So, something big happened yesterday. There was a massive viral article, and do you know how big the something big is happening? How much that grew?
It went big. I was getting texts about it.
It has over 100,000 likes now. It's quickly becoming the most read thing in the world. I don't know, it's crazy.
Okay. And they're saying that only 200 people actually read it, but everyone has an opinion. Everyone has an opinion. I had an opinion. I wrote a piece about it yesterday comparing saying AI is not like CO. I kind of got one shot by the CO analogy just from a mathematical perspective. But the best reaction to something big is happening, it's got to be John Palmer. Something small is happening. I was reading this this morning actually laughing out loud. So we'll read through it. John Palmer says,
Yeah, you were reading it to yourself and then you just started belly laughing, and so then I started filming, and then I asked you what you were reading. I was filming you. Then I sent that to John.
To get his reaction, you reacting to his.
Think back to 2014, a little over a decade ago. If you wanted a burrito, you had to get in your car and drive to the restaurant. You had to stand in line. You had to talk to a person. If someone told you that one day a stranger would bring you a burrito to your front door because you pressed a button on your phone, you would have said, 'That sounds dystopian and also incredible.' Then over the course of about three years, it just became normal. And this was before Chipotle started skimping out on the portions, too. I think we're at that same inflection point right now, except instead of burritos, it's your entire job.
I've spent the last two weeks reading about AI and watching clips of people building things with it. I live in this world now. The future is being shaped by a remarkably small number of people. A few hundred researchers at a handful of companies. I'm not one of these people, says John Palmer. I work in crypto, but I am close enough to feel the ground shake. I follow several AI researchers on X, and I also recently purchased a Mac Mini. I haven't set it up yet, but I think you can understand what I'm saying. I know this is real because it happened to other people first. Here's the thing nobody outside of tech understands yet. The people sounding the alarm aren't making predictions. They're telling you what already happened to them. I am relaying this information to you secondhand with conviction. For years, AI had been improving steadily. I absorbed these improvements fine because I wasn't really paying attention, to be honest. Then apparently in 2025, everything got much faster and then faster again. I know this because I watched a three-hour podcast about it last Wednesday. The host was visibly shaking. I could tell because he wasn't even able to finish a single bite of Guinness. Then on February 5th, two major AI labs released new models on the same day and something clicked. Not for me personally. I was at a birthday dinner for my wife's coworker, but for the people who tried the models that day, it was apparently a huge deal. One guy said he tells his AI what to build, walks away for a few hours, and comes back to find the work done. This will very likely be what my life is like as well. Once I set up my AI will be doing things while I'm in the other room. That's potentially huge. Let me give you an example so you can understand what this looks like in practice. The guy on Twitter built an entire app just by describing it. The AI wrote the code, opened the app, tested the buttons, decided it didn't like the layout, fixed it, and only then said, 'It's ready. I'm going to do something. I'm going to do something like that with my Mac Mini. Some type of cool app or something.' I'm still figuring it out, but I have all the hardware and I'm definitely going to optimize my setup with all types of skills and plugins. I'm not exaggerating. That's what my Tuesday could look like potentially.
But I tried AI and it wasn't that good. People will say this. They say this, I tried AI, it wasn't that good. Here's John Palmer's response. He says, I hear this constantly. I understand it because I also thought this, but the models today are apparently unrecognizable from what existed six months ago. Some of the people who are saying this have hundreds of thousands of followers, and make sure you're not on the free plan. You have to get on the paid plan.
How fast is this actually moving? Let's answer that question. In 2022, AI couldn't do basic math. By 2023, it could pass the bar exam. By 2024, it could write working software. And by 2025, some of the best engineers in the world had handed over most of their coding work to AI. In February 2026, you can make AI send you a morning summary of the top posts on Reddit from your Mac Mini. What should you actually do? I'm not writing this to scare you, says John Palmer. I'm writing this because the single biggest advantage you can have right now is being early. Early to understand it, early to buy the hardware, early to subscribe to the paid tier and ask it to make a meal. You can start using it for real work. If you're an engineer, give it a meal. A meal plan.
A meal plan. Oh, a meal plan. Yeah, actually make a meal. That would be cool.
If you're an engineer, give it a GitHub repo. If you're in finance, give it a messy spreadsheet. If you're in other industries, definitely figure out what you can give it, then give it, but give it something for sure. I just need to think about it more. Mention it to your co-workers. Right now, there's a brief window where most people are still ignoring this. You can be the person ignoring it less, which puts you ahead. Be the person who walks into a meeting and says, 'I used AI to do this analysis in an hour instead of three days.' It's fine if it actually took you three days. The point is people will respect you. Get a Mac Mini. This is point number three. This
And by the way, you can actually get a Mac Mini on DoorDash.
You can just like a burrito.
Every single account I've seen posting this stuff has a Mac Mini. It's the ultimate tool for this stuff. You got to get your financial house in order. I'm not a financial adviser, says John Palmer. That said, I did just spend $599 on a Mac Mini plus Apple Care and $50 on Claude tokens. But if it gets too expensive, have a backup plan to switch to an open source Chinese model. Fourth, watch the podcasts. Actually, don't just watch. Study TBPN, DoorDash, etc. The good thing is that these shows are several hours long, so you can basically fill up your entire day just watching them or making them. In our case, rethink what you're telling your kids. The standard playbook, good grades, good college, stable job. It's over. Tell them it's likely not looking good for them. Go to your four-year-old and tell them it's not looking good, buddy.
It's not looking good. What I know, I know this isn't a fad. The technology works and the richest institutions in history are pouring trillions into it. I know this because I've seen it mentioned pretty frequently over the past few months. I know the people who will come out of this best are the ones who start engaging now, not with fear, but with curiosity, a sense of urgency, and ideally a Mac Mini.
The future is already here. It just hasn't knocked on your door yet. It's about to. And when it does, I will be ready. My Mac Mini will be unboxed. My agent will be configured. I will describe what I want in plain English and it will appear. I just have to optimize everything first and make sure my setup is super legit. Most people won't want to hear it until it's too late. You can be the reason someone you care about buys a Mac Mini.
Such a banger. So, so good. I went and I described in plain English to Codex 5.3. Build me an app. An amazing app. The best app ever. One that will win awards and make me a legend in Silicon Valley.
An award-winning app. And it appears to be working.
More seriously, John Palmer's post feels like a response to Will's essay, Toolshaped Objects, which is fantastic. Anthropic is nearing the completion of a deal to raise more than 20 billion in a funding round led by investors including Peter Thiel's Founders Fund, Dario, and Dragon. Let's dig in. Being a little bit silly with the pronunciation. Sydney over on X said British AI researcher be like Anthropic. Anthropic is nearing the completion of a deal to raise more than 20 billion in a funding round led by the group I just mentioned. The deal is set to value Anthropic at about 350 billion, nearly doubling its prior value. The funding round includes a range of investors including CO2, Singapore's GIC, Microsoft, Nvidia, and it has become a who's who of Silicon Valley and Wall Street.
So Founders Fund is known for the monopoly thesis, like there's power law. You want to be in the best company in the category. You want the best ownership.
Put all your money in that one.
Yeah. I'll go all in. Famously, they did some other social networking deals, but obviously like Facebook was the big one at seed and that was enough to return that fund many times over. And so the whole thesis of the fund for a long time has been like this concentrated bets, concentrated bets. AI has been an interesting one. They're now in three major labs because they have a huge position in SpaceX, which now owns XAI. And so they're in that lab. Sort of coincidentally, Peter Thiel, I believe, is listed on Wikipedia as a co-founder of OpenAI or at least he was like an initial donor and got like a co-founder title and so has been involved in OpenAI for a long time and then they invested I believe in like the $60 billion round right after right when ChatGPT was launching and it was an odd deal because the company had not yet converted. It was very unclear what Microsoft would get, what the employees would get, what the investors would get, what the nonprofit would get because the nonprofit was going to keep a stake. And so it was this like wild bet. And I know a lot of investors, it was a leap of faith. A lot of investors were scared off because there was, you know, what am I investing in? Like I'm not getting shares, I'm getting units. And it was like a very odd structure. And I know many investors that were scared off by that.
I'm an absolute unit, but I'm used to buying shares.
Yeah. Exactly. So you had to do a lot of sort of leaps of faith. The thesis at the time, how big can this be? What is the market? What are the opportunities for really really huge categories? A lot of Founders Fund's investments like SpaceX. It's like this massive category that's very very hard, but if you crack it, it's going to be an entirely new market. Social media was certainly that as well. That made a lot of sense. Obviously, they're in X, now they're in Anthropic. Sequoia is also in all three. And the interesting thing is like PT has given a number of talks where he's sort of been anti-AI. You see this whole thing about like crypto is libertarian, AI is communist and that AI is like a consolidating force. He's never been one to say like oh the tools are so amazing. It's been a lot of like is this just doing your homework for you?
He's not putting out threads about using his Mac Mini.
I don't even know. I don't even know if he has a Mac.
You should send John Palmer's piece to PT. So yeah, I mean reading into this, does this say that the Founders Fund team thinks that there's like a divergence in the market that there's actually two markets going on here? One in consumer AI advertising, knowledge retrieval, the other in code and developer tooling and codegen and that Anthropic has actually carved out a separate space. So they're less competitive than people think. Maybe they're just more competitive and it just makes sense to get into both.
And we have some breaking news.
What's the breaking news?
Anthropic has formally announced the round.
They said we've raised 30 billion. So they upsized it.
They are now at a 14 billion run rate.
With this figure growing over 10% annually in each of those past 3 years. We'll see if they got another 10x in them.
The biggest update for me is that the forward deployed engineer thing is like very real. Even if the tools can build themselves, there's still so many blockers to actually rolling these out that you're going to need a lot of people inside these organizations who are actually playing with Mac Minis on the weekend and getting obsessed with these tools and then bring them into the enterprise because it's very easy for a lot of people in Fortune 500 companies to just show up and do their job the way they always do it because no one gets fired for switching things up. But if you roll out some tool and it doesn't work the way you intended or it sucks up a bunch of your time even if you're just I have eight hours of work to do in my traditional system where do you get the extra hours to automate your work? It takes time and that's where the forward deployed engineers come in. So great gig for anyone who wants to go implement enterprise agentic workflows because that will be the theme.
We can see Anthropic's run rate revenue growth. See here, quite the pop from $0 in 2023 to 100 million in Jan 2024 to a clean 1 billion in January 2025 and now sitting at 14 billion.
This is acceleration. This is true. Are you feeling it yet?
True acceleration. Dario sort of underplayed 2027 I guess or I guess he said that he was like I don't think we have another 10x if they go to a hundred billion. Well, he said like we've been 10xing. Yeah. You know, we'll see if we do it again. Yeah. Like wink wink.
If they hit 140 billion at run rate by the end of the year, that would be insane.
But strange times. Strange times. You love this effect.
I mean, Tyler's been saying stuff along these lines for some time. So more in Anthropic, Nathan Lambert was quoting their announcement from yesterday saying, 'We're committing to covering electricity price increases from our data centers.' This is Anthropic.
To ensure rate payers aren't picking up the tab. We'll pay 100% of grid upgrade costs. Work to bring new power online and invest in systems to reduce grid strain.
Maybe this should have been their Super Bowl ad. When you kind of look at clearly how much energy the industry put into the Super Bowl and what the net effect was for the average American. I don't think the average American came away from that like having more positive feelings towards AI. If anything, they were like, 'I'm sick of this stuff.'
Yeah. Yeah. Yeah. The meme was like, 'Oh, beer ads like refreshing.' It's pretty easy to say, 'Hey, my electricity bill is going up. I don't want a data center in my county in my city.' And you call your city county representative and they say, 'Yeah, I don't like the slop either. I see a bunch of junk that doesn't seem that valuable. Why do we need it here? It's not going to create that many jobs. It's not going to be that good for taxes. Blah blah blah blah blah.' And other companies that do this, too. Google had their clean transition tariff and Amazon also pays a surplus above electricity costs already. The anti-AI issue is remarkably bipartisan. Data Center Watch claims that 55% of Republicans and 45% Democrat in affected districts. So 55-45. So it's not really like a right issue or a left issue. Like there's push back on both sides. So Anthropic is putting $20 million into a super PAC operation that is meant as a direct counter to the OpenAI super PAC operation. The midterms are a battleground between these two rival AI labs, says Teddy Schleer at the New York Times. And so we're going to see some attack ads. It's like Sam and Dario are running for election and we're going to see some be the president of AI.
Yeah. Yeah, but I mean we just said that like the overwhelmingly on both sides there's like a lot of push back, right? So it's like what are the actual differences between their approaches? Yeah, that is interesting.
Maybe I guess maybe Anthropic will be more like about the China chip issue.
Yeah, it feels like there they should be more in alliance than against each other.
Yeah, it's like crypto there's a crypto super PAC. Yeah, there's like the big one. There's no like big AI one right now. Yeah, it would be weird if there was like a USDC super PAC and then a USDT super PAC and they were like fighting each other and just constantly surfacing bodies that are buried in the closets or skeletons of the closet because every time there's a successful attack on OpenAI yes people might be like okay I like Anthropic a little bit more but most people just come away being like I don't like AI. If you were falsely accused of a crime and it went to trial, who would you prefer to listen to the arguments and give the verdict? And the options are a jury of your peers or Claude Opus 4.6. 63% of people prefer Claude.
Might be the audience here.
I think the jury of your peers will be around for a very very long time. I mean, they still have courtroom stenographers and we have
I think GLaDOS is going to be around a long time, too.
Yeah. Okay. If you're thinking in centuries, maybe, but I mean, the courtroom sketch artist, we've had cameras for a hundred years. Like, the juries are
I'm not saying that I think this will ever happen.
Okay. But you'd prefer it. You would ideally I mean, how many jurors will be sitting there being like, 'Okay, Claude, how should I vote? I haven't really been paying attention. Like, I've been playing Candy Crush.'
