Episode 102
RECAP: KEY TOPICS OF 2025
Keywords entrepreneurship, technology, AI, data centers, insurance, personal growth, literature, startup ideas, environmental impact, search evolution Summary In this episode of the Venture…
Keywords
entrepreneurship, technology, AI, data centers, insurance, personal growth, literature, startup ideas, environmental impact, search evolution
Summary
In this episode of the Venture Step podcast, host Dalton Anderson reflects on key themes from 2025, including the challenges of discerning reality in an age of advanced technology, the evolution of search with AI, environmental concerns surrounding data centers, the insurance industry's struggles with risk management, the need for return on investment in AI, and personal growth insights from literature and personal experiences. Dalton emphasizes the importance of accountability and leveraging expertise to generate ideas and drive success.
Takeaways
It's becoming increasingly more difficult to discern what is real. Search is rapidly changing to the point where people are preferring AI for search. The environmental effects of data centers are significant and often undisclosed. There's a 200 billion insurance gap that needs addressing in the industry. We need to focus on AI ROI to justify investments in technology. Start with what your team is especially good at to generate ideas. The hard thing about hard things is a must-read for entrepreneurs. If you're going to eat shit, don't nibble; face challenges head-on. You have about another 40% left to give when you feel exhausted. Hold yourself accountable and be blunt with yourself to improve.
Episode content
Explore every layer of this episode.
Each article, guide, analysis, and field note has its own focused page and stays linked to this source conversation.
Articles & stories
Narrative and editorial pieces that carry the conversation forward.
Research & analysis
Evidence-led work that tests and expands the claims in the conversation.
Synthetic Media Verification Research Note
Media verification should separate origin, integrity, manipulation, and truth. A file can be authentic to a camera and depict a staged event. A generated image can be acc
Startup Idea Generation and Testing Research Note
Jared Friedman's original Y Combinator talk controls the seven idea-generation routes. E102 correctly attributes the framework to Startup School after referring to it imp
Insurance Protection Gap and Mitigation Research Note
The protection gap is the difference between economic loss and insured loss. A number requires a hazard, geography, period, method, and source.
Honest Personal Review Research Note
The reflective guide separates facts, identity, control, influence, constraint, responsibility, response, support, recovery, lesson, changed behavior, and review date.
Hard Problem Escalation and Recovery Research Note
The public guide takes a brief attributed lesson from *The Hard Thing About Hard Things* and turns it into an operating sequence. It does not reproduce the book or glorif
E102 Annual Recap Research Note
E102 is a solo annual recap recorded after the topics it revisits. The transcript is authoritative for Dalton Anderson's views and personal experience. It is not authorit
Data Center Community Diligence Research Note
National data-center estimates provide context but cannot establish one project's local impact. The project record must control.
AI Search Publishing Research Note
Google's generative AI search guide was last updated July 10, 2026. It says SEO remains relevant because AI Overviews and AI Mode use the existing Search index, ranking,
Field notes
Focused observations and durable ideas worth carrying into other work.
How to Tell If an Image or Video Is AI-Generated
You usually cannot prove media origin from appearance alone. Trace the source, inspect credentials and watermarks, search earlier copies, and report uncertainty.
Seven Ways to Find a Startup Idea, Then Test It
Use seven Startup School routes to generate startup problems, then test the user, alternative, urgency, frequency, buyer, economics, advantage, and kill criteria.
SEO for AI Search Without GEO Theater
AI search changes query paths and measurement, but durable visibility still comes from original, crawlable, people-first content, clear entities, evidence, and links.
Why Better Models Will Not Close the Insurance Protection Gap
The insurance protection gap reflects physical risk, exposure, mitigation, affordability, capacity, regulation, and take-up. Better models address only part of the system
How to Conduct an Honest Personal Review
An honest personal review separates facts from identity, control from responsibility, recovery from avoidance, and one changed behavior from punishment.
Data Center Community Impact: Records to Request
Communities should evaluate a data center with project-specific power, water, cooling, air, land, incentive, workforce, emergency, and reporting records.
How to Confront the Hard Problem First at Work
Confront a worsening problem by establishing facts, containing harm, choosing the right channel, assigning authority, communicating honestly, and measuring recovery.
Full episode
Read the complete record.
The show notes, transcript, and source trail remain on this canonical episode page.
TranscriptRead the full conversation.
E102 RECAP_ KEY TOPICS OF 2025
Transcript
Dalton Anderson (00:01.134) Welcome to Venture Step podcast where we discuss entrepreneurship and two trends and the occasional book of view. In this episode, we're gonna be doing a 2025 recap. That's right. We're gonna be resurfacing topics that I spoke about all throughout last year that I think are important to bring back up. And maybe you might be curious and dive into that episode if you haven't listened to it prior. But this episode is gonna be structured in themes.
The first theme is going to be tech. The next theme is going to be related to literature takeaways. And then the last theme is going to be personal. Try to get the juicy stuff up front. And if you're still around, you can learn a little bit more about me. But before we dive in, I haven't done this in a while, but I'm your host, Dalton Erison. Today's episode is special because it's a one year recap of 2026.
2020 I'm in 2026 2025. I'm already time traveling. This is me from 2027 recording that pretend that I'm in 2026 for the episode that's in 2025. Regardless, this is a special episode close to two years now on the podcasts, consistently releasing episodes, which is incredible. Appreciate everyone's support and
love how much the podcast is growing. It's impressive, the numbers. And yeah, I just wanted to call out that I appreciate it and it's awesome. Thank you for listening. That's it. But if you didn't know who I was, my name is Dalton. I come from insurance and have a technical background, data scientist. In my free time, I like to run, work out, cook, do all sorts of wacky stuff.
and read. And I've been reading, but I just haven't done any book reviews. So I'm working on it. I've got an interesting book that I'm almost done with that I would love to share with the show. I'll be doing that soon, maybe in a couple of weeks. Who knows? We'll see what we have on a guest schedule. OK, so the first theme I said was tech. So let's get into it. So the first theme is about where we're at. And I think there's a couple of things to talk about.
Dalton Anderson (02:27.416) The first one being it's becoming increasingly more difficult to discern what is real and what is not real. And that's coined the synthetic Rubicon. And once you cross it, then distress is created and it's really hard to understand what is going on. There are some initiatives to prevent this from being widespread, but in the future there's going to need to be sophisticated
validation of what is real and what is not real. And right now, photos are what's currently hard to discern. But eventually, pretty soon, probably within a year, we'll get to videos and videos will be once you get to videos and pictures. It's really hard to know what what is what is up, what's down. Like, am I seeing a simulation?
is this podcast simulation? And this is one of the main causes or reaction all. Actions that I had about a video podcast because I wanted to make sure that people knew that it was me and it wasn't some A.I. avatar speaking and everything being artificial, which will become a thing. One hundred percent. I podcasts and these.
social medias will get flooded with these AI generated folks and it's gonna be some kind of governmental agency that's running all the accounts. We've seen this before with Facebook pages or the top Twitter meme pages where Russian accounts or this, all sorts of crazy. I don't wanna get so conspiracy theorist here, but we've seen it time to time again where.
If there's an ability to have influence and influence people, then large organizations or governments alike will take that opportunity to exert and extract the value. It's economics. I don't know. mean, whether or not you think it's ethical or right, it's still going to happen unless
Dalton Anderson (04:51.534) there is clear guidelines and validation parameters to prevent these things from happening. So that's an interesting dynamic. Wondering how that's going to play out. The stuff that they are working on, which is great. mean, some of the stuff that Google has created about having a synthetic key that's generated within the pixels. So the pixels, once the pixels are created, there's a synthetic key that's registered
with Google. And so you could tell or you could ask Google whether or not this is AI generated or not if Google made it. And I think that they're trying to expand that into other image generators. But I think the problem that I talked about last time was if you make any edits to that image or you
rescale it or you crop it or make any edits to the pixels, then I'm pretty sure the code is lost. And so then it doesn't match the code that's in the registrar. And so then you don't, you don't match up. And so you wouldn't be able to tell if it was AI or not.
So then there's a couple things that need to get solved there. But I think the main thing to watch out for is this discernment factor, the lack of the ability to discern what is real and what is not real when it comes to images. And eventually we'll be getting to videos. There were some serious, serious models coming out last year or came out last year. The Sora 2 model and then the VO3 model were very good.
The fact that they did voices, they had different accents, languages. They were incredible, still not to the point. And they also incorporated physics models into the video so that I think I had an example about somebody paddle boarding and then jumping off the paddle board and.
Dalton Anderson (06:58.754) them understanding the shift of weight, the bend of the knees and the water rippling before the jump and then the jump causing a large ripple of the water and then the person splashing causing a different type of ripple and they can bind and then it's settled out. Very detailed work. Something that wasn't a thing two years ago is now a thing. And before you know it, a year from now,
it's going to be really hard to tell what's going on. So not to be too loom and gloomy, but that was the uncanny valley section. The next section is something that I want to dive deep into. And I've experimented with myself, but I did an episode about how search is changing. And if you think about yourself and if you're listening to this episode, you're probably experimental or curious.
and you probably use AI to help you search things across the web or help you find items on the web for you.
You're not the only one. And search is rapidly changing to the point where people are preferring and utilizing AI for search. Predicted to trend. To be like the primary search. Search factor. Form factor, yeah, the primary form factor of search is AI search. With that changes how websites are structured.
Because websites always had these crawlers and these crawlers would prioritize what websites that users would see first and then how long those users stayed there would determine whether or not that was a valid suggestion, which would change the weight of the authority of that website plus the backlinks, healthy backlinks, all this stuff. And that's how you get your ranking, your page ranking. That's not necessarily the same.
Dalton Anderson (09:06.009) ball game as like generative search, like generative engine optimization versus search engine optimization, because you're no longer searching yourself. It's a computer. It's always been a computer, but you know what I mean? It's this AI agent that people like to call it. Anyways, that changes how the websites need to be structured and the information needs to be much more detailed than a human.
would necessarily need because AI wants all the information to understand the background, the context, the theories, the resources, the references, everything. And then it compiles it back to the user, whereas the user would be turned off by that having that much information. It would. I said it when it's the user. The human would rather have a more streamlined approach in providing information.
and we're always taught to be direct, concise.
and try not to overcomplicate things. Whereas the AI wants it to be complex, not too complex, but hierarchically complex, I would say, where it has information that it can pick and pull from, and it's able to understand the structure of the website from a site map. And then it's able to understand
the structure of that webpage, which you would use hierarchical headings and bodies for AI to understand.
Dalton Anderson (10:45.549) What's really going to happen is you'd have to change how websites are structured. not saying like this, website needs to be restructured. It's more of like you need to restructure your content and you're optimizing for different things. Like are you optimizing for references? Are you optimizing to be part of this AR search? Are you optimizing for authority to get referenced or do you want direct links? And those change what you're trying to do and how you want AI to.
incorporate your website or brand. It's the whole thing. I personally have incorporated this philosophy into my website, the one that has the VentureStep articles and the videos of the Spotify podcast.
And I personally think it's works. I structured it to be more AI friendly and it was actually less work for me. I basically turned my podcast, almost that subscription, transcription to.
a what I like to call a smart transcription. So it's a transcription that has little snippets, references and quotes and sections that are hierarchical. And so when AI goes on there, it can look through everything and see what sections to look at and where to go and what it wants to do. And I have seen a meaningful uptick in unique visitors.
given that
Dalton Anderson (12:29.359) I'm not optimizing for people, but AI and I that's fine. I don't necessarily care if humans are on there as much. It's just more or less getting people to search. And when they're searching, me getting brought up as potentially an option to listen to or read about or little excerpts from here and there to help the guests that are on the show and help myself.
That's all it is. It's fine. It's no big deal. But I personally think it's works and I think it's been going splendid. Okay. So the next theme is related to tech, but it's more related to the episode that I had about AI and the data centers and the GPUs used for
the computing and training of AI models.
And if you're not familiar with this, it was pretty insane. The environmental effects of the centers on some occasions, and I'll explain that in a second. And then also the shady under the table deals that the city slash county had with these companies.
And then the undisclosed water and electrical use that they put estimates down, but it wasn't necessarily clear to the city government that they would be using that amount of water or electricity. And I think they had a fairly good idea about the electricity piece and that there was most of the time capital arrangements or agreements.
Dalton Anderson (14:31.681) to increase the infrastructure capacity of the electrical grid due to the strain that their warehouse or like where their data center would have on that area. What wasn't understood
was the amount of water that they were gonna use. And we really found out about all the water use through a breakthrough environmental investigation in pretty sure Arizona with Google's.
data center and Google hadn't disclosed how much water they're gonna use. Of course, no one was doing it. So they're like, we don't need to do it either. And then there was these people researching this like water aquifer in Arizona. And it was draining way faster than it should because no one was using it. And they're like, well, what's going on here? Why? Like, is the water leaking or how are we losing so much water?
And then they tracked it back to Google's operations and they were like, what the heck? Like our estimations are saying that you're using this much water. And they're like, yeah, about that. We are. We're using we're using a good amount of water. And they're like, you can't use that. You can't use this well. This is you can't do that. That's a no. And so I think they stopped doing that and.
And then they had some kind of agreement to get water somehow. But yeah, there was there was operations where it needed a lot of water and they didn't disclose it. And then they're already in a water scarce area using a lot of water, which was brutal. And when I'm saying a lot of water, like I mean, like certain places, they're going into these rural towns and they're using they're using like 60 percent of the water, like.
Dalton Anderson (16:45.069) water usage is like 60 % of the data center and the rest of the people. Like insane volumes, like 800,000 gallons, like crazy stuff.
And a lot of the water is wasted at that point because of the chemicals that it runs through. It's not just water. It's put into like a solution that prevents a corrosion and these other things and lubricant like lubricates the machinery that runs through. And so there is a lot of water instantly evaporated. And then the water that is left over is not usable, nor is it.
I'm gonna say potable, but I know that's wrong.
It's not drinkable. OK, so. It's not drinkable water and it's wasted and a lot of it becomes water vapor. So the next question is, what about the environmental factors besides the water usage? Well, it's interesting thing is since these data centers are scaling so fast and the electrical grid infrastructure typically takes a long time. They have created a workaround and the loophole is like, OK, well,
Given that we can't get this, we need, we already agreed to all these jobs and we've already made our billion dollar investment or however much it was 500 million to a billion, multiple billions. They're like, well, we're going to need to get more electric electric capacity now. ASAP. And so the work around is they just did.
Dalton Anderson (18:28.442) turbines, diesel turbines, and they're massive, massive. They're huge. They're like three stories by like...
Dalton Anderson (18:39.162) I know, like three stories by 30 feet wide. I mean, they're really big. I was impressed how large these things are. And they're on backwater, complete backwater, because all these data centers are getting these turbines. And they're basically not up to environmental standards.
there was a huge issue where you have to get permits or special things you have to do. You've got to take the fumes out of the air because they're low atmospheric fumes and it causes breathing issues, asthma, sickness, all sorts of things. And it's really bad to breathe in. And the big case on that was with Elon Musk XAI data center in a already
environmentally challenged area, they decided that they're going to do like an insane amount of turbines, I think 20. They got approval for like two and then they were like, OK, well, we got those. So let's do like 18. And at a certain point, they're running like 20 at once. the. results came back and they're like, what's going on in this area? And then they did investigations after.
They told them that, we're only running a couple found out that they're running a lot of them.
And then the sad part was the impact on some of the folks that were nearby, because it's been ongoing for some time.
Dalton Anderson (20:19.066) that they were sick, they're hospitalized, they've developed all sorts of breathing issues with breathing in these fumes. And so now they're all they're all jacked up from the turbines. And it's just not cool. Some of the documentaries that I showed in the episode were pretty brutal. I try to be politically neutral, but I know people get pretty fired up when they hear Elon Musk, regardless.
I was pretty bummed out about the turbines and how they approached it. And they definitely could have been more tactical for sure. And upfront. Let me get a drink of water real quick while my mouth is getting dry.
Dalton Anderson (21:05.968) All right. Well, I'm out of water, so keep that in mind. OK, so then the last thing about this section that I think that would be interesting to dive into just briefly is I discussed an episode, the 200 or the 200 billion insurance gap. These historical models that we're using to measure our data don't necessarily reflect what the current environment is, and there's been some recent
refactoring of the models to help reflect this new world that we live in.
But it's a really difficult situation given that everybody is getting squeezed. the insurer wants cheaper prices. The reinsurance group wants smarter risk selection. The MGA wants growth in profits. The carrier wants growth in profits.
And no one wants really give an inch and they all want to take a mile. so there's just and then you've got the watch, the watch groups, and then you've got the regulators. All tussling once and once among each other, and it just causes massive issues about this situation that is really boiling to the surface. And then the last thing is like there's not a lot of incentive for the insured to harden.
there's not a lot of incentive for the insurance company to get the insured to harden. And so then you're kind of stuck in the middle. It's like, okay, well your risk sucks. And they're like, Hey, I know, I don't know what to do about it. And then the insurance company is like, well, let me know when you know, and when you fix the problem. And so there there's a big push with companies trying to provide insight on what to fix and how to fix it and providing resources to fix it. And then there's also
Dalton Anderson (23:11.45) private slash public funding like private funding were at reduced rates and then Public funding from grants with your state if you invest into your home To reduce your risk But just overall there's just problems and in the whole the whole industry like why are we? Why are we developing on land that constantly floods?
The same situation in Florida, like massive flooding and homes being just absolutely decimated apartments, complexes, all sorts of stuff. If it consistently floods, it's in the floodplain. Build a park. Do do that for activities, soccer fields, football fields, tennis courts, whatever it may be, community activities. Cool. And then repurpose the land doesn't need to be some kind of beach house slash like.
multi-use area.
Okay. The next piece is related to AI and trying to get value out of AI. And I've been hearing this a lot is we need AI ROI, return on investment on this AI baby. We need ROI. And so we're, out of the experimental phase and we're into the implement, the implementation slash pushing for ROI. And something that was pushing for ROI was Instacart's
AI implementation or implementation of AI pricing. And it went from value of service to willingness to pay. And they were trying to minimize customer surplus. And they were dynamically changing the price in the same geographic area, depending on how much you like the brand, how urgent the request was, whether you were disabled and
Dalton Anderson (25:14.02) had less mobility than somebody who was not. All of those put together created your maximum extraction value for every product that you were shopping for, which to be frank is pretty messed up. But that's why they got fined 60 million dollars from the FTC when the article came out. Like literally three days later, FTC was like, boom.
Here's $60 million. Please pay at the end of the day.
Okay, moving more personal.
Dalton Anderson (25:58.367) SUS Learnings, which is the startup school, wanted to share, and I'll just read this because I don't want to mess it up. And I have my notes on this page. OK. So seven recipes for generating ideas. Start with what your team is especially good at and think of ideas that take advantage of your expertise.
Look into ideas or look into ideas in your idea space and leverage your expertise and you will automatically have product founder fit. Start with a problem you have personally encountered, have unique insights for.
Think of things you personally wish existed. Look for things in the world that have changed recently that might have created a new opportunity. Look for new variants of successful companies. Go and talk to people and ask them what problems they have. Look for a big industry that seems broken. And the bonus idea, find a co-founder with a good idea.
But you should really iterate through idea or recipes one through two.
Dalton Anderson (27:25.197) Yep. So that was the section about.
generating ideas. I just thought I would share that. I think that was a good framework for generating ideas. Even if you're not trying to do a startup or found a company. What about at your company? What about at your position? What are what are things at your position that you are exceptionally good at? And then what are things that you could leverage your expertise within your company? Generate value and then the company will
tell you and also give you more knowledge, better financial outcomes for yourself.
It's simple, like if you provide a lot of value, you'll get compensated for it. And if you don't get compensated for it, then you can go somewhere else and get compensated for the value you're bringing. Simple, simple,
But I did think that that would be good to add to this show so that you have it and that the world can have it. It's not mine, it's from the startup school. They offer the course for free. Once again, you don't need to be an inspiring founder, you just need to be curious and it does have a lot of great life advice. And if you take the
Dalton Anderson (29:04.32) owner approach at your company, you will be recognized. And if you're not recognized, eventually you will be recognized by somebody else because you'll have better. You'll just be better at everything. Like you would be more self-confident. You would articulate yourself. I think when you're self-confident in your abilities and
feel that you are adding a lot of value that will resonate through the way you speak, the way you act, the way you hold yourself, through new ideas. You'll get more excited, more robust with the ideas you'll come out with. And before you know it, you've blossomed into a new human. So I encourage you to build, explore, and be curious about the world.
and the people around you, and especially people at your company.
All right, moving on. So the next section is literature, literature. And this is probably my book of the year for 2025. And this is the hard things, hard things. Man, now I'm blanking on it. That's crazy. Hard things about art. Yeah, I was going to say the hard thing about hard things.
but I was messing it up. I was going to say hard things about hard thing, but it's the hard thing about hard things by Ben Horowitz. Ben Horowitz, amazing book. Brilliant. I loved it. Loved the book. It was. For me, book of the year of 2025. Book of the year for sure. And 2025 was able to read a couple of books, and that was my favorite.
Dalton Anderson (31:06.072) Ben Horowitz, he was CEO and he was one of the famous wartime CEOs. And there's wartime and there's peacetime. But Ben was someone who led a company through multiple pivots, multiple hard times, an IPO, the whole nine yards. And Ben had many challenges.
and he explained how he dealt with them in the book and he was able to offer advice, frameworks, and overall how you should think about and attack problems, how you should lead an organization, how you should present yourself.
great book, but one of my favorite sayings wasn't even from Ben. It was from Ben's CFO at one of his companies. And I don't even remember the person's name. I'm sorry. But the saying of the year is if you're going to eat shit, don't nibble. And it was from a boardroom meeting when they were having issues with their estimated earnings against their first quarter.
on their IPO. And if you're not familiar, you pretty much bake in the speculative value of the stock price with
your earnings and your first earnings are most important. And so if you don't perform on your first earnings call and your earnings report, it's like, OK, well, you said you were going to do all these things and you're failing on the first report. What's going on? This company is not what they say they are. And then the stock tanks, which is fair because you don't have any historical performance and you're new on the block.
Dalton Anderson (33:01.071) so people don't trust you. So they are running into earnings issues and they had the decision to don't deal with it until later and kind of chip away at it and make excuses and slowly, slowly eat it. Eat the poop in this scenario. Or they could just grill up a full on shit sandwich.
Eat the whole thing. All at once.
Which is great advice. Like if you're have to eat bad food, like if you cooked your food and it sucks, why would you wanna eat it six for six days a little bit at a time? It's just gross. I know every day you're eating bad food. What if you just ate all your food all at once? And it's related to a product, not product, project management task methodology called Eat the Frog.
do your worst thing first. But in this scenario, it's like, don't don't make excuses. Don't put it off. Don't try to minimize the fallout of what is going to happen anyways. Just do it all at once. Get it over with. Pull the bandaid off. Move on. And it's a it's a great, great saying, and it has a lot of meaning. And it was funny. One of my colleagues, John.
had recently had to deal with some stuff and he was like, I got one of those days, gonna eat shit, don't nibble. And that was like legendary saying, But just a great, book. Highly suggested. Okay, the next one was The Accountability Mirror and the 40 % Rule by David Goggins, book You Can't Hurt Me. So David has this,
Dalton Anderson (35:04.401) the that he talks about, two things, the 40 % rule and the accountability mirror. 40 % is just like when you're really feeling it and you're wanting to give up, you got about another 40%, which is kind of crazy to think about, but I think it's true. A lot of people don't push themselves hard enough, and when it does get hard, people like, is too hard, I need to give up.
maybe give up is the wrong word, but I need to give in to my thoughts. And well, that can create, to be frank, a weak mind where you are not in control. Someone else is in control of you and it's not you, it's your subconscious and your subconscious has insecurities, some kind of blocked, you know,
childhood memories, whatever it may be, that you're not, you're not really control those things as much as you are with your mind. There's some, parts of their subconscious that you can control, but it's your subconscious. Like you don't necessarily have full control of it, but in your mind, you can control how you operate in your personal space. And if you remember correctly, on one of my books that I talked about on these stoics,
stoicism books is your personal space is your mind. Your mind is your personal space. That's your only place where no one else can hurt you. At least for now. I'm sure 20 years from now they'll be able to like throw some kind of sticky thing on your mind and fray your fry your brain like hack into your brain. But that's not a problem we have to worry about now. Right now personal space is your mind. No one can control or hurt you within your mind. So that's what he's saying. It's like
When you, when you're tired, when you wanna give up, when you're not in it, when you don't wanna record a podcast like right now, and I'm not saying I didn't wanna be here, but I definitely didn't wanna record the podcast today. It's like, I all this other stuff I gotta do. I have 40%, I got so much more energy, I've got so much to give you guys. I could give everybody. She can have some, you can have some, we're all good. I got plenty energy.
Dalton Anderson (37:30.234) When I'm at zero, I've got another 40.
So that's a general concept of the 40 % rule. And the accountability mirror is, it's just like holding yourself accountable and being mean and being blunt to yourself. Like if you're fat, you're fat. And you either decide that you want to change it or you don't. If you are disappointed,
with yourself and you think that you're not smart, you're probably right. Go do something about it.
If you're bad at cooking, you're bad at cooking. Just submit it. Move on. Do something about it. But don't sit there and complain to people, all my life sucks. I can't believe it. I got laid off my job again. Or, man, in this generation, there's no good partner around. Okay? You sure about that? Because there's millions of married people.
There's millions of people in relationships. What are you talking about? And it's just like, hold yourself accountable. Don't say these things and then complain to people or complain to yourself, woe is me. When you're the problem, you are the problem. You control yourself. You control how you feel about yourself. You control all of these things. It's all in your mind.
Dalton Anderson (39:13.073) Hold yourself accountable. It's simple. It's so simple that if you have a problem, just attack it and you don't have to do all that at once. Like if you're out of shape, don't try to run a marathon the next day. Be reasonable with yourself. But I think the first step is holding yourself accountable and recognizing where you're at and where you need to go and closing the gap. But that's
That's a crucial thing, because a lot of people, they'll tell you, your resume's great, don't worry about it. great episode, darling, phenomenal, you're such a great speaker, you're so good. Okay, what do I do with that? don't worry about it, your episode, awesome, it's so good, the best.
or, that business idea, that's amazing. Like people are worried about chat GPT or these AI chats from being too positive, but then they have these people around them that don't tell them how it is. I think that's the true reflection of a friend is someone who legitimately has course.
in frank correspondence with their colleagues slash friends. And I try to find another word, amigos. There we go. Amigos.
Because if not, you're just in this bubble world and everyone is just bubbling around each other and no one's really having any friction and feedback. And there's no way for you to improve if you can't see it and then other people don't tell you. Because there's some of the flaws you have you don't necessarily recognize unless somebody points them out. It's true. It might be the way you run. It might be the way you eat your food or cut your knife or...
Dalton Anderson (41:23.761) whatever it may be. You don't know everything. mean, you're having a hard time. A lot of people have hard times with correcting themselves with things they know about, not necessarily things that they don't know about. So I think that's the first step is just holding yourself accountable and making it clear on what is the goal and attacking that and recognizing what you need to work.
Okay, so my personal reflections, I'm super thankful to be here. When I moved to New York City, I was very excited. I was starting a monster chapter in my life that I'd prepared for for a long time and I was exhilarated.
And I had a lot of things going on, a lot of things planned, and I was ready for them.
It just didn't work out. It just didn't work out. I got really sick. I lost 30 pounds. My clothes didn't fit. The bones in my back were visible. My ribs were showing out, basically out of my shirt. I was skinny. Skinny, skinny.
And I'd lost my ability to be mobile. I'd lost my confidence to walk around, run, work out, lost my energy levels. I lost my interest in pursuing other things. I lost my spark. Food didn't taste very good. It was bland. Life was dull. My hormones were all jacked up. My hair was falling out. I had to go to the hospital.
Dalton Anderson (43:12.05) I mean, I had all sorts of problems after I got sick and it took me a long time to get better. After I wasn't sick anymore, I had to take one month off of work, which is insane. And then from there, I couldn't work for hours at a time. I could only work at, like the first week, I'd only work at like 45 minutes at a time. And then I'd have to like lay down because I would get too sick or something.
So all these things and me having to consume like rice for like a month afterwards, so like two months I was eating rice and no other foods like rice and bread, which sucked. No salt, no nothing, no cheese, no dairy, no milk. I miss milk so much. I was worried I wasn't gonna be able to drink milk anymore. Because that's like one of the things like if you mess up your stomach really bad, I had vertigo, my stomach was all jacked up.
you can lose your ability to drink milk. And I would have been super bummed. I'm a big milk guy. Love milk.
Recently discovered, I think it's called baby chai, which is a little more milk. It's really good. I was impressed, but I'm just really happy to be here. I'm back. I've got the spark, the fire for life. I've got bounds, I'm bounds of energy that I've just, you know, pull out another one and I just, bam, I've got another 40%, % times four. I'm everywhere. I am everything. I'm just messing around, but.
I just have a lot more energy and just joy and want and desire for life than I did five months ago for sure. And I'm super thankful to be back and I feel mentally stronger than I did prior. I was just surprised how long it took me to feel good again. Which it is what it is. You can't plan for these things. You can only react to them. Then I felt that I reacted very well.
Dalton Anderson (45:22.362) It's just the next step. The next step is getting serious about my future, getting serious about myself, my pursuits, my dreams, doing that in the city of dreams.
Dalton Anderson (45:39.282) couldn't name a better place. So I just got to lock in, do the things I need to do, execute my plan, and everything will fall into place. I will will it into existence that I can assure you. Well.
I appreciate you listening in to this episode and of course, wherever you are, good morning, good afternoon, good evening. Thanks for listening in and talk to you next week. Goodbye.
SourcesFollow the source trail.
E102 Sources
Preserved episode evidence
[[E102 - Transcript - dalton-take-04 (Dropbox copy 1)]] is the canonical raw source. It preserves Dalton's 2025 recap, opinions, personal experience, and recording-era understanding of synthetic media, AI search, data centers, property insurance, algorithmic pricing, startup ideas, books, accountability, illness, and moving to New York.
The transcript is a discovery source, not a factual authority for technical, legal, environmental, insurance, financial, medical, or psychological claims. Several statements require correction before publication.
Synthetic media and provenance
deepmind.google/models/synthid
Google DeepMind's current SynthID page says the watermark applies across supported image, video, audio, and text systems and is designed to survive common modifications. This corrects the transcript's suggestion that cropping or ordinary pixel edits necessarily destroy it.
deepmind.google/blog/identifying-ai-generated-images-with-synthid
The original technical announcement describes capabilities and limitations. It explicitly says the system is not foolproof.
The Coalition for Content Provenance and Authenticity defines its technical-standard mission. C2PA records provenance and integrity assertions; it is not a universal detector of truth.
spec.c2pa.org/specifications/specifications/2.3/explainer/_attachments/Explainer.pdf
This is a current primary explainer for Content Credentials and should control technical descriptions during drafting.
NIST AI 100-4 separates provenance tracking, watermarking, metadata, content authentication, and synthetic-content detection. The report was updated in April 2026 and supports the article's evidence hierarchy.
Search and AI discovery
developers.google.com/search/docs/fundamentals/ai-optimization-guide
Google's guide, updated July 10, 2026, says SEO remains relevant to AI Overviews and AI Mode. It recommends original, non-commodity, people-first content and rejects requirements for llms.txt, special AI markup, forced chunking, a separate AI writing style, or special structured data.
developers.google.com/search/docs/appearance/ai-features
Google states that the same foundational SEO practices apply to AI features and that no special AI markup or separate machine-readable file is required. This directly limits any claim that GEO replaces SEO or that pages should be written for machines instead of people.
developers.google.com/search/blog/2023/02/google-search-and-ai-content
Google's guidance emphasizes helpful, original, high-quality content regardless of how it is produced.
developers.google.com/search/blog/2026/06/gen-ai-performance-reports
This current Search Console announcement documents dedicated generative-AI visibility reporting in a limited rollout. Availability and interface details are dynamic.
Data centers and community infrastructure
energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers
DOE summarizes the 2024 LBNL report and its U.S. electricity-use estimates. Forecasts are scenarios, not certainties.
energy.gov/powering-americas-ai-future-data-center-resource-hub
This is DOE's current resource hub and contains later projections. Figures should be dated and tied to the named report.
gao.gov/assets/gao-25-107172.pdf
GAO reviews generative AI's environmental and human effects, including energy, cooling, water, and uncertainty. It is the appropriate federal source for broad resource categories.
EPA's stationary-engine resources provide federal compliance context for certain backup generators. State and local permits still control the named project.
fema.gov/emergency-managers/practitioners/lifelines
FEMA's community-lifelines framework supports review of energy, communications, water, transportation, safety, and other connected services during disruption.
The transcript's specific statements about water volumes, chemical contamination, turbine dimensions, permits, health outcomes, and named facilities require direct utility, permit, environmental, public-health, enforcement, or court records. They are excluded until verified.
Property insurance and resilience
Swiss Re currently estimates a 2025 global natural-catastrophe protection gap of $424 billion. This makes the transcript's unsourced $200 billion figure unsuitable as a current fact.
NAIC's strategy supports the connection among home hardening, catastrophe models, risk communication, and resilient insurance markets.
content.naic.org/research/resilience-data-visualization-map
This current NAIC resource documents state mitigation grants, discounts, and related programs. Eligibility and benefits vary by jurisdiction.
fema.gov/emergency-managers/risk-management/hazard-mitigation-planning/requirements
FEMA explains mitigation planning and grant-program relationships. It does not promise funding to an individual homeowner.
nibs.org/wp-content/uploads/2025/04/NIBS_MMC_MitigationSaves_2019.pdf
The National Institute of Building Sciences report estimates benefit-cost ratios for several national mitigation categories. Its averages do not establish the value of one property project.
ibhs.org/wildfire/wildland-fire-embers-and-flames-home-mitigations-that-matter
IBHS describes system-based wildfire mitigation across roofs, vents, decks, fences, walls, and the near-building zone. It does not promise that a particular retrofit prevents loss or lowers a premium.
Startup ideas and books
ycombinator.com/blog/startup-school-videos
This identifies Jared Friedman's Startup School session on getting and evaluating startup ideas. The transcript's "SUS Learnings" reference should be corrected to Y Combinator's Startup School.
ycombinator.com/people/jared-friedman
This current profile links the original idea talk. The seven recipes should be checked against the recording before publication rather than copied from the transcript without attribution.
ycombinator.com/library/8g-how-to-get-startup-ideas
This is the original Jared Friedman talk and controls the seven idea-generation routes.
steveblank.com/2010/01/25/whats-a-startup-first-principles
Steve Blank's practitioner account supports the idea of a startup as a search for a repeatable and scalable model. It is not empirical proof that one discovery method predicts success.
harpercollins.com/products/the-hard-thing-about-hard-things-ben-horowitz
The publisher page establishes Ben Horowitz's book. The public guide uses a brief attributed lesson and does not reproduce the work.
csrc.nist.gov/pubs/sp/800/61/r3/final
NIST's current incident-response guide supports preparation, containment, communication, recovery, and improvement as transferable operating disciplines. It is a cybersecurity guide, not a universal management standard.
dash.harvard.edu/entities/publication/13a7b031-0fdd-45ec-a7e0-2b80e2bc679f
Amy Edmondson's 1999 study introduces team psychological safety and reports its relationship with learning behavior in 51 work teams.
pubmed.ncbi.nlm.nih.gov/33749936
This systematic review and meta-analysis covers self-compassion-related interventions and self-criticism. The authors report moderate study quality and heterogeneity.
pubmed.ncbi.nlm.nih.gov/31864153
This meta-analysis reports an association between higher pretreatment self-criticism and poorer psychotherapy outcomes. It does not turn the public guide into therapy.
pubmed.ncbi.nlm.nih.gov/22708885
This randomized study examines self-distancing during reflection on negative experience. Its population and clinical context limit generalization.
cdc.gov/niosh/bulletin/2020/fatigue-work.html
CDC NIOSH guidance supports the principle that insufficient recovery can allow work-related fatigue to accumulate. It is not medical advice for Dalton or the reader.
The Hard Thing About Hard Things by Ben Horowitz and Can't Hurt Me by David Goggins are copyrighted books. Any quotation must be brief, edition-checked, attributed, and necessary. The public pages should favor Dalton's application of the ideas over book summary.
Corrections and exclusions
The transcript's claim that the FTC fined Instacart $60 million days after an article was not substantiated and will not be republished. E099 contains the dedicated pricing evidence ledger.
"Synthetic Rubicon" will not be described as an established technical term without a traceable source. Visual artifacts and detector scores will not be presented as proof that media is authentic or false.
GEO will not be presented as a replacement for SEO, and content will not be optimized for machines at the expense of people. Current official Google guidance says helpful, reliable, people-first content and ordinary SEO fundamentals remain relevant.
The insurance page will not offer coverage, pricing, engineering, or mitigation advice for a specific property. The personal pages will not give medical or mental-health advice, glorify self-abuse, or present a difficult recovery as a universal performance method.
Draft-time checks
Recheck every figure, product capability, law, grant, and program immediately before publication. Find primary local records for any named data-center case. Verify book wording against the exact edition. Let Dalton decide how much personal health detail belongs in the Episode Story.