at 6AM central time.
[31:11] We thank you so much for watching, and we'll see you tomorrow morning.
[31:14] Let's go.