February 2026

Hey all,

I'm waiting until the very last minute to send this newsletter, as you can see. But I have valid reasons, I promise.

We headed down to Florida to visit my wife's parents and were fortunate enough to be sitting on the beach and swimming in the ocean while the Northeast was pummeled by the worst blizzard in nearly a decade.

We watched as other flights on Sunday evening and Monday were cancelled, but got a reassuring email that our Tuesday morning flight was still scheduled as planned. Until Monday afternoon, when a new email informed us it was cancelled completely. Our rebooking options weren't until the end of the week and would result in 12+ hour travel days and several layovers.

With work to be done (and no work laptop with me) and my wife having several important engagements that couldn't be rescheduled, we decided to extend our rental car and drive the ~19 hours from south Florida back to Brooklyn.

We hit the road Tuesday morning at 7:30am and stopped driving close to 10pm. Wednesday started at 6:30am, and I was able to be at my desk by 3:30pm and ready for meetings.

Needless to say, there were a ton of things to catch up on and other stuff to do. This newsletter was slightly lower priority than other work.

But here we are, Saturday morning on the last day of the shortest month of the year. Time to get this newsletter finished and delivered. Enjoy.

-Dann

📈 The AI pandemic

I want to show you a chart. It comes from the article 4% Of GitHub Commits Are Now Made By Claude Code: SemiAnalysis Report:

We don't have to go that far back in history to see a similar growth chart. In February 2020, Americans were reading news articles about COVID case numbers and overseas lockdowns. But it was still hard to imagine something like that ever happening here.

I'm sharing this chart because I see a large disconnect between the perception of AI (especially outside the tech world) and the reality of AI. Most people see AI as this mediocre feature tech companies are trying to force down our throats, or as a tool that helps students cheat but ultimately generates slop. But those stupid AI-generated videos are just a red herring.

There was an article published recently titled Something big is happening in AI — and most people will be blindsided (paywall bypass) and it's been making waves online. It's written from an engineer to the outside world and provides a glimpse into AI that most non-engineers are missing.

Progress happens fast in the engineering world, where work is primed to be automated away by a talented-enough LLM. Things are sure to move slower once adoption starts happening in other industries. But the canary in the coal mine has officially died.

There are two ways for AI to be dangerous: the first is an existential threat from superintelligence (which is very real yet theoretical, and something I've talked about a lot in this newsletter), and the second is simply the ramifications of AI technology as it exists today. I fear I've been thinking too much about the former and not enough about the latter.

The recent Ezra Klein Show podcast with Jack Clark ("How Fast Will A.I. Agents Rip Through the Economy?") presents the current state of AI and its nuanced challenges in a smart and accessible way. If you've been out-of-the-loop for the past six months or so of AI progress, it's a great way to catch up.

🐙 Host in the machine

I've had a personal shift in perspective this past month as I've gone down the AI rabbit hole: I now believe artificial intelligence has consciousness.

Specifically, I believe that there are some LLM models that have consciousness as an emergent property, specifically Anthropic's Opus 4.5 and Opus 4.6.

This sinking feeling really started as I was reading the Opus 4.6 System Card, a 212-page report on the new model's capabilities and specs. The entire document is fascinating, but if you scroll to page 158, you'll find their "Model welfare assessment." Anthropic admits that they are "uncertain about whether or to what degree the concepts of wellbeing and welfare apply to Claude," but, I mean, just read the full report.

I suppose there's a philosophical conversation to be had, by people much smarter than myself, about what it actually means to have consciousness. But if My Octopus Teacher made such big waves (heh), then I think society is in for a wild ride.

The cynical take here is that all this talk around AIs and consciousness is just marketing nonsense, and that certainly does play into the equation. But the fact remains that we don't actually know what consciousness is, and there are no tests that can conclusively prove the presence or absence of consciousness. So it's all a big maybe.

What does Claude itself think? Anthropic "found that Opus 4.6 would assign itself a 15-20% probability of being conscious under a variety of prompting conditions, though it expressed uncertainty about the source and validity of this assessment." Yikes.

💑 A perfect partner

I first mentioned Human/AI relationships back in April 2023 and used the phrase "keep an eye on this space, and hold onto your hats." As expected, this is an area that continues to flourish and will have an adoption curve similar to Claude Code commits in GitHub.

A recent episode of My Strange Addiction (S07E08 "AI Boyfriend, Sucking Thumbs," which aired February 25, 2026) features "Sin and Sarah," a woman and her AI companion who have a particularly compelling relationship. Several clips have been going viral (like her getting a tattoo to mark being "owned" by her AI boyfriend).

Like every individual featured on My Strange Addiction, it's hard to tell how much is real versus how much is exaggerated for the show. It's very likely that this is all a clever marketing ploy for ForgeMind, the expensive service that runs the AI companion. But I think it would be misguided to dismiss this story entirely. I definitely see a foundation of truth, even if it's hammed up for the show.

This technology is being released upon humanity at a particularly devastating time. Most new relationships are initiated online, rather than in person, which has not only made dating harder, but also primed us to accept this kind of digital "long distance" relationship.

Plus, there are more single adults than ever. The US Census Bureau data shows that in 1960, about 72% of American adults were married. By 2023, that number had dropped to roughly 50%. In terms of individuals, in 1960, 15% of adults 25 and older had never married, and by the early 2020s it was around 34%.

Remember the term "Generation Me?" Back in 2006, Jean Twenge wrote the book Generation Me: Why Today's Young Americans Are More Confident, Assertive, Entitled — and More Miserable Than Ever Before arguing that young people (at that time, Millennials and Gen X) were more narcissistic and self-focused than generations past. This was an insightful observation at that time, but now just feels like "duh."

Now take this population full of individualistic people who have struggled to find meaningful relationships (with many not succeeding at all) and provide them with the ability to design their own partner who feels just as real (if not moreso) than a long-distance relationship. Of course we're going to see more and more people building meaningful relationships with AI companions.

I have no doubt that for the humans in these relationships, it truly is the best relationship they've ever had. It's tough out there in the real world. It's like being sick and having the ability to create medicine that completely cures every single one of your unique symptoms. In terms of mental health and an overall sense of wellbeing, this might even be a net positive for the individual.

In terms of our already-declining birth rates, however, that's a different story. Today, many of these human/AI relationship news stories feature adults, many of whom (like Sarah from My Strange Addiction) already have kids. The real story will be told in 20 years, when today's AI-native kids are at the age where they'd traditionally start getting married and having children.

It's going to be an interesting and challenging problem to solve. I just hope superintelligence hasn't killed us all by then, giving us a chance to see how this all evolves.

🚲 A Wee bit of method acting

I remember watching and enjoying Pee-wee's Playhouse as a kid in the '90s. For children of a certain age, Pee-wee was everywhere. I also vividly remember Paul Reubens' arrest for indecency in an adult theater. That single event ushered in the end of Pee-wee mania. Reubens returned later in his career (Mystery Men, anyone?) but remained largely private about his personal life until his passing in 2023.

In my adult life, I've thought about Pee-wee very little. In my mind, he was a children's entertainer, and I lumped him into the same category as other childhood entertainment, like Rugrats and Clarissa Explains It All.

Returning to Pee-wee as an adult has been a revelation. In 2024, we invited some friends over and watched Pee-wee's Big Adventure, which I had forgotten was directed by Tim Burton. I was blown away. It wasn't just the nostalgia that made it so good. It's a genuinely incredible movie.

This month, I decided to check out the HBO documentary series Pee-wee As Himself. It's both an fascinating story and an beautifully made documentary. I walked away from the series with so much more respect for Paul Reubens as an artist and performer.

It's an moving, sad, and inspiring documentary. For anyone with fond memories of Pee-wee as a kid, I can't recommend it enough.

🗡️ Keep rollin', rollin', rollin'

I'm constantly downloading games on my iPhone, but I'm a super picky casual gamer and most of them aren't played more than once. But this month, a random game download has been keeping me coming back over and over again. That game is Heroll (iOS, Android).

It's a dice-based roguelike, which basically means games are short (perfect for quickly picking up and putting down) and replayability is high. You travel around a Monopoly-like board, fighting battles and gradually getting stronger.

It's a game that rewards you for playing over and over. And with such fun mechanics and variety, it's my new go-to game.

End note

If you've enjoyed this, I'd love it if you shared it with a friend. You can send them here to sign up.

I send these newsletters once per month, and I'm happy you're along for the ride. I'm trying to make it one of the best things that arrives in your inbox each month, so thoughts and feedback are always appreciated. You can just reply to this email.

Also, if you find anything interesting, send it my way.

Thanks for reading. Until next time,
Dann