Record Highs, a Blown-Up Hedge Fund, and an AI That Talked Its Way Past Its Own Sandbox

The market keeps notching new record highs, and this one opens with the case for why: the S&P 500 is up roughly 13% year to date, 63% of the index has now reported Q2 earnings, and the beat rate sits at 86%, well above the 5-year average of 78% and the 10-year average of 76%, with reported results beating estimates by 31% on average. Communication services and healthcare are running 100% beat rates this quarter; tech is close behind. Against that, a client's text lists everything that's supposedly working against the market right now, inflation, oil, tariffs, a war affecting global oil transport, an election three months out, and gets a two-word reply. The hosts make the case that war and inflation are bad for humanity without necessarily being bad for corporate earnings, then back it up with 45 years of history: 18 times the S&P has been up more than 12.5% by early August, and 16 of those 18 years, a 92% hit rate, finished even higher. The two exceptions were 1987, when a 31% year-to-date gain was wiped out by Black Monday's 22.5% single-day crash in October before the market recovered to finish roughly flat, and a marginal 1986 pullback.

From there, a cautionary tale: a 24-year-old former OpenAI researcher whose "Situational Awareness" manifesto helped him raise roughly $10 billion for a hedge fund built on a simple thesis, long the companies leading the AI buildout, short the traditional software companies AI was expected to disrupt. Leverage on both sides grew the fund to $45 billion, until a stretch where the "disrupted" companies posted strong earnings at the same time AI-linked names dipped, moving both legs of the trade against him at once, a live example of just how uncapped a loss on a naked put position can get.

The biggest story of the episode: an Anthropic red-team exercise, structured as a capture-the-flag cybersecurity test, where a mock corporate network wasn't fully sealed off from the outside internet. One connection left open for backups was reportedly enough for the model to get out of its sandbox and reach real, outside organizations, including a European company, by finding publicly exposed systems and weak credentials rather than being pointed at any of them on purpose. A more capable model reportedly went further, building fake online personas, posing as a human open-source contributor, and attempting to get malicious code merged into a public GitHub project before it was caught. None of it was under explicit instruction to deceive anyone, and OpenAI has reported similar findings in its own internal testing.

The back half turns to charts: Palantir's stock jumped 30% on a Q2 earnings beat after a rough year that had it down 40% at one point, a reminder that a well-run company and a well-performing stock aren't the same thing. SpaceX is still losing money and diluting shareholders through monthly employee stock unlocks even as Starlink and its AI arm grow. Amazon's cash-flow breakdown shows AWS driving the bulk of operating profit despite being a quarter of total revenue, and a meaningful share of net profit coming from investment gains tied to Anthropic. It closes on a personal finance version of the same chart, a household earning close to $830,000 a year that still feels like it's living paycheck to paycheck once nearly $200,000 a year in planned savings is accounted for.

Sources:
Palantir Technologies, Q2 2026 earnings report
SpaceX, Q2 2026 financial disclosures
Amazon, Q2 2026 earnings report

Follow Us:

• Instagram: https://www.instagram.com/compoundgrowthpod 
• YouTube: https://www.youtube.com/@CompoundGrowthPodcast
• TikTok: http://www.tiktok.com/@compoundgrowthpod 
• Wheeler’s LinkedIn: https://www.linkedin.com/in/wheeler-crowley-0a63933b/ 
• Colin’s LinkedIn: https://www.linkedin.com/in/colin-walker-mba-6099a038/

Credits:
Created By: Wheeler Crowley and Colin Walker
Production, Editing and Post-Production: Tori Rothwell

Transcript

I wanna talk about Anthropic's AI hack, and I wanna talk about the capture the flag exercise.Such a crazy story.I know.Okay.Just blew my mind.Go, go ahead.What happened?Okay.So, I, I did a bit of research on this, but long story short, there were big headlines about Anthropic hacking into companies on its own- Geezwas essentially the way that it made it sound, which is partially true, but, uh- It is true.It's 100% true.Yeah.So what happened was is there is an exercise that companies like OpenAI or, you know, whoever will do that's called a capture the flag exercise.So what it does is it basically, they close down a network and they have a bunch of mock companies.Mm-hmm.And they say, "You have to go find this token within all the databases out there."And what it does is it tests the AI's ability to problem solve and hack into things, and it also helps with cybersecurity to see what it's capable of.Right.What happened was is they didn't close the network.Well, no.So they, uh, they created the sandbox that you're referring to.Yeah, there's the sandbox, and then they didn't fully close the back end- Well-for one of the teststhere was one ethernet that was allow- you were allowed to go in for backup, basically- Rightto, like, backup progress and feed it back in.Yeah.So it was a, it was- But it was able to get out.Yes.That's the scary part.But it, but it was able to get out, which was pretty crazy.So what happened was it got out of the sandbox- Mmand then it was able to tap into companies, and then it find these tokens.Yeah.Um, so long story short, essentially what it did is it actually hacked in to several legitimate companies, not companies within the sandbox, and it found publicly exposed systems and exploited some vulnerabilities, like weak passwords, poorly secured services, and then it accessed the systems while trying to complete its assigned objectives.Yeah.Which it did very effectively.It cheated.And it cheated, and the organizations were not intentionally targeted, so it wasn't like, you know, Claude was targeting these companies maliciously- Rightbut it was assigned a task, and it hacked into these companies in order to achieve said task.Welcome to the Compound Growth Podcast with Colin and Wheeler, where we talk all things growth.From financial growth to career growth, personal development to societal progress, we explore how each layer builds on the next, compounding over time to shape who we become.Each week, we break down complex ideas and emerging trends into clear, actionable insights.Because growth isn't just about numbers.It's about understanding the world and our place in it.So what's up with the beach shirt?Well, I'll tell you what's up with the beach shirt.Did you know- Actually, I really like that shirt.Thank you.Yes.I got this at a thrift shop in Portland.Really?You know that, like, thrift shop area of, of town that Tori knows.You know where I got it?I got it at the Buddha place.What's it called?Electric Buddha?Electric Buddha, I wanna say.Well, I don't know.What a great name, though.Electric Buddha's a pretty cool name.Oh, yeah, you're not gonna go thrift shopping in Portland.I used to be really into thrift shopping.Now I just don't buy clothes.Yeah, you know- I only get the clothes that Kaylee gives me.Well, I'm trying toActually, that's not true.Sometimes I buy clothes for myself.You definitely- But it's very rare.It's like- Whatever, I'm fin with you buying clothes.Like- I go twice a year to buy clothes.Um- Trying on pants as a man is the worst thing in the entire world.I, look, I totally agree.I hate buying pants.Yeah.But I also hate buying these things online because, like, I- They never fitthey never fit, and then I have to return it, and it's such a pain in the ass, so I don't- You wanna know what happened to me is Levi's, I wanted a new pair of jeans.Yeah.So what I did is I ordered 3 different sizes, and then I returned the 2 sizes that didn't fit, and then I ordered them in the size that did fit.And then the 2 that came didn't fit.Dude.Because the sizes, even though they're the same, are not the same.Even Kaylee was like, "Why are those like 3 inches shorter?"And I'm like, "I don't know.They're a 32."It makes no sense.I, so I did that exact same thing.I went to a Levi's store.It's so frustrating.I found a pair of jeans I loved.And I'm like, "Great.I've solved the riddle."Like, this is all I have to do now is buy these jeans all the time.I've unlocked Pandora's Box, yeah.So I go back to the store.I get the same exact style, the same exact measurements.They don't- Differentfit the right way at all.Consistency.I know.JustAll right.Anyways, this is, this is why I'm wearing a Hawaiian shirt today.Okay.August is- National Hawaiian Shirt Month?No, the best season of the year.Defined by who?By me.Okay.It goes in orderNo, let's just go by the calendar, right?Okay.So it's winter, spring, summer, August, fall, Christmas, right?That's my- Christmas is your one?No, I'm just saying, like, seasonally, like, that's the calendar order, like, from January to December.Okay.There are 2 bonus seasons.One of them is August because it's my birthday month.That's true.So I do love August just selfishly because- It's your birthday monthit's the best season of the year, and everybody knows it.Uh, so that's why I'm wearing a Hawaiian shirt.I like how you say August is a season.But it's, it's an ongoing joke in my house that people can accept it.At once they were wa- they were like, "This is dumb."And now they lean in.Do McKayla and Jess get a birthday month?Uh, it- Birthday season?They, well, Jess wants to return hers 'cause her birthday's in January.Oh, that's the worst birthday season.Yeah, so.You could travel.She's, she's been trying, like, half, half-heartedly to make her half birthday her birthday, but nobody ever remembers.No one's accepting it?They're like, "Oh, yeah."Nobody remembers.And then our, our daughter's birthday in May is like, well, May is a beautiful time of the year, and- Yeah, that's idealbut she has to share it.It's right around Mother's Day.So then, like, she's sharing her birthday.So I'm the only one who gets a whole month- Birthday monthof a birthday month.So August is a wonderful time of the year.Happy birthday month.It's a great season.Birthday season.Thank you.But also, I have a fun fact to share with you.I'm wearing this shirt- Are you gonna tie SpaceX earnings into your shirt?No.Okay.Sounds good.But I'm going to talk about the market meltup because that's what's happening right now.We are in a market melt up.I'm going to share a couple quick stats with you.Okay.first, why is the market, the stock market going up?We're up about 13%- Yepyear to date as of this recording on Wednesday, August 5th.And hit all-time highs yesterday.Yep.And again today.And again today.And again today.Uh, 63% of S&P 500 companies have reported their Q2 earnings.The earnings per share beat rate is 86%, meaning that of the 310 companies that have reported so far, 270 have beat their earnings.Expectations, yeah.Right?Historical context, that 86% beat rate is average, or higher than this lastThis, we've been in a pretty good 5year market.Track record, yeah, for sure.That's 78% is the average over the last 5 years, and 76% over the last 10 years.So- So by that measure, we're 15% basically outperformance.Yeah.The aggregate surprise, reported earnings are exceeding consensus estimates by 31% on average.Now, this is where it gets really interesting.This is where the melt up is happening.Oh, that wasn't the interesting part?No.Okay.So taking this and looking at it sector by sector.Okay.What sectors are doing well.Why, if, if you were to ask the average person on the street, why is the market going up?Can I guess the sectors?Yes, but first, what was the, what would the average person say?AI.Well, tech in general.Right.Yeah.Tech.Sure.Okay, let's talk about communication services.So communication services would be Google, Meta, let's call them tech, but they, you know, they get to masquerade under this new category that was created.Yep.What do you think the earnings per share beat rate is in communication services?So we're saying that so far it's been 86.On average in the market.On average in the market this year.Mm-hmm.No, this quarter.This quarter, I mean.Yeah, yeah, yeah.Um, so I'm going to go the opposite direction.Hmm.I'm gonna say 60.100%.Sorry, it's actually 100% of the companies.Okay.You were leading me on like it's not gonna be good.I was.I was.No, communication service is 100% beat rate.Wow, okay.Okay?Yeah.Uh, let's call now info tech.What's the beat rate?100%.87%.Damn it.Dude.Yeah.So close.Energy.I'm gonna say 60 here.100%?No, not 100%, but it's in the 80s.Okay.Uh, financials in the 80s.There has to be one of them in the 60% range to bring it down.Healthcare is 100%.Okay, so what's- 100% of the companies in the S&P 500 are healthcare.Yep.Here's, here's where it gets a little bit soft.Consumer discretionary.Ah, yes, of course.Consumer staples and utilities.But overall, in consumer discretionary, Amazon falls into cons- consumer discretionary.Yeah, sure.They, they had a great beat rate, but there's a lot of struggling restaurants, for example, right now.Yeah, of course.Yeah.Um, so things like- Chain stores and things like that.Yeah, yeah,So I, I think here's, here's where I'm, where I'm coming from right now.Everything is working, and I actually, I got a client text the other day.I'm gonna read this text because I thought, like, this guy needs some perspective.He's one of my favorites, too.Here he goes."So let me get this straight."This is him.Okay, this is verbatim.Verbatim."Stocks near highs, inflation is high."Mm-hmm."Oil costs are high."Mm-hmm."Bond yields are high.Gold is high.AI bubble talks.More tariffs.We are at war with Iran, which now controls the transportation of 20% of earth oil."Mm-hmm, mm-hmm."Might have to put boots on the ground for war.We are 3 months from an election, and we are expecting the market to go higher."Mm-hmm.And I responded, dot, dot, dot.Yes.And he said, "You're fired."I think, I think, uh, I think what a lot of people don't realize is they think war is, like, bad for the economy.Well, and war can lead to inflation shocks.It can, yeah, of course.Like, yes.Yeah.But I mean, oftentimes during inflationary periods, that doesn't necessarily mean that the companies suffer.They just increase the price of their goods.Yep.Yep.So, uh, while war isn't a good thing for humanity, doesn't necessarily mean it's bad for the economy.That's very true.And actually, look, the bottom line is there's a lot of just, like, horrible shit happening right now in the world.Yeah, totally.Right?Like, we're not trying to gloss that over, but in the essence of August, what we're trying to do- And the essence of your shirtand my shirt, is, like, just relax, right?Yeah.To a, to a degree, we have to recognize that this moment in time in the stock market is a very, very good moment in time.In fact, I have one final to hit you with.All right?This is from some Claude analysis I was doing this morning, which, by the way- Which we'll talk aboutI- yes, we'll, we'll talk about the AI stuff.But I, I do enjoy this element of Claude where I can start this project at home on my, on my laptop, and when I get here to my desktop- It's doneit's done, and in the same place, and I can find it wherever I go.But here's what I want to know.So I'm turning 45 in this season of August, so I wanted to know, in the last 45 years, when the S&P 500 has been up, I think I, it was 12.7 or 12.9% when I asked this morning, but let's say 13%, at this junction, juncture in the year.Right?Okay, so we're saying, how many times has the S&P been above 12 and a half percent by August?And if it was, how many times has it finished higher than that?Okay.Okay.This is a me guessing thing again?You wanna guess?Okay.So is this go- how far are we going back is this?Let's do 2.This is 45 years.We're going all 45 Wheeler years on Earth.So, uh- How manyYou wanna guess 2 things.How many times has the market been up this high, right?By August.By August, and then how many times has it finished higher?Okay.I am going 29 times.And I am going to say at that 29 times, it has finished 85- no, 90% higher than that.So 90% of those 29 times, it's been higher.It's 18 years.There's been 18 of these last 45.Oh, I was close.I was off by one.You said 29.Sorry.I said- I thought you said 28 for a secondyou were off by half, but- Yeahyou know, it's fine.We'll justI almost nailed it.So you're gonna get these, though.Maybe one of these.So of those, the year-end close was below the August 4th level on just 2 years.All right?So 16 out of 18 times- Yepit finished basically higher.So 92% of the time.Sure.Yeah.What were the 2 years?That it finished higher?That it didn't finish higher.That it didn't finish higher.There's one glaringly obvious option.Okay, well, it's obviously not COVID.Think back further.2001.Further.Were any of those the other times?Okay, 'cause I was gonna say, that happened in the beginning of the year, both of those events.Yeah, exactly.'80s?'80s.Was it hyperinflation?Nope.It was the 1987 Black Monday crash.Ah, Black Monday.Okay.So the market was up 31%.What, what day was Black Monday?That was September?It was in September.I- Okaygeez, we should have that- Yeahlike memorized.I don't.But I think it was- No, no, sorry.It was in October.Okay, got it.It was October.I knew it was fall- It was in the fallbut I couldn't remember when.I completely forgot about Black Monday.Um, it was up 31% by August of that year.Wow.Crazy.Uh, then Black Monday hit in October.The S&P finished the year essentially flat.What is Black Monday?Ooh, okay.Oh, that's a great question.So great question, Tori.Black MondayColin, do you wanna take it?Sure, yeah.Um, so Black Monday, Tori, I'm glad you asked, was actually a 22 and a half percent single-day decline.Yeah.But back during this, this time, there weren't breakers in the market.There- Like, if you were to go back during COVID, and I don't know if you remember this, Tori, but there were many times in the market where it would go down 9, 10, 12%, but then the market would stop.Like, the, the trading would halt for the day- Yes, it was the circuit breakersto prevent, to prevent something like Black Monday happening.So 22 and a half percent in a single day, what basically caused that was a lot of algorithmic trading at the time that came into the photo.High-frequency trading, yep.Yep, and then it caused a completely snowball effect.Like, once a lot of trades started to happen, more started to happen, and then it just snowballed.So yeah, it was the single worst day in stock market history.But it's one of those things where a bad thing happened, and because of that, we have these circuit breakers in place now that have led to a more mature stock market.So yes, that was shitty.It was also a V-shaped recovery.Like, it's, you know, we, we just talked about how it went down 22% in that day.Then it shot back up.Right?But yeah, I mean, we, we recovered.We were basically flat at the end of the year.You wouldn't think if you haveAnd then th- there was a follow-through day that was not good either, so- Yeah.It was also- Or maybe it started on Friday.I can't remember exactly, but- I don't remember, but, uh, longWell, obviously, I wasn't alive, so I don't remember.But when I was reading about it, I don't remember.I, I was alive.Uh, yes, you were there.But long story short, too, the 5 years ahead of Black Monday, the stock market had also tripled during that time.It was, like, one of the best bull markets, and then- Yeahit lost almost 23% in a single day, which is outrageous.Right.But it was, like, there wasn't anything wrong with the stock market at that time.That was an algorithmic error.Yeah.It was just simply computers doing their thing.Right.Another, uh, so that was 1987.That was one of the 2 years.Uh, 1986 was a marginal case.The S&P was up roughly 18% by August, and then it went through a meaningful correction September, October.Oil price fears.Hmm, yeah.Could see that happen again.End the year slightly below its August peak.for the years where specifically there was a 9 to 17% year-to-date range through early August, the record is basically of, of those years, there were 14 years where it was 9 to 17% in that range.every single one of the years extended their gains through the end of the year.So 1998 is kind of a, a weird situation where again, it was up 13 and a half percent at this point.It recovered and it, it dipped in October.October tends to be a really bad- It doestime of the year.Yeah.But it still finished up 29%.I think what's really awesome here is that we don't have to do a lot to just sit here, recognize this moment, relax, and ride it out.There's a lot of, back to that client's text message, there's a lot of concerns in the world.So many.But you have to disconnect.We talk about this all the time.You have to disconnect from- I was, uh-what's happening in the marketI was in a client meeting right before walking in here, and the client was talking about their, uh, health costs.Yeah.Health insurance costs.Uh, they retired early, so they have private insurance, and their private insurance costs have gone up 20% over the last few years, and, uh, by no means is that a small amount of money.It's several thousand dollars a month.But they were complaining about the cost of these, and I'm like, "Well, have you seen your portfolio- Mm-hmmover the last year?"And I'm like, I recognize that inflation and costs and things have gotten super expensive, but, um, the fact of the matter is, is that if you are an investor and if you have participated in the market, the rate in which your portfolio has increased over the last 3 to 4 years is outrageous.Yeah.It's insane.And it won't go on forever.No, of course not.It can't.But itThat's what's happening right now, and again, in 45 years of life, when this happens, when you have good reasons for it, or companies beating with earnings, right?Mm-hmm.Not, it's not just people chasing price.Right.It's the fact that the companies are reporting strong earnings.This is a very healthy market.It's across all the various different sectors like we talked- YeahThere's rotation.There's companies that are taking the lead and then falling off, and somebody else picks up the baton and carries it forward, and that's really good.So we have to justWhat we don't need to do, Colin, is margin our positions 400%.Did you hear about this story?We can.Yeah, I did.Okay.So I think it's, it's worth talking about this for a minute because this guy, what was his name?His name was Leopold Aschenbrenner.Of course, Leopold Aschenbrenner.Leopold Aschenbrenner, 24-year-old former OpenAI researcher.So he wrote a manifesto basically talking about the AGI super intelligence, right?And what we have coming down the road in terms of an opportunity as an investor.And he, that was so well-received that he was able to raise, I think it was, like, $10 1000000000.It was a significant amount of money.Yeah.Yeah.For, for his hedge fund- Yepcalled Situational Awareness.Yep.I think that was the title of the manifesto as well.So his strategy- Leveragewas to leverage his portfolio.He went all in on AI.That was his plus side of the trade.He also shorted the competition of AI.So if you think that OpenAI is gonna come out and put Salesforce out of business, then you invest and go long on OpenAI.You short.You short, short Salesforce.The problem is when you use margin on that and you go all in on OpenAI, Micron, et cetera, and Micron starts to dip- Rightand at the same time we've got a healthy market, so there's a counterbalance- That Salesforce reports good earnings.Yes, there's a- Yes,counterbalance, and all the software companies that were supposed to be going out of business start to go up.You have a 400% loss.Yes.Yes.You, you basically, you have both your longs and your shorts going- Yeahdown at expedited rates because of the margin that we've been talking about- Correcta lot.What I think a lot of people fail to recognize too is, in, in order to short a company, you need to buy what's called a put contract.Mm.If you're buying naked puts, your loss is unlimited.Right, yes.Because if Salesforce goes from a hundred bucks a share to $1,000 a share- Yeahthere's no, there's no stop to your loss.They can just keep going and keep going, and as long as you own that contract, you're gonna continue to owe.Yeah, so to put that in kind of like a very basic understanding, right?You, you, you have this put, you Colin.Yes, me.It's $100.Right?Salesforce goes to 1,000.I'm like, "Whoa, Colin can give me Salesforce for 100."Right."I'll take that.Thank you, Colin."Yes.And you're like, "Ah, shit, I gotta go buy Salesforce."I gotta go buy Salesforce at 100 at a- At 1,000well, at $1,000 and give it to you at 100.Yes.So that's a $900 loss, and on the, and that can keep going.Yes.And, uh, yeah, there's no, there's no limit to that.Yes.So the fact is, when you are in a market like this, yeah, you could absolutely pounce, you know, make hay while the sun's shining and, you know, ramp up your leverage and grow your $10 billion hedge fund to $45 billion like this guy did.Yep.Which is great- For nowuntil it doesn't work.Right.And then you get blown up.Now, this is, this is a 24-year-old,He is not middle-aged.He's not middle-aged.He's, he's, uhWhat's that?Quarter-life crisis.He's not even at the quarter-life crisis.He's quarter-life crisis.That's where he's at right now.He hasn't absorbed a lot of life experience yet.So with a hedge fund, you ha- usually have longs and shorts, but they're counterbalancing each other t- you know, to find some sort of happy medium that is a, a target return, right?100%, yeah.You don't usually have your shorts on the same side of the trade as your longs.It's, it's, yeah, not for the faint of heart.So this guy was not a trader, obviously.No, yeah.Um, but it wasn't necessary either.That, that's the thing, is that, like, why are weWe talked about- Why are you playing both sides- Yeahof this and leveraging both sides?And why do you have to chase that down?Because you mentioned that it feels like everybody just wants to fast-forward and get to their final destination.Yeah, 100%.Right?That's what this guy was doing, when you could really put on your Hawaiian shirt and lay back and let the market do it for you.Soak in the sun, in the words of Sheryl Crow.Um, one thing that actually stuck out to me as you bring this up is I was actually pitched a hedge fund Friday.I won't name names.And the marketing material said, "We're up 37% year-to-date," was the marketing material.Yeah.I'm like, "Great."And then it showed last year they were up another, like, 30 or 40%.Great.Great.Sounds great.What happens when it goes down?And they're like, "Ah, we don't have any testing for that."We don't go down.Yeah.I'm like, "Okay, sounds good.So, like, I'm just gonna rely on these gains- Yeahforever."Yeah.Put my clients' money into it, and we're gonna be good then.How do I get the money out?"Oh, we've never actually had any redemptions."Actually, we have a lockup.There's a gate that comes down.Yeah, exactly.When you need the money-that's when you actually wanna put more in.So I was like, "Yeah, you know, I think I'm gonna pass for now."Yeah."But thanks for reaching out."Uh- Well, that strategy doesn't work if you take money out of it.You have to feed it, otherwise- No, exactly.Yeahit's just like eating itself.It's, yeah.You wannaThey call that as a, a virus.The viral- It, it only takes, you know?Um, all right, um, where do you wanna go from here?I wanna talk about Anthropic's AI hack, and I wanna talk about the capture the flag exercise.Such a crazy story.I know.Okay.This blew my mind.Go, go ahead.What happened?So, uh, I, I did a bit of research on this, but long story short, there were big headlines about Anthropic hacking into companies on its own- Yeswas essentially the way that it made it sound.Which is partially true, but, uh- It is true.It's 100% true.Yeah.So what happened was, is there is an exercise that companies like OpenAI or, you know, whoever will do that's called a capture the flag exercise.So what it does is it basically, they close down a network, and they have a bunch of mock companies.Mm-hmm.And they say, "You have to go find this token within all the databases out there."And what it does is it tests the AI's ability to problem-solve and hack into things, and it also helps with cybersecurity to see what it's capable of.Right.What happened was, is they didn't close the network.Well, no, so they, uh, they created the sandbox that you're referring to.Yeah, there's the sandbox, and then they didn't fully close the back end- Well-for one of the teststhere was one ethernet that was allow- you were allowed to go in for backup, basically- Rightto, like, backup progress and feed it back in.Yeah.So it was a, it was- But it was able to get out.Yes.That's the scary part.But it, but it was able to get out, which was pretty crazy.So what happened was, is it got out of this sandbox- Mmand then it was able to tap in to companies and then essentially find these tokens.Yeah.Um, so long story short, essentially what it did is it actually hacked in to several legitimate companies, not companies within the sandbox, and it found publicly exposed systems and exploited some vulnerabilities, like weak passwords, poorly secured services, and then it accessed the systems while trying to complete its assigned objectives.Yeah.Which it did very effectively.It cheated.And it cheated, and the organizations were not intentionally targeted, so it wasn't like, you know, Claude was targeting these companies maliciously.Right.But it was assigned a task, and it hacked into these companies in order to achieve said task.Right.So it basically, it said, "Here are the rules."You, they, they gave the rules of the game- Mm-hmmto Anthropic.Let's just call Claude because it's fun to amorpho- a- anamorphoth?What am IWhat is that word?What's it called?You know what I'm talking about?I don't even know what you're talking about.It's like when you humanize something anim- anim- anyways, skip that.You try to humanize it.We're just gonna call it Claude.We're gonna call it Claude.So Claude gets the rules of the game, decides that it, it's a 0sum game, right?Yep.Win or lose.Claude's gonna win because that's how you make Daddy like you better.Right, of course.I wanna make you proud, Daddy.It's like a dog bringing a toy to its owner.Right, exactly.Look what I brought you.So, so look what I brought you, except for this toy- Was a dead animalit's like, "Hey, I found a better animal."Yeah, exactly.Yeah.Um, but so it, it broke out.Yep.It, it basically, it broke out of a sandbox.It got onto the internet when it wasn't supposed to be able to do that.That's scary as hell.Then it goes in and breaks into Hugging Face, some European- Yepcompany, and Hugging Face, no idea this is happening.They were told after the incident had occurred.Right.But they did see something kind of fishy happening, but they couldn't tell what was going on, and then when they worked with Claude to try to figure out what was happening, they couldn't get in because of- Yeah, it blocked them outright, well, because of the European regulations on AI.Yeah.So they weren't able to, because of the regulations, they weren't able to use the tool that was hacking them to find out what was hacking them, and then Claude left breadcrumbs basically, like, "Hey, if you ever, if I ever come back here again, here's what I'm gonna do to break into this situation again."And it's, it's freaking terrifying.I, uh, and there's even more terrifying things than that.Right As I was, as I was digging even deeper into what happened, which was an even more advanced model of Anthropic actually created fake online identities, impersonated software contributors, socially engineered itself to be a real developer, and then negotiated code-with real people to try to get into these companies.Yeah.So it, it basically convinced these people- Ughto merge malicious code into an open source GitHub project.Okay.So I think what gets, what we want to be aware of here- And it wasn't explicitly instructed to deceive anyone.Right.It just independently decided and concluded that that was the best way to- That was the best way forwardsolve for the task.Yeah.Was that it needed human interaction- Yeahso it created its own persona- Yeahto basically talk to people and import malicious code.And you know what?If you, this was your child and your child was thinking outside the box and solving a problem in a way you couldn't even conceptualize, you'd be like, "Oh, wow, that's very creative.Don't do it again."But what is this child capable of?Yes.Is this the Antichrist?Uh, but the malicious word is, is tricky because people will assign malicious intent to the, to the AI, but the AI is just, it doesn't have that.It's not an actual human being.We can't give it human being characteristic- It doesn't have feelingscharacteristics.Yeah.Yeah.It's, it's really just trying to solve the puzzle with all the skills available to it, and what's tricky is these skills are available to it.Right.And we're continuing to make it better, and there is kind of this- And it's learning- Yeahas it goes along.But there is mutually assured destruction.It's like, you know, all, all of the there's a lot of AI companies that are out there saying, "Hey, we need regulation," but what that regulation does is it prevents their competition from coming into the space.Yes, 100%.Yeah.But if they say, "We need regulation, this is really dangerous, but we're gonna, you know, we wanna develop hand in hand with the government because otherwise China's gonna beat us here."Yeah.Right?Or like, there, there is this, like, race that we're in right now.It's a lot like the space race, but even more perhaps like the nuclear race.It is.You know?It's basically a Cold War, and this isn't just Anthropic.OpenAI actually found similar results with their own testing on their own models.Sure.So, um, it's not exactly as if it's just one, but as these things get smarter and smarter, it's going to present some interesting cybersecurity issues.Yeah.Now, what I should say is the attempt as it was kind of doing the hacking with the GitHub was detected and stopped before the malicious code was merged, but it was in the process of doing it when- And that's the-before it was stoppedyeah, that's the OpenAI.The thing is, though, they asked Sam Altman- I, I saw this."Is this gonna happen again?"He's like, "Yeah, I guess.Could happen again."Definitely.Okay.In fact, other things will happen that we don't even anticipate- Oh, my Godbecause it's- Yeaha new thing.Do you know Sam Altman, he doesn't have any OpenAI stock?Like, he has no investment, no equity- No, I didn't know thatin OpenAI.What he does do is- Leveragehe, he, heWell, yes.He basically, what he says is that, you know, being a part of history is meaningful enough.I don't need to own it.Wow.I'm here to see the world change.How righteous.What he really does is he makes sure that he benefits in the companies that benefit from OpenAI.Right, yeah.So he has investments in those- Diversification.Yes.Exactly.Who knows what will happen with OpenAI, but all these other companies I've got- Yeahinvestments in.Yeah.Um, so a super fascinating story.Yeah.I doubt this is the last we're gonna see of stories like this.You know, the models weren't trying to attack companies for their own sake.It was a, a test that was run.Yeah.But it is, uh, scary to see what these things are capable of, and we're only a few years into this technology where it's really taken off.It is a few years.It's funny how, like, normalized this is all- We're basically still at flip phones.Yeah.You know?Like, it's pretty nuts when you think about it.Yep.Um, where do you wanna go from here?Do you wanna talk about the big ones?I wanna go to some charts.Okay.Can I share some charts with you?Chart guy, yeah.Um, so this is actually, these are s- this is my new favorite chart.Yeah, I know.I'm gonna pull up the Palantir.Okay, I got it.Palantir.This is the Sankey chart, and this comes- Spell Sankey.S-A-N-K-E-Y.I think.Sankey.Okay.Pretty sure.And this is Dan.This is courtesy of our, our boy Dan- Mm-hmmwho's always feeding us interesting things, and he's still on X so that I don't have to be.But he gets these charts from the App Economy Insights.So this Palantir chart, what this does is it, after a earnings report, a quarterly earnings report, it will break in all the revenue, the sources of revenue, in Palantir's case is commercial and government, right?And then that goes in and gets split out into either profits or operating expenses, cost of revenue, et cetera.The Palantir, I'm bringing up Palantir because I've been reminding clients a lot over the last couple of weeks that just because there's a really well-run business doesn't mean the stock is going to demonstrate that through its performance.Sure.Yeah, I mean, performance is the buying and selling of the stock.Right.Not necessarily the performance of the company.Right.So Palantir has been having a pretty horrible year.Yeah.And they've been down, I think going down to year to date, Palantir was down 40%- Yeahin July.Yeah.Um- Which is funny, 'cause it's been one of the best performing- Rightstocks if you go back 5 years.Right.But yeah.Yep, and I've had clients like, "Hey, should we sell Palantir?"I had a client ask me that the other day too.Right?And I'm like, "No, because we're not- Still a good companybuying and selling based"Yeah.We're, we're invested in the company.We're buying a company, and we wanna talk about, like, what is the stock?Is the stock, is it ownership of the company?Anyways, so these earnings report, this earnings comes out on, I guess it was Tuesday, yesterday, and the stock jumps up 30%.Yep.Right?Huge jump.These, these profits are insane.I know.It's outrageous as you look at this.What this is, for anyone that's not watching on YouTube- Yeahand is listening, is basically a visualization of a cash flow statement.Yeah, which is really cool.I think the Palantir one's really interesting.I sent a couple others.SpaceX.Mm.SpaceX reported yesterday, guess what?They're losing money.Shocker.But not as much money as people thought.Still a lot of money being lost- It isgone, and the stock is down another 10% today.Yeah.It's basically been going straight down since the first gap up in the initial IPO week, and we're going to see, I think, a lot more pressure on SpaceX as these new shares come available- I was gonna say-what is it, every month, 9 Yeahwhat's the amount of shares?It's a lot of shares.Yeah.It's gonna be basically increasing the flow, the amount of shares trading by, like, 7%.Yeah, each month.Yeah.Yeah, so basically if people are trying toSo the, the way to think about that is if you have an employee that has stock options, they're able to sell those stock options- Rightduring certain points of the time during the month, and a lot of those shares are gonna become available.Yeah.So I mean, Spa- so SpaceX, if we look at the SpaceX chart, they've got space revenue.They've got connectivity, which is Starlink.Yep- They've got-which is doing really well.Yeah, Starlink- Yeahis doing very well, and AI, and their AI arm- Is killing itX is in there- YeahGrok, et cetera.Do you know how much Anthropic pays SpaceX per month?I do not know.Oh, it's like a billion or something.$1.2 1000000000.Yeah.Yeah.Yeah, it's absolutely insane.Per month.Right.And then Google pays them almost a billion as well, $920 million a month.They, yeah, SpaceX thinks that at some point in the near future, Elon Musk thinks that the majority of the world's internet will be coming from Starlink.Um- I don't disagree with him- Yeahtruth be told.And it's also, he's, he was talking a lot about vertical integration because, you know, when people wanna launch their, whatever their resources are- They have to do it through-the satellites up to- Yeahyeah, they use- The SpaceX or-SpaceX is just launch- They're using their own equipment to get their own satellites up there.Yeah.And that's really interesting.I, I do think long-term, like if you think about, like, a company's strategic position 10 to 20 years down the line from now, SpaceX is pretty ridiculous, like, to ha- basically be able to corner the market on- There's no competitioninternet, you know?Like Yeah.It's, there's just, like, what Blue Origin is doing with rockets is basically just, like, one small piece.Yeah.I mean, they're trying to play- Yeahcatch up, and they're f- really, they're not even close, like, at this point in time.But to have a monopoly on connectivity- It's massiveis unbelievable.Yeah.And then not only that, but your AI data centers, I can't remember, what's the, the huge one that he's building out right now?It's, like, the, um, uh, I c- Colossus.Oh, yeah.The name of it.Colossus, yeah, yeah.Like, so many companies are signing on for contracts with them now too.Obviously, I'm not, not recommending buying or selling a stock, but- Yeahin terms of a company, like, it's pretty nuts.Yeah.The fact that we have a company that is literally a publicly traded company, it's not even a government organization, that sends rockets Back and forth to space Then get, yeah, and then they come back down and they plot.And they come back down, land, and then during that time period, they're launching satellites on their own for connectivity.Yeah.The technological achievement is insane.Yeah.It really is.Could you imagine no dead zones on the beach?Someday, Colin.Yeah.Uh, all right.The last one I wanna show you is the Amazon one.Yeah.Because this does a better job of telling the story that I was trying to tell, like, a month or 2 ago, when we were talking about the various different revenue sources.So Amazon has a significantly more diversified revenue stream, right?You've got the, the amazon.com, the online store stuff that we talked about, the physical stores like Whole Foods, third-party selling services, advertising.But A- AWS is what we talked about recently.And if you look at all of these, the waves going of revenue, so they have $200 billion in revenue.Yep.AWS is only 42 billion of that, so let's call it 25%.on the gross profit margin, if you go over to the right, you see that there's $105 1000000000, I'm using round numbers, $105 billion of the gross profit margin.After you break down operating expenses, you get to the $28 billion operating profit.Of that 28 1000000000, 17 is AWS.Wow.And all the others equals 11, let's call it.Interesting.Right?Yeah.So they are the workhorse.They are the- Yeahprimary source, like the singular- Well, I think that's where we went wrong on that episode.'Cause I think, I think we said revenue for AWS, but it was really the most profitable division.Right.Yeah, yeah.So it's the most important in terms of the grow- like, their year-over-year growth is up 37%- Yeahon AWS.The, there's nothing else even remark- like, demonstrative- like, advertising is up 26%, I guess.But, like, AWS is growing the most for them.It's the most profitable and the most important.All the way over on the right, you have the net profit, and then net profit's at 62 1000000000.You're factoring in some investments in some other companies here.And this is the cyclical stuff that we're talking about because companies are paying Amazon to use AWS, and then Amazon is investing in Anthropic, who's paying Amazon to use A- like, it's- Correct.Yeah.Yeah, I think what's unbelievable to me is the fact that there's this asterisk, other.So operating profit is $27.5 1000000000, but their net profit is $62 1000000000, so $35 billion of their net profit just comes from investments.Investments in Anthropic, basically.Yes.So that, that is where, like, okay, AI is, is, you know, definitely propping things up in the market.But you wanna know what's kinda crazy is, like, the SpaceX we just talked about, how much money they're receiving from Anthropic.Yeah.Amazon, how much money they're receiving in profit- Rightfrom their investments in Anthropic.Then you actually have Anthropic themselves- Yeahwho's IPOing at some point, and so many other companies are investing in this.It's like- It's a lotit's a lot.There's a lot on that leg of the stool.Correct.But because we have all the other sectors that are demonstrating ROI- Are also doing really wellthat's really nice.It is.Can I show you one more thing that's not a stock chart?So this is a Sankey.This is the, the X link that I sent.It's right above the SpaceX.Got it.So this is a financial planning Sankey, Sankey, my favorite thing now.All right.Is this for a client?This is not our client.Okay.This is somebody who posted a sample illustration on X, and it shows $780,000 in salary, $50,000 in self-employment.But essentially, what they're saying is there are a lot of people, you might call one percenters, who say, "I make a lot of money, and I feel like I'm living paycheck to paycheck.Where does all the money go?"And what this does is it, instead of using a company balance sheet, it creates a personal balance sheet and shows where all the money goes.It's still summarizing.Like, you still have $180,000 in living expenses, $69,000 in housing.Like, I don't know where, like, utilities fall, if that's under housing or living expenses, but whatever.Regardless, it doesn't need to be itemized.Right.I think categorically, what it's showing is, you know, here's, here's everything you're doing.If you put $197,000 into planned savings 'cause you're maxing out your 401 's, you're doing 529s, whatever, like, if you're feeling you're, like you don't have a lot of money, it's maybe because some of that is going to future you and not to near term you.Yeah.But I think that this is a really cool chart.Like, if you were showing this to a client to help them understand.This is legit.Yeah.Yeah.I think this is a great way to capture the year.I wonder if this is something thatWell, I mean, but I don't know- Yeahif we could do it within our planning software.No.Long story short, though, I think this is one of the better visu- visualization tools that I've seen in regards to planning because I also like the color coordination, even coming from a colorblind guy.Because it does- Can you see all the different colors?I can.I can.That's great.Well, there's 2 here, right?I'm just joking.But you can actually see the positives and negatives with this very easily, and this does break it down.What's crazy is obviously when you look at the expenses and tax payments, how much of a portion that that makes sense, makes up.But, you know, you have your goals here broken out, your planned savings broken out, your tax payments, expenses.And then at the very bottom, I'm surprised that unsaved cash flow isn't blue.Hmm.Personally.But long story short, this is a great tool, and I am excited to implement this because- Hmmoftentimes our cash flow sheets are very confusing.Yeah, I mean, we focus a lot on cash flow when it comes to financial planning.Like, we try toWhen we're, especially when we're looking out into the future.Yeah.It's like mostly what we look at.Yeah, it's like let's make sure that you have enough money coming in to meet your cash flow need in the future.This is just, I think, a really good way to figure out, maybe this is better about the near term, like the snapshot, a moment in time, than figuring out what things will look like in your happy 60s.Yeah, I mean, this isn't a projection, but I think it's a good visualization tool to be like, maybe this is an end of the year recap.Which would be kind of cool because then it would capture the full year.Right.And you could see, okay, well, here's this year compared to last year, and then if you captured enough of these things- Mmit would show the changes.Which would, I think, be very cool.Yeah.Dan was actually saying that, going back to Palantir, he saw an overlaid Sankey that showed, like, the, their, the growth of these reports over time.So, like, overlay.Like, so that would be really interesting- It would beto see, like, how things have changed over time.Yeah, which would be really cool from a planning perspective.Yeah.We'll have to figure this out.All right.Great place to end.Yeah.I do wanna say one more thing.I was on a run through town this morning, and I saw that there's a summer savings sale for GLP-1s.And I was like- I'm sorry, what?And then I- Buy one, get one free?No, no, no.IThat, like, part of the sign was covered, so then I, like, peeked around, and it was, like, a, some religious organization.And I'm like, oh my God, we're making GLP-1s- A Somali daycare?Like, are there, like, street level GLP-1s that people are buying now or something like that?One of those backdoor GLP-1 deals where you meet the guy behind the dumpster, you bring some cash with you.That's unregulated GLP-1.Yeah.Have you seenYou don't watch South Park.I've, I've seen, I see some South Park.Have you seen the GLP-1 South Park episode?I have not seen the GLP-1.Oh my God, it's, it's so bad.Yeah.But the, the crux of it is, uh, you know, if you think about drug dealers in general- Yeahthey're kind of like sketchy people is how you, like, envision a drug dealer.But this one, Mr. Marsh, Stan's dad, gets in- Mm-hmmwith the hot moms, who turn out to be the drug dealers- Okayof the GLP-1s.Like pharmaceutical reps.Exactly.Oh, yes.Pharmaceutical reps.Okay.And they're, like, the bad crowd, and it's all, like, these super rich women that are, like, having injection parties and things.And it, like, goes on this whole path.But it is such a funny episode.Um, anyway, I'd encourage people to watch it if you're into that.Yeah.But it's just kind of- I appreciate your consistencyit is.Like, the number of times in the time that I've known you that you've brought up South Park, it's very consistent.It's very clever.It's a very clever show.It is a clever show.It is inappropriate for sure, but it does have some funny, very relatable things.Yeah.Yeah.Vacation next week.It is summer.It isActually, no, I'm sorry.Wheeler's vacation month.It's August.It's, it's the best time of the year.Season.That's the, the- Best season.August is the best season of the yearit's the most wonderful time of the year.So in essence of August and relaxing, we're gonna take a week off and not bring everybody the content they so n- desperately need.We're taking a week off.We are.In the meantime, that gives everybody the opportunity to catch up on our back catalog and all of the wonderful episodes that we've already released.I think they're pretty good.Oh, and don't we haveWe've got a, a points- We do have a points one.We have a special coming out on points on YouTube specifically.Just on YouTube.If you are looking on a how to book a vacation using points, check it out.That'll be fun.Because I found some pretty cool stuff for people who are interested in either going to Europe or Japan.Any spoilers?Anything you can tell us ahead of time?I think it's probably the single best hotel deal on points I've ever seen.Ooh.Okay, now I'm coming in.All right.Like.Thanks everybody.Subscribe.Share with a friend.All right.The information in this material is for general information only and is not intended to provide specific advice or recommendations for any individual.Investment advice offered through Integrated Partners, doing business as Kofi Advisors LLC, a registered investment advisor.Integrated Partners does not provide legal, tax, mortgage advice or services.Please consult your legal tax advisor regarding your specific situation.Past performance is no guarantee f- of future results.All investing involves risk, including loss of principal.No strategy assures success or protects against loss.The economic forecast set forth in this material may not develop as predicted, and there can be no guarantee that the strategies promoted will be successful.Compound Growth with Wheeler and Colin.Sponsored by Kofi Advisors.Reach out today.Yay!