UL NO. 448: TSA SQLi, NYT Github, NK RPM, NVIDIA Mystery...

relies on. Join the 5000 fast growing companies that leverage Vanta to manage risks and prove security in real time. Vanta scales with your business, helping you successfully enter new markets, land bigger deals, and earn customer trust. Vanta Unsupervised Learning listeners get $1,000 off Vanta at vanta. Com slash unsupervised. That's vanta.com/unsupervised. Welcome to Unsupervised Learning, a security, AI and meaning focused podcast that looks at how best to thrive

as humans in a post AI world. It combines original ideas, analysis and mental models to bring not just the news, but why it matters and how to respond. All right, welcome to unsupervised Learning. This is Daniel Miessler episode 448. Let's do this. Okay. I want to ask everyone to resubscribe to the podcast. I don't know all the different channels that are listening to this. This could be on YouTube for example. There is a podcast feed which you

can get on Apple or Spotify or whatever. So I would just say go sign up for that because the URL has changed and I'm putting out a lot more content and putting out a lot more clips of different, shorter pieces in addition to the weekly episode. So I would say go check that out. And if you're listening to the podcast, then, uh, apologies, because obviously you're subscribed,

otherwise you wouldn't be listening. Okay. My new favorite fabric pattern is one I just made up called Extract Primary Problem. And I'm so excited about this thing. So it takes any text input or any body of work. So you could basically say like Noam Chomsky or you could literally like echo the words Noam Chomsky or like the Unabomber or whoever. You can send basically any person's body of

work or like Russian literature or whatever. You send anything in there, and it will turn it into a single sentence that summarizes what that work believes the biggest problem in the world to be. And the reason I'm doing this is because I'm trying to understand people's work, or an essay or something based on a concise description of what they think the problem is. I think this is a very powerful way to see the world, is to understand what do they think the biggest issue is that

we need to solve, because that's where you start. And then you look at, okay, are the solutions that they're providing, are they actually good strategies for addressing that? But most importantly, what do they think the actual problem is? Because if they think the actual problem is like something you don't care about, I mean, why are we listening to their work? It doesn't mean there isn't like some other way to

get value from their work. I could think of examples where there is, but I would rather listen to Viktor Frankl's work, who believes, you know, that there's a lack of meaning in the world, for example, which is definitely my view. Uh, I invented, uh, I invented basically the same thing that Viktor Frankl did. Um, well, probably lots of people have done it before Frankl as well, but I was basically going down and am still going down

this path. Then, like a year ago, I read Viktor Frankl's work, and it is I mean, it's 100 years before I'm doing this work, and it basically found that anyway, it doesn't matter. That's a diversion. He nailed it. And his problem statement is essentially the same as my problem statement. So it raises the question of like, okay, for a given person, for a given person who's producing value and work in the world, what do they believe the problem is?

So that's what this particular pattern does. The other one I created recently is a create story explanation, which explains a really difficult piece of content in a flowing story style. So it basically is like, okay, so they believe this, which means this and a few facets of what they believe is this here and also this here, which means this part here. And the reason this is important is

because of this. So it turns it into a story based, conversation based explanation of the topic as opposed to like extract wisdom, which is like, you know, listed out in a very clear and concise way, but it's still kind of sciency or scientific like, oriented as opposed to like story and conversation Oriented, so I really like this one as well. Highly recommend everyone check out the. We've been Lied To About Work essay that I put out last week.

Getting a lot of positive feedback around that. So appreciate that. Security. Researchers have discovered a SQL injection vulnerability in a critical air transport security system. This is actually Sam Curry, um, and another one of my buddies who actually put this out. Really cool. I mean, the vuln is, uh. Yeah, remarkably, I don't know how that vuln still exists, but they are basically able to use SQL injection to add themselves to a list of authorized individuals to get cockpit access

to a plane. The New York Times has revealed a significant breach involving GitHub tokens, and attackers are already using it to gain access to various repositories. Researchers from OpenAI, Microsoft, MIT, and Harvard have proposed personhood credentials to verify real humans online without revealing their identities. This might also be needed for voting. This might also be needed for handing out UBI. This might also be needed for social media accounts. There's

so many reasons to have this. I think this is really cool tech. The other thing that's interesting about this is Sam Altman has a world coin, which I think is crypto based on authenticating yourself. So you could see world coin like folding into this as well in a way that ties in with UBI. Um, yeah. So it's zero knowledge proof, which is a crypto concept. Essentially, you don't have to know who it is, but you know

that it is them. Yeah. One thing I worry about is the back end database probably get eventually hacked, because at some point you have to have a database that you're actually validating against. Thanks to Nudge Security for sponsoring the show this week, Record a future has announced a new integration with Google's security operations, enhancing both Siem and Soar components. North Korean hackers are back to targeting the

NPM code repository with malicious packages. The US Army Special Forces showed their hacking skills during the Swift response 24 military exercises in Sweden. They used a remote access device to hack into a building's Wi-Fi, disable security systems, and then stormed the building, leaving leaving behind signal jamming equipment and a laptop playing Never Gonna Give You Up, so

they rickrolled them on the way out. Chinese companies are planning to launch over 15,000 low Earth orbit satellites with Mercedes page from the Australian Strategic Policy Institute warning that this can enable countries using Chinese broadband services to control information flow, monitor user activity and even shut down the internet. Internet during unrest. Yeah, I think that's the whole purpose

of it. Las Vegas police are pushing back against a new NFL policy requiring officers working Raiders games to provide photos for facial recognition. They don't want potential misuse of that data. I mean, the biometric data is a picture of them. I don't know. I don't know why they're so worried about that. Maybe there's more to it. I don't think so, though. I mean, you call anything biometric data and people are like, oh, I don't want that out there. It's like, yeah, that's just your face, man.

It's just your face. And it's not actually your face. It's a picture of your face. Google has announced new variants of its Gemini 1.5 model, including the smaller Gemini 1.58 B and improved Gemini 1.5 flash and a stronger 1.5 Pro. I have not messed with this yet. And by the way, we are in the AI and tech section. I've not messed with the new Gemini 1.5 Pro. I hear it's really good, so I'm going to do that

within fabric soon. OpenAI keeps making everyone wait spelled w e I g h t. I don't know why I think that's funny. It's very stupid for their new model. Half model, whatever is going to come out soon, maybe called Orion and maybe using the new strawberry technology. So we're all waiting for that. California's AI regulation bill, SB 1047, has passed the state Senate with 29 to 9 vote,

and now is going to Gavin Newsom. And this is basically a set of rules saying companies making models of a certain size must have certain safety measures in place, like you have to have a kill switch and stuff like that. It seems rather logical. Elon actually backed it, which was surprising. OpenAI is reportedly in talks to raise a new funding round at a valuation exceeding 100 billion, led by Thrive Capital, with Microsoft also expected to participate.

And Nvidia and Apple have also been rumored to be involved in potentially wanting to invest. OpenAI has enhanced its assistance API, making it easier for developers to fine tune how AI assistants handle file searches. So this is basically more controls and power Around. Rag companies like JP Morgan and Walmart are shifting from restricting generative AI tools like ChatGPT to developing their own internal assistants. Basically, they know they can't use the cloud versions, but they know they

also need to use AI. So there appear to be deciding to build internally. Cisco is acquiring robust intelligence, which secures AI applications. Not close to the details at all here, but feels desperate from Cisco. Like we know we're screwed because networking is not where it's at anymore. And we need to do something in AI. So let's buy a thing that secures AI applications. I don't have high hopes.

Today's new note pin is a wearable version of its previous credit card form factor, which I have and have used. I don't use it that much. I don't like the form factor of having to pull it out. It's actually the friction of using the app that's that's hurting my use of it. So hopefully that's better with the pin. Also, I don't like the credit card version of it. Actually, I need to take it out of my wallet because

I don't really use it that much. Amazon is set to release a new version of its Alexa voice assistant in October, and it will be powered by Anthropic's Claude AI models. I am getting the new one. I have it on order though. The new note pin. Sorry to go back to the previous story, but I'm getting the one that goes around the neck, similar to the limitless one, which I also have on order. I hope it has a red button though, because I want people to know, and if I don't know the person, I'm going to

ask and say. My thing is like recording everything. Not specifically for this conversation, but it's kind of recording everything just to let you know. So I could turn it off if that bothers them. Amazon is set to release a new version of Alexa Voice in October, so very soon, and it will be powered by Claude. That is going to be actual. I am very excited about this. I can't wait. I don't actually have any Alexa stuff, but

I can't wait for the Apple version. Nearly half of Nvidia's revenue comes from for mystery customers, each spending over 3 billion on AI chips like the H200. Heavy reliance on a few major clients raises concerns. Yeah, and and they just got hit with an antitrust kind of like a probe yesterday. And I think they crashed like 10% like the stock went down like 10%. That was painful. And and tech is shutting down after 27 years. They

said they are passing the torch to Tom's Hardware. Kind of sad I kind of grew up on an antique. China's Wukong game just sold 10 million copies in three days. Huawei posted record profits in the first half of 2024, 7.7 billion in net profit, despite ongoing US sanctions. That's insane. Imagine what they would have done without the US headwinds. A woman in California used an AirTag to track down her stolen mail, and the suspects were found with mail

addressed to lots of other people. So they're now they're facing multiple felonies for disrupting federal mail. Humans. Anarchy in Sudan has led to the worst famine in the world seen in 40 years. Nearly half of New York City bus riders are skipping the fair. 48% of people are not paying for the bus, presumably because they just like, what are you going to do? What are you going to do? You're not going to fight me, you're not going to kick me off. There's not enough police. The

police don't care. So hopefully they're going to start clamping down on that. Researchers at the University of Kentucky have found that long Covid patients show brain changes similar to those seen in Alzheimer's. The study, published in Alzheimer's and Dementia, highlights shared issues like neuroinflammation and abnormal brain activity, suggesting common underlying mechanisms. That is super interesting and exciting. We'd

love to address both. Scientists have discovered that the interaction between two molecules, pkmzeta and kibra, is crucial for maintaining long term memories, so this appears to be a pretty major breakthrough on how long term memories are formed. Ozempic, a drug for type two diabetes and obesity, might also slow aging. Are you kidding me? It's like they keep finding new things that this does. Researchers found it could treat illnesses like heart failure, arthritis, Alzheimer's and cancer, even

reduce deaths from cardiovascular issues and Covid 19. Oh yeah, that's a new thing I just saw. Covid 19 death reduction by 34%. That's not a story in this one, but it's a story I just read. So it's in the podcast. So I wonder how much of this. Oh yeah, I put it right here. All this as news of it addressing more and more issues feels like it's hitting something extremely fundamental, like inflammation, which has long been pointed to as a meta cause or like a meta symptom.

I think meta cause more than meta symptom in lots of other diseases. I'm not saying it's actually inflammation, just that it seems to be affecting something very fundamental. But what if it's just that being thinner and having less visceral fat or something related to that is actually the thing. Or maybe being thinner. And the less visceral fat leads to less inflammation, which the inflammation is causing all the

other things. So like maybe the Covid 19 and the aging and the dementia and all this stuff is related to one or both of being thin and healthy, maybe physically active, maybe inflammation, I don't know, but I'm very curious. I can't wait to see, like this mystery unfold. A CIA deep cover operative known as Anthony Laguna spent years infiltrating Islamist extremist groups, even reaching al Qaeda's broader network.

But he died in 2016, seemingly seemingly of like mental health issues, which is calling attention on the CIA for like, how do you how are you taking care of your people when you put them in these traumatic situations? How do you handle their mental health? And more people are going no contact with their parents, driven by a mix of personal growth and unresolved conflicts. I think the gulf between Gen Z and what are Gen Z's parents? Are

those exes? I think those are exes. Maybe the difference between Gen Z and Gen X is just so large, like ideologically, that maybe that's why this is happening. It's definitely a factor. Discovery. Whoosh is a command line tool for transferring files and opening shells over a peer to peer WireGuard connection. Super cool fire crawl crawl sites using Claude or GPT and turn the output into LLM ready markdown. History.

For feed, Jozek developed an open source tool that creates a complete historical archive of full text posts from any RSS or atom feed. I absolutely need to use this. Yes, I need to use this like immediately. Uh, which is why I put it in there. So the most dangerous email I've ever sent. This is a great read. Who wants to be hired? A Hacker News thread for people looking for work. My buddy Clint Gibler did an epic post summarizing every AI talk from Blackhat and Defcon 2024.

Just an amazing piece of work. It is so like so deep and like involved. Yeah. Very comprehensive. Definitely check it out. Using GPT four for web scraping. Three questions candidates can ask to invert the power dynamic in technical interviews. The Hatch Restore two is a smart alarm clock designed to mimic sunrises and sunsets to help you wake up

more naturally. I am so close to getting one of these things, but I'm stopped by a simple fact I get natural light in my windows in my bedroom when the sun comes up, so the sun provides like a sunrise like experience. But I wear an eye mask when I sleep to block out the light. So I feel like this is like the worst kind of overengineering when you have actual sun, you buy a fake sun, but you can't see either of them because you're wearing an

eye mask. Uh that's silly. Okay, ideas. Beware of commodified incuriosity. This is a really good piece. So it's talking about the act of researching and thinking being replaced by focus on efficiency and productivity. I think it's a great way of looking at things and a reason to be cautious with the overuse of something like extract wisdom, which I

talk about a lot. Basically, I summarization basically a big part of learning something is struggling with it, and that's why I tooling focused on learning should be used, in my current opinion, to help you find things that you read slowly and absorb and you grapple with and think

about critically. And then you could use something like extract Wisdom to help you make sure you don't miss things when you take notes, but don't think that anything other than future learning implants or something can substitute for the hard work of actual thinking and processing and grappling with content. Depression as a hand on a stove. This argument suggests that instead of trying to eliminate depression, we should see it as a signal to make life changes. Attention to enchantment.

This piece argues that being enchanted by the world comes from learning to pay attention to it. I feel like learning to meditate, which is really just paying attention in my Sam Harris based school of meditation combined with music festivals. For me, I've truly learned to appreciate small things in

daily life. It's made me a very happy person. I really feel like it's a cheat to be able to extract this much joy from, like, my neighbors, and seeing people go for a walk with their little kid, or hearing distant like children playing and like roughhousing and screaming or whatever. And this also relates to framing, but the action or the attention piece is really, really key here. Rarity and beauty. I often wonder how much of beauty is just rarity. There are lots of seagulls at this

lake that I visit often. The local lake here in my town. I barely noticed them by default, but they're like big white birds and I actively try to look at them sometimes as if they are rare to try to trick my mind into noticing their beauty. Ralph Waldo Emerson has a quote if the stars should appear but one night every thousand years, how man would marvel and stare. So if the stars only came out every once in a while, there would be like eclipses, organized like eclipse

watching things. They would be organized just to go and watch the stars. And the opposite is true as well. If the eclipse happened all the time, nobody would look at it. It would be like the most boring thing ever. So it's the fact that it's rare that makes it special. That's really it. And I often think about how much beauty we have in our lives that we ignore because it's either too common or too omnipresent, or maybe those

are the same in this context. I feel like the Stoics had it figured out when they taught the exercise of imagining your life without certain things. I try to do this, and it does help me appreciate them. Two bad choices in November. This is a political one. I'm not going to do it, honestly. It's just that I think both candidates are both bad in different ways, and I'm very afraid in both cases of either of them winning, but for very different reasons. And I think both of

them could do some good as well. Recommendation of the week. Here's a frame for you to try on. This goes along with the previous one of the previous points. You become what you pay attention to. It's election year in a few places, including the US. The world is on fire. Politics are a mess everywhere, and it seems like the very fabric of Steven Pinker's last couple of books has completely unraveled. But maybe we have an option other than

staring directly into the toilet. Maybe all the beauty in the world is still there, and maybe we could focus on that instead. Or at the very least, not on the toilet. I'm not saying to be cowardly. If you're one of the few people who can actually change lots of minds and help the world in some sort of way, maybe you should serve time in the toilet. That is

our current situation, but that's probably not the case. Most of us won't be missed on those front lines, so maybe instead we could use those few months to make a list of the best poetry books, or start a D&;D campaign with your friends, or learn to play piano. Maybe we could focus our attention on the great stuff in the world that we always said we'd look at later.

Now is a pretty good time to do that. It's a good time to fill our attention with that instead of the ugliness in the world, so we can become the goodness for someone else. And the aphorism of the week. You become what you pay attention to. So be very careful what you pay attention to. You become what you pay attention to. So be very careful what you pay attention to. Unsupervised learning is produced and edited by Daniel

Meisler on a Neumann U87 AI microphone using Hindenburg. Intro and outro music is by zombie with a Y. And to get the text and links from this episode, sign up for the newsletter version of the show at Daniel miessler.com/newsletter. We'll see you next time.