All right. Welcome to unsupervised learning. It's Daniel Meisler. All right. What do we got here? All right. So got a new definition of prompt engineering that I'm basically using. Now when I talk to people. And I can share that a little bit later. AI's impact on the job market. Working on a number of new talks. I didn't think I was going to do this. I thought I was actually going to be like 2 or 3 for the whole year, but it's like turning into a lot more.
So curious to see where that's going to end up. It's going to end up being like 4 or 5 talks. What I like about this system that I'm using, though, is that I'm only speaking about stuff that I normally talk about anyway, and that's what essentially what the newsletter is and what the podcast is. It's like stuff I'm already talking about and it's the same for talks, it's the same for presentations, it's the same for blogs. It's the same for tweets. Like I'm not doing a strategy
that relates to particular media or syndication destinations. I'm not thinking at all about that, actually. And I've talked about this multiple times before, but people keep bringing it up, like, how do you come up with ideas or whatever? First of all, the way I come up with ideas is I read a crap ton of stuff, like all the time. I'm always reading mostly books, but also articles and stuff like that, and social media and other builders and creators
in AI and security and stuff like that. So that is the main way that I get ideas is by having tons and tons of inputs. And the metaphor that I use for this before was like a particle accelerator, like I'm just like the atom, the tiny little brain in the middle. But I'm being bombarded by the coolest ideas from like Greek classics to like ancient philosophy productivity books to AI stuff to whatever. And when that stuff
hits my brain. If you've ever seen those outputs from one of these particle accelerators, it's like this super colorful, amazing, like explosion of things that come out of it. And those are all the subatomic particles. And what's interesting about this analogy is those only last for this very small
amount of time. And that's why I'm so obsessed with capturing, because if I'm having a conversation or I'm out for a walk and it's funny because going out for a walk or taking a shower with no tech is very similar to a particle accelerator, where you're actually capturing ideas from really smart things, because what happens is your brain will start freaking out and it's like, hey, let's think
about some stuff. Hey, what about that one time where you read that one thing and it starts just churning and coming up with cool ideas, which is why you have ideas in the shower, because your brain is freaking out because it doesn't have any input. So it starts like just generating things. I wonder if we're doing the same things in dreams or whatever. I'm not sure. Maybe a similar mechanism, but the point is, I have ideas because I consume a lot and the ideas don't even
necessarily have anything to do with what I'm consuming. They tend to be peripheral to what I'm consuming, and that's why I read deeply but broadly. Right? Because I'll switch to fiction and there'll be a cool idea in the fiction, and then suddenly the fiction is the peanut butter. And this nonfiction AI thing that I'm working on or coding on is like the chocolate, and they mix together and
all of a sudden I have a cool idea. And this is why I really believe that this whole idea of like, oh, people are super smart or people are super genius or whatever, look at this person who has all the ideas. Nope. The ideas come from inputs. If you stop reading and I know this for a fact for myself, if I stop reading within a few months, even a couple of weeks, like my brain will just slow down. Like I will not have the same number
or quality of ideas coming in. And this is why people like Buffett and Charlie Munger and all these people, they're like, look, if you want to be smart, you need to read. You need to be consuming really high quality stuff at a very high rate of speed. And that's not always the case with like the rate you can actually get really high quality ideas. Riva has talked about this where like slow reading, a deep book like the Bible, or like a classic from the Greeks or something, whatever.
You could slow read and try to memorize and like go super deep with the thing, try to fully extract all the wisdom from it and do that on a regular basis. And you still get lots of ideas. But what you can't do is do nothing and think nothing, and consume nothing and just consume. You don't want to just be like watching Netflix or something. Of course, there could be some good stuff on Netflix where you actually get some ideas, but in general, that is a thing
that calms your mind and like sedates it right? You're not getting high quality content coming in. It's more like relaxation. So all right, cool giant diatribe. And we've got through one bullet in the damn newsletter. All right. Publish a new official pattern template in fabric. So official Pattern Template is the name of it. And it's basically all my latest formats and instructions. And like the structure of the of the pattern or a prompt really an I prompt
that works best for me. And I did a few updates here. I've got a lot more, I would say comments or descriptions of the sections. I broke out more sections. I added a examples and negative examples section, and I also broke out the identity and the goal. So a number of upgrades in there and people have been asking like, okay, how do you make a perfect prompt? So I made this, that basically shows that. And of course that's going to be updated constantly. Right. Because this is like there's no
testing for these things. So everything I'm doing is like it's like it's like placebo. If it works for me, I'm like, yeah, this is really cool. I like it, and I continue to do it because it works for me. And when I test it against other things, I am doing a B testing. But first of all, LMS are non-deterministic, so I can get different answers for whatever it's like. In order to do this properly, you need to run it like tens of times, dozens of times, hundreds of times.
And of course you have to worry about cost here, but then you have to have a really good objective rating system that's looking at those and actually finding the differences between these prompt techniques. And a bunch of papers have tried to do this. I have not necessarily agreed with the results of those papers. And again, that could be because the whole thing is a moving target in the first place, right? Bottom line is this what works for me? It is the official pattern template in fabric
right now. Did my buddy Jason Haddix red blue, purple I class last week. It was fantastic. It was Monday, no Thursday and Friday of last week. It was so good. Jason is definitely the best teacher that I know and it was like, you're learning stuff. But he's also just hanging out. He's talking or we're talking all his friends. Different people like myself were also contributing to the class a little bit. He would ask a question, we would give an answer. Jason would chime in on that. And
the people taking the class were also super smart. So they were contributing lots of stuff, and it was just like it was just great. It was absolutely great. Both days of content really enjoyed, and the next time it comes out, you should definitely sign up. So I got a new sponsor conversation here with Abhishek from Material Security. He's the CEO over there and we talked about a bunch of cool stuff. It was really a great conversation. One of my favorite topics was like, why are product
managers so good at being CEOs? And so he gave a great answer to that. I got another piece up here on getting a perfect sound from your microphone. And this sound that you're currently listening to is a good version of that, although I don't do post-production for this particular thing because this is an OBS video going directly into YouTube. So I'm not doing post-production on here, but it should be a pretty decent initial sound. I got
a political one here. It's called The left's. It's political, but it's centrist, so I don't think it should bother you unless you're crazy left or crazy right? But you should check that out. Lots of people from the left and the right have reached out and said that they loved it. So that should be a good metric for you. Yeah, it's called Left's Brexit security got a bad checkpoint. Von massive botnet was busted up. Ukraine is using AI against jamming.
So yeah, it's basically like kill decision from Daniel Suarez where, um, the drones were getting jammed, GPS jammed, and also just any RF was getting jammed. So they made them autonomous so they could just navigate by vision and looking at landmarks. And then the idea was you have a drone swarm. And this is from the book and probably also from reality, also brought to you by the Pentagon, a drone swarm with tiny little explosives on them. And they would just
all fly and navigate. They would have a target that they were trying to kill, and they would navigate by the landmarks. And when they find the person, they could do facial recognition on board, don't need to call out anywhere and then just crash into them, kamikaze and kill the person. And jammers don't matter because they're not using navigation based on RF smart home tech in warfare. So Home Assistant is being used in Ukraine to warn people
of alerts of attacks incoming. Yeah. Incogni is another data deletion service similar to delete me moves your info from over 170 data brokers. They are not a sponsor. Otherwise you would say sponsor on the top of it Ticketmaster. So they got hit by a group called Shiny Hunters, and it looks like around 560 million people's data has been stolen. 560 million. I'm so glad they were getting broken up, or at least we're trying to break them up. And so glad that Taylor made that happen. I think
she did. A few people spread the most information about fake news about vaccines and stuff like that, going, this is back to like 2020 and 2021, but, uh, older Republican women. 80% of the misinformation. Yeah, I wrote this myself. Small, persistent groups can have a significant impact on propaganda. You wouldn't think that. You would think that small groups like that would just be like a glass of water in the ocean, but turns out if you're persistent enough, it'll
get out there. NSA released a guide on how to keep your phone from getting hacked, and they basically said you should reboot your phone once a week. Frightening. What do they know that I don't know? Or they're like, yeah, you should reboot your phone every week. And I'm like, why? Never mind. Don't want to know technology. All right. So grok, I talked about this. How fast it is wickedly fast this thing okay. So it can now do be lama
3 to 1200 tokens per second. That's essentially like it just fills out the page as if it just instant. And so here's my question. Why doesn't Google or OpenAI or Microsoft or someone just like, walk up like, you know, a clearinghouse sweepstakes or something? And just like, here's a box of $1 billion, please give us your technology and disappear. Right. This is like Magic Beans proprietary chip tech that lets him basically run AI. What, tens of times faster than
anyone else? Ridiculous. Looks like Satya Nadella is freaking out about a potential partnership between OpenAI and Apple. I think he's right to freak out, and he should make a phone before it's too late. Or a personal assistant on some kind of device. He needs a personal device. Anyone who wants to play going forward need a personal device and an OS, and it needs to be able to see and hear and basically be integrated with all your context. And anyone who's not doing this is screwed. And people
have a massive advantage. Are people who already have access into a robust operating system, namely Android and Apple. Researchers have new attention mechanisms that actually outperform standard multi-head attention. Cool. Well, let me know when it fixes haystack performance. Need I need haystack performance. I need to be able to give you giant text documents and you don't miss a single thing. So wake me up when that happens, all right? Recall
a privacy disaster. This is the Microsoft thing that records everything. Stealing everything. This is what I wrote. Stealing everything you've ever typed or viewed on your windows PC is now a reality, thanks to a feature in Copilot plus recall that is a privacy nightmare. Yeah, disagree with that take. I think in ten years almost everyone will think computers who don't have this are basically worthless. And in fact, ten years most people won't be talking, will only be talking,
will mostly be talking and gesturing to their computers. Computer will mostly be like monitors and HUD displays or whatever, and your computers are basically doing the work. So the idea that it's going to forget something you worked on or forget something that was said to you, or that you said to somebody else is completely asinine. Of course your AI is going to remember everything. What else would
it do? All right, Llms aren't just internet simulators anymore. Yeah, so increasingly llms are being trained on custom non-internet data that's there's still a data wall there as well. But I don't know, I think a lot of people like social media companies. Meta is in great position for this, for example. I mean, because they have Instagram, they have all this thing. So basically whatever people are posting and doing and talking about, that becomes a massive source of data.
Anyone who's like really close to the OS and especially in messaging apps, social media apps, that is a massive source of data that they can then sell to other AI companies or keep it for themselves to have an advantage. In the case of like Lama3, Nvidia is getting ready to overtake Apple in market cap. That's insane. Completely insane. But here's a way to make sense of this. Imagine that there's a giant filter on how many people are capable of creating a hit movie, writing a book, starting
a business, sharing their art with billions of people. In other words, imagine the current number of major builders or creators is something like 0.00000000000 17 of our 8 billion people. I just made up that number. It's not a real number, despite how precise it is. So just work with me. So the reason Nvidia is rising so fast is because one of the primary pieces of tech is the GPU. Okay? It is like center mass of this entire thing. And what we're talking about is removing 3 to 5 zeros
from this big zero number. We're talking about multiplying humanity's creativity output by thousands in the next decade. And that requires chips. And thank you for coming to my Ted talk about Nvidia. That's essentially why they are skyrocketing. Google released an SRE handbook. Basically saying simplicity is a core principle for reliability makes total sense to me. They should make one for product management. That's mean. But it's funny.
All right. Karpathy's GPT two 124 million parameters looks like LMK and 90 minutes for. $20. Karpathy is an absolute beast. Tiny Grad's latest update. This is from Geohot. Slashes Python time and ditches external dependencies. Terminal animations. These look really cool. I'm not sure I'm going to use them, but they look really cool. And humans little known intelligence bureau that
nailed the outcomes in Vietnam, Iraq and Ukraine. I want to build something that basically programmatically goes and gets these predictions and incorporates them into my life model and how I see the world, because it's essentially like these prediction markets are the smartest people doing the best kind of predictions, and they're doing it in a very transparent and logical way. And I learned about this from the book called Superforecasters. Is that Tetlock? Bill Tetlock, I think, might be the
name of the author or one of the authors. Unbelievable stuff. Really, really good stuff. New study says sleep inequality in the US is tied to economic stress. Makes sense to me. Mice with PTSD like behavior showed significant improvement after having access to a running wheel. Basically, exercise induced neurogenesis like
healing trauma with exercise. That was a bit of a stretch, but that's where they're going with this Ozempic and wegovy changing eating habits, fine tuning your taste buds, making sweets taste sweeter. Mm. Yeah. I mean, it's definitely having an effect on me. Like I had pizza. I've been having pizza because of watching this damn pizza influencer who samples
pizza and is just making me want to eat pizza. Anyway, I get a pizza, I eat, like three slices and I'm just like, yeah, it was good, I guess, but I'm full. Whereas before before I took Wegovy, I would be like, I would eat three quarters of the pizza and be like, I should not eat any more. And then I would eat the rest and be like, what's up with ice cream? John Carmack dives into bullshit jobs. Yeah, really good book by David Graeber. Got to read that book. I need to read it again. No, I don't I
should just summarize it. I have a fabric pattern for that. Imagine navigating high school social maze without a smartphone, missing out on chats and memes, beginning a unique perspective on life and friendships, the social lives of teens who don't have phones. I don't have to imagine this. I grew up in the 80s. It was glorious. Yeah, very much on the tip of Jonathan Haidt basically saying getting phones
out of schools highly support that effort. Comprehensive review finds I exercises might actually help in preventing and controlling myopia. This is actually an incorrect read. This was my fault. This is an incorrect read. The study. There are a whole bunch of studies saying it helps and a whole bunch of studies saying it's not. But this particular study actually says they couldn't find much evidence, strong evidence that it did help. So that's that's boo on my part.
Lung cancer breakthrough unprotected. Oh yeah a lot of people are talking about this. I don't get excited until it comes to market and everyone loves it like Wegovy. But lots of people are talking about how this thing is insane. 60% of advanced stage patients progression free for five years. Marc Andreessen says to stay in school. But then he says, if you're the type of person to drop out, you also won't listen to this advice. And I love this letter.
Love my Wife is dead letter from Richard Feynman to his dead wife. Very touching ideas and analysis. My favorite way to explain prompt engineering. Basically, don't think of it like I think of it in this way. Instead, prompt engineering is how to explain to a superior intelligence what your problem is and what you would like to happen. And if you can't do that well, you will lose to people who can't. That's it. How to explain to a superior intelligence what your problem is and what you
would like to happen. That's prompt engineering. All right. My X thread on why it's so hard to find tech jobs right now. All right. So this this is the vibe here problem for the job market isn't that AI is happening. The problem is that it's happening at the exact same moment that most companies are figuring out that 80% of their employees are worthless. We need to stop expecting
things to go back to the way they were. The new reality is companies mostly hiring super ambitious, exceptional, proven people who are total gods. With AI, this means most formal education becomes a waste of time and money, because a degree doesn't certify that you are super ambitious or exceptional or proven, it doesn't certify any of those things. So in this model, only elite schools will matter because the filtering for being exceptional will have happened just by
being accepted into the school. That doesn't count for like people who got in because their mom already graduated, right? But for people who got in via merit to a top school, that is a good filter. So really this is two separate things. One, you're exceptional enough to be accepted into the top school and two, you finish some classes. Problem is, number two, you can get anywhere, right? The actual training in school between one like a state school
and like Harvard is not that much different. Number one is the thing that employers actually care. The filter to be able to get into an elite institution. So the result will be companies hiring the top 10 to 20% of people in competence. And this will be filtered by elite school attendance and proof of competence via something you've put into the world, like on your website, YouTube as a tool or a company you built, or an AI algorithm rating you. This is going to be huge as well.
So the hyperbolic form of this is if you're not 19 and at Harvard, or if you don't have your own projects or companies you've built and talked about online, you are not going to be interesting to employers. You will be part of the 90% basically fighting for scraps and the recommendation of the week for everyone you know who's going to be thinking about their career in the future, which is basically anyone who's not independently wealthy try to get them to think about things in the following way.
AI is going to take out most executors of knowledge work. The people who will survive are the people with ideas who are actively building that thing, actively building that thing. So it's not just ideas. So the safest thing to be is a creator or a builder. That means a founder or an entrepreneur, a programmer, somebody with ideas who is productive and ambitious, like an artist like you could
be an artist. It doesn't have to be some hardcore engineer, but it's got to be some kind of category like that. And the trick is you have to be able to have the ideas, be able to make it yourself, or attract the talent and the AI and harness the AI to be able to make it. And then you've got to be able to market the hell out of it, which is a lot of writing. So a lot of like communication and persuasion. So bottom line, the winners in this new game will be the people making new things
and bringing them into the world. I'm going to say that again, the winners in this new game will be the people making new things and bringing them into the world. And the aphorism of the week. The most beautiful people we have known are those who have known defeat, known suffering, known struggle, known loss, and have found their way out of the depths. Elisabeth Kubler-Ross.