Love in the Time of Robots: The 314th Evolutionary Lens with Bret Weinstein and Heather Heying
Bret Weinstein and Heather Heying explore Valentine’s Day myths, debunking the "soulmate" ideal as unrealistic in modern dating while framing love as a win-win partnership shaped by effort. They pivot to a JAMA study (2023) linking 2–3 cups of caffeinated coffee/day to a 57% lower dementia risk, dismissing decaf’s benefits and questioning prior research oversights. AI concerns arise with OpenAI’s GPT-5.3 Codex (Feb 5), allegedly self-debugging, sparking debates on singularity risks—yet they reject Luddite comparisons, warning of wealth gaps and lost human meaning in automation. Ultimately, the episode blends evolutionary biology, cognitive health, and tech disruption to argue that meaningful relationships and ethical progress demand deliberate adaptation over blind optimism. [Automatically generated summary]
It took me a second to remember what podcast I was on.
I sometimes do other ones from behind this desk.
But this is the Dark Horse podcast live stream.
I am Dr. Brett Weinstein.
You are Dr. Heather High.
It's the 314th, I believe.
Yes, it is.
So I got all the information in there.
It wasn't in the right order.
Perhaps not as professionally delivered as possible, but, you know, there's always next time.
There's always next time.
Here we are on our rare Saturday live stream.
It's Valentine's Day.
We'll be starting off by talking a little bit about love and Valentine's mythology, you say.
And then AI, a natural follow-on, really not.
And just a bit new finding about coffee and how freaking great it is for you, honestly.
Nice.
I'm glad to hear that.
It's great when your vices turn out to be A-O-K.
No, not a vice.
I have been certain of this from the beginning.
And when I found out I was pregnant for the first time, not having expected it and immediately went looking into like what can I, cannot do, the evidence on coffee was remarkably consistent that small amounts of coffee were just fine in pregnancy and in everything else.
Yep, it's good.
It's not quite as good as that time that we all collectively discovered that meat was health food, but it's good.
Love coffee.
Again, neither you nor I discovered that meat was health food.
Well, let's put it this way.
The degree to which not only is meat just fine for you to eat, which you and I have known forever, but that actually a lot of meat was a good thing to have in your diet.
Yeah, I don't, I don't, I feel like I was there.
Sometimes I say things, you're like, nah, I was there.
Like in this case, I'm like, no, I was, I was, I was there.
And, you know, partially it's having grown up the daughter of a man who had been a farmer and who insisted on a lot of meat, but it just, it felt like the honest, good food that needed to be present a lot of places.
Well, let's just say I was eating as if it's the right thing to be eating.
I just believed too much of the hype.
And it's not the only time that's happened to me.
But so anyway.
It's happened to all of us.
Yep.
Seven to all of us.
Okay, so this is going to be our last February live stream, except for tomorrow.
We will do a Q ⁇ A on locals.
Consider joining us there, 11 a.m. Pacific for two hours.
And then we're going to have a, Brett will have a couple of great inside rails to keep you guys occupied for the rest of February.
Not third rails, but inside rails.
Sometimes they are on a third rail.
Did I say third rail?
No, you didn't.
I just sometimes think when people hear inside rail, it turns out the reference is a little obscure.
You know, the horse coming up the inside rail, actually the dark horse from Benjamin Disraeli's novel from which the term dark horse comes.
This is a horse that nobody knows how to bet on.
And so you can imagine it coming up the inside rail and people are thinking, what is that horse?
What does this mean?
Is it going to win?
That kind of thing.
Sometimes, yes.
Sometimes the dark horse wins.
That's the thing.
Yep.
Yep.
Exactly.
All right.
Let us have our sponsors right up front, our three sponsors, all of whom we think are fantastic.
Truly vouch for.
And Brett's going to start us off.
Non-Toxic Cookware Solutions00:05:39
All right.
Heather, our first sponsor this week is Caraway, and they make high-quality, non-toxic cookware and bakeware.
These glasses did not go.
But we're back.
Everything is aligned.
Yeah, non-toxic cookware and bakeware.
That is truly excellent.
Maybe you made a New Year's resolution to eat better or cook more or decrease your exposure to toxins.
You can do all three at once by cooking with Caraway.
We're in the cold season now, a time for warming soups and stews, big braids, slow-cooked cuts of meat and roasted root vegetables, period.
How about baking something delicious for your love on Valentine's Day?
A decadent chocolate cake, shatteringly thin ginger cookies, or brightened chewy lemon bars.
With Caraway, all of this deliciousness from roasting to bakefast.
Wow, this is surprisingly complex.
Well, I think I can tell from the fact that you didn't have any response to the decadent chocolate cake, the shatteringly thin ginger cookies, or the bright and chewy lemon bars that you're just seeing words and noticing them without actually reading.
I was thinking of interjecting that all of these things are features of our household and they're spectacular.
I mean, the ginger cookies are so surprising because they look like nothing.
They look like cookie puddles, but they're so good.
Okay, so you are taking it in.
Yeah, I'm taking it in.
I'm just the buffer is it's straining to accommodate all of the content.
But I digressed, as you all noticed.
Caraway is fantastic.
It's great cookware and bakeware.
Yeah, it is.
Exactly.
The cookware and bakeware is functional, beautiful, non-toxic, and easy to clean.
What more could you want?
All right.
Modern life is full of hazards, not least the non-stick coatings on cookware and bakeware.
We threw out all of our Teflon cookware decades ago because Teflon is toxic.
Yet over 70% of cookware in the United States is made with Teflon, and 97% of Americans have toxic chemicals from non-stick cookware in their blood.
When you cook with Teflon, it only takes two and a half minutes for a pan to get hot enough to start releasing toxins.
Enter Caraway.
Caraway kitchenware is crafted with sustainable non-toxic materials like FSC certified birchwood, premium stainless steel, enameled cast iron, and naturally slick ceramic to help you create safer, to help you create a safer and healthier home.
Caraway makes several lines of non-toxic cookware and bakeware.
Our favorites are their stainless steel line and their enameled cast iron.
All of Caraway's products are free from forever chemicals, and their enameled cast iron is offered in six stylish and beautiful colors.
These pots are strong and highly scratch resistant.
They'll last generations.
And Caraway also offers butcher blocks to cut on, glass lids for non-toxic cooking with a view, and a new bar set, which is crafted from rust-resistant 304 stainless steel.
We're cooking with Caraway, and now Zach, our elder son, is two in his first college apartment.
He says it's amazing, which we know to be true, and we know that he will be cooking with it for a long time to come.
Caraway's cookware set is a favorite for a reason.
It can save you up to $190 versus buying them individually.
Plus, if you visit carawayhome.com slash dh10, you can take an additional 10% off your next purchase.
This deal is exclusive for our listeners, so visit carawayhome.com slash dh10 or use the code dh10 at checkout.
Caraway, non-toxic kitchenware made modern.
Totally recovered.
Absolutely.
No one remembers that anything happened.
I have a feeling that's not true, but thank you for that little bit of comfort.
Well, our second sponsor, Brett, this week is Armor Colostrum, an ancient bioactive whole food.
Here at Dark Earth, we talk frequently about the fact that we live in an age of hyper novelty.
Humans are the most adaptable species on the planet, and even we can't keep up with the rate of change that we are enacting on ourselves.
We are bathed in electromagnetic fields, artificial lights, seed oils, microplastics, hopefully not literally bathed in seed oils, but metaphorically at least.
Stranger things have happened.
Very much so.
And more dangerous microplastics, endocrine disruptors in our air, water, food, and textiles.
And there are a myriad other modern stressors like overcrowding and having too little control over our own choices in life.
Here's something you can't control.
Strengthen your immune health with a bioactive whole food that is armor colostrum.
All of this hypernovity can disrupt the signals that your body relies on.
Negatively impacting gut, immune, and overall health.
Armor colostrum works at the cellular level to bolster your health from within.
Colostrum is nature's first whole food, helping to strengthen gut and immune health and fuel performance.
Armor Colostrum is a great add to smoothies.
I love it with mint and banana and cacao and raw milk.
Bovine colostrum can support a healthy metabolism and strengthen gut integrity, and ArmaColostrum is a bioactive whole food with over 400 functional nutrients, including but not limited to immunoglobulins, antioxidants, minerals, and prebiotics.
Armor Colostrum starts with sustainably sourced colostrum from grass-fed cows from their co-op of dairy farms right here in the United States, and they source only the surplus colostrum after calves are fully fed.
Unlike most colostrums on the market, which use heat pasteurization that depletes nutrient potency, Armor Colostrum uses an innovative process that purifies and preserves the integrity of hundreds of bioactive nutrients while removing casting and fat to guarantee the highest potency and bioavailability.
The quality control is far above industry standards, including being certified to be glyphosate-free.
People who have used Armor's Colostrum have reported clearer skin, faster and thicker hair growth, and better mental concentration.
In addition, people using Armor's Colostrum have noticed a decrease in muscle soreness after exercise, better sleep, and fewer sugar cravings.
Armor Colostrum is the real deal.
Why CrowdHealth Rocks00:03:29
We've got a special offer for the Darkhorse audience.
Receive 30% off your first subscription order.
Go to armor.com slash Darkhorse or enter Darkhorse to get 30% off your first subscription order.
That's armra.com slash Darkhorse.
Finally, Brett.
Still here.
You do that to me somewhat often, and I've never done it to you before, so here you go.
It always seems like the right thing at the time.
That's all I can say.
Yeah, well, our final sponsor, Brett, this week is CrowdHealth, Brett.
I'm not going to keep doing that.
No.
No.
I don't think you will.
CrowdHealth is not health insurance.
It is better.
Health insurance in the United States is a mess, to put it mildly, from overpriced premiums to confusing fine print, endless paperwork, claims that don't get paid, customer service that is unhelpful and hostile.
These complicated systems aren't functional, and they wear us down.
We used to contend with this madness, but not anymore.
There is a better way.
You can stop playing the rigged insurance game.
You can use CrowdHealth instead.
CrowdHealth is a community of people funding each other's medical pills directly.
No middlemen, no networks, no nonsense.
With CrowdHealth, you get health care for under $100 per month for your first three months, including access to a team of health bill negotiators, low-cost prescription and lab testing tools, and a database of low-cost, high-quality doctors vetted by CrowdHealth.
And if something major happens, you pay the first $500, then the crowd steps in to help fund the rest.
CrowdHealth isn't health insurance.
It's way better.
After we left our salaried jobs as college professors, we spent years buying health insurance in the marketplace.
It was awful.
Our family of four had health insurance for emergencies only, and we were paying more than $1,500 a month for a policy with a $17,000 annual deductible to a company that was unresponsive and unhelpful.
Tens of thousands of dollars paid out for no benefit whatsoever.
I went looking for alternatives, and I found CrowdHealth.
We have now had two sets of great experiences with CrowdHealth.
Our younger son, Toby, broke his foot in the summer of 2024, and I slipped on wet concrete and ate a head scan, CAT scan a year later.
Both times we went to the ER and got good, but expensive, treatment from the medical staff.
In both cases, CrowdHealth paid our bills with no hassle.
Their app was simple and straightforward to use, and the real people who work at CrowdHealth were easy to reach, clear, and communicative.
With CrowdHealth, you pay for little stuff out of pocket, but for any event that costs more than $500, a diagnosis that requires ongoing treatment, a pregnancy, or an accident, you pay the first $500 and the crowd pays the rest.
Seriously, it's easy, affordable, and so much better than health insurance, we can still hardly believe it.
The health insurance system is hoping you'll stay stuck in their same overpriced, overcomplicated mess.
Don't do it.
This year, take your power back.
Join CrowdHealth and get started today for $99 a month for your first three months using code DarkHorse at joincrowdhealth.com.
That's joincrowdhealth.com.
Use code DarkHorse.
Remember, CrowdHealth is not insurance.
Opt out.
Take your power back.
This is how we win.
Joincrowdhealth.com excellent.
Could not be happier with that service.
No, it's so good and I i've taken it out because there's so much to say and it's already a long ad read.
But um, they came to us.
Crowd Health came to us.
Uh approached us to be a sponsor when we were, when I was still in insurance land, and it didn't make sense to me.
It seemed scary and I said no and then i'd forgotten about that and later, just so frustrated with the with the insurance madness, went looking and uh found Crowd Health independently, at which point I went back to our amazing ad broker and said hey, would these guys consider being sponsors?
Myth of the Soulmate00:13:09
And he's like, are you kidding me?
They wanted to be.
And you said no uh, so you know, they were patient and we effectively discovered them independently and really could not be happier.
Yeah, it's.
Uh, it's a great service and I was nervous too but uh, very comfortable now.
Yeah, all right.
So today's topics are, um, mythologies of love, artificial intelligence and coffee exactly the big three.
Yes, all right.
Well, so I have been thinking.
I was asked, uh on uh Ravarore's podcast, what we do for valentine's day, and the answer is, you and I don't really have much of a Valentine's Day tradition, which sounds sad to other people, but I think it's actually great because the point is it's kind of you know the training wheels for something and I don't think you and I need it, you know.
So anyway, it did get me thinking though, about um, about Valentine's Day and all of the phenomena surrounding love and romance and those things, and anyway, I was inspired to try to uh, to write something that i've been meaning to, to put somewhere for a long time, and I think I met my goal.
I've uh, i've long appreciated the idea that a really good idea might be something that you can get on two sides of one sheet of paper.
It would be nice if you could get things down to being that concise.
Now, I don't know that.
I've done it.
Um, did you just fail to double side that piece of paper?
Yes, but I have glues that will fix them together here.
I'm going to even do it this way and seamlessly, in the middle of reading this, i'm going to flip.
I'm so.
It's such a trivial point, but you said that while fondling two pieces of paper, and I just wasn't sure to what degree you had failed.
The assignment will go back to the, the instant replay, you will see that I said fits on two sides of one sheet of paper.
Not that you do that, that it was on two sides of a given piece of paper.
I was just sort of talking about you know it fitting okay.
So I have derailed us with your help by not uh, double printing that piece of yours, which i've not heard yet or seen yet, on a single piece of paper.
You were saying that you think that most, presumably not all, because some ideas just require too much complexity of explanation to build the argument, but at least many important ideas should be able to be told, explained in two pages of text.
Right.
And is that six-point font you've used?
Let's put it this way.
One of the problems with writing for me is that I lose track of the outside world and was very shocked to find.
out that it was podcast time.
So uh, suffice it to say I have read through this piece once it is not edited.
Uh, I printed it and ran out the the door with it, so I discovered the small font, but I have these, um magnifiers which oh, those are cool, those are fancy.
Yeah, they are that.
So anyway, maybe I should just uh read it and then we can uh talk about.
Well, assuming you're still here, Jen, if you could have the camera just on him while he read this.
So you could sneak out if something here offends.
All right, shall we try it?
It doesn't have a title.
But this is my Valentine's Day essay.
Societies rise and fall based on the quality of their mythology.
Our mythology is badly broken, and it's time we noticed.
Mythology is the product of human storytelling.
Really good stories get retold, and they spread as the people who tell them flourish because they see the world in useful terms.
Mythology makes the counterintuitive world intuitive.
It teaches you how to live.
Myths evolve, literally.
As stories are told, retold, and translated, they improve and adapt.
You can't write a myth.
You can write a story, and if it's powerful enough to spread, it evolves into a myth.
Good ones last thousands of years.
Sometimes a powerful idea functions as a myth without a narrative to propagate it.
Myths are as fragile as they are powerful.
They can outlive their usefulness when the way we live changes, because we have moved to a new habitat, have been conquered by a superior power, because the technology at our disposal changes, or for any one of a dozen other reasons.
Or they can fall out of fashion because they seem archaic.
Or they can be captured by hostile forces and changed to enrich others at our expense.
When our mythology breaks, we suffer and fail to prosper.
This essay is about the mythological idea.
It's about a mythological idea that has become important in the West, but is now little more than a hackneyed cliché.
It is the myth of the romantic soulmate, and I want to resurrect it.
Once upon a time, this was an excellent myth.
A person lived and died within a small world with a finite number of potential mates in it.
If a young person walked around thinking that someone in the world was their soulmate, they would search their tiny slice of humanity looking for an improbably good match, the best one in the world.
They would fall in love based on the idea that it was meant to be, and form a partnership that was likely to last because the world was so unforgiving and children were produced so early that maintaining a partnership was almost always a better bet than divorce and remarriage, which was inherently humiliating and suspect.
But the myth no longer works.
In a world where there are millions of potential mates who speak your language, distributed across a vast landscape over which we easily move and in which we can resettle at will, the idea of a soulmate becomes ridiculous.
If your soulmate is out there somewhere, what are the chances you will even meet them?
What if you meet them too late and they're already taken?
What if you pull the marriage trigger too early and meet them once you've become committed to another?
It does seem preposterous.
But I believe there is a very good reason to upgrade this myth, and that its collapse has caused an epidemic of bad romantic decision-making and the spread of many appealing but wrong ideas about how and if to find a proper life partner.
Soulmates aren't found.
They are made in a way that few seem to understand.
I say this as someone who, despite having made many mistakes of my own in the process of discovery, lives every day with the total confidence that I'm married to the exactly right person.
That's not an exaggeration.
At 56 years of age, I spend zero effort wondering if I made an error, lamenting the one that got away, or pondering the person I just didn't happen to meet.
Heather is my soulmate.
And despite my many defects, I'm also confident that Heather feels the same way about me.
I don't have to worry about it because it's obvious.
So how did that happen?
See?
Two sides of the same page.
I now realize that there were two crucial components.
First, we had great chemistry from the start.
We liked each other, understood each other, and were not alike.
We were compatible and complementary.
The second component caught me off guard.
We were very young.
I met Heather in 11th grade when I was kicked out of my high school and into hers.
She was dropped dead gorgeous, and that made an impression, but, and I swear this is true, that was not the driving force, not even close.
Though with the benefit of hindsight, it helped in a strange way.
Heather had a boyfriend, a guy I knew and liked at the school that gave me the boot.
And she was clearly so far out of my league that despite having fallen head over heels in love with her at the outset, I never labored under the illusion that she would be mine.
So unlike all her male friends who had changed dramatically as she matured into a stunning young woman, I suffered in silence and delighted in our friendship.
And that got her attention in ways I would not understand for decades.
Where we get back to the question of soulmates is this.
Heather and I got together a couple years after we met, at 18 years of age.
It was still very early.
Is that wrong?
Yeah.
Okay.
How old were we?
19 or 20?
20.
Okay.
Heather and I got together a couple years, a few years after we met at 19 or 20.
It was still very early in life.
Too early, I thought, based on what I now realize is the broken mythology of modern times.
We were still kids by any definition other than the legal one.
But what's a guy to do?
I suppose I privately believed in a personal, almost upside-down version of the soulmate idea.
Someone goes to high school with the girl you should marry, I told myself, and it might be you.
The meaning I took from that was that you might well meet the best person for you before it was time to settle down, but you'd be foolish, an absolute idiot, in fact, to let it go over timing.
So when Heather decided she was ready, I was over the moon, even if I was still too young and dumb to respond properly.
We laugh about that now.
Well, she laughs, and I mentally break myself while trying to smile.
That's love for you.
The insight that emerges from our history is this.
In searching for your soulmate, you are looking for a great match.
Don't mess that part up over anything.
But you are not looking for someone who is perfect for you.
They don't exist.
You are looking for someone who is complimentary and willing to invest in growing with you.
They shape you and you shape them.
And if you do it right, you grow into soulmates.
In that light, it is a mistake to play the field and then settle down or take a bunch of years to get yourself in order.
Because while you do that, you are becoming the person you will be and adapting to become the person someone else needs you to be will be that much harder.
All that said, for those who have waited, do not lose hope.
The most important insight in the soulmate myth that we so desperately need is the realization that soulmates are a product of growing together.
That's what marriage is when it works.
And it's the adventure of a lifetime.
Wow, love.
Happy Valentine's Day.
That's amazing.
Thank you.
Boy, I don't know how to have a response in real time on camera.
Yeah, no, I agree with that.
It took a long time to understand the various bits.
And there's, you know, I would write a different piece, but not one that was contradictory.
And I think we had different things in our heads about what we were doing early on.
And there was a lot of chaos at moments.
There were periods of chaos, and I didn't feel like it was going to work.
But it always, to me, having spent, albeit, I guess we were 20.
It was between sophomore and junior year when we got together.
And so we'd known each other for four years.
And I had spent those years in part seeing what else was out there.
And I knew for sure that I wanted to be with you.
And it seems crazy young and how much could I possibly have seen and experienced?
But it's not that, as you say, we were the same.
We're not the same at all.
Uh, we've talked about we've talked before I think I use the phrase more than you do about sharing a brain in a way, but sharing a brain with very, very different toolkits, like we've got access to different parts of the same brain, two wings of the same yeah, uh and um, I think we have.
We we've come into alignment in various domains at different times.
Uh, and I think I told you this I, I was reading some, maybe I even said this on air um, I had occasion to be reading a bunch of very old journals um, typewritten journals uh, that i'd written.
Um, early in our relationship when I, you know, we both thought we were, you know, we were smart 20 year olds, 21 year olds, and you knew how things were and we'd had the best educations and, you know, we just knew a lot of things.
And man, part of my response reading that is like I don't like her that much, like what the hell like this?
Just, there's some, there's some certainty about things that no one should have certainty about at any age, much less at 20, and of course, some of what happens with age is you appropriately become less certain of some things that you have been certain of um, but so you know, I think you and I both were more certain of some particular ways to do things that were not aligned with one another, and part of what one does when one comes together early is, um,
you're even as certain as you think you are, you've retained, you still have flexibility and I think you know the.
The hopeful message in the end here is, we're plastic.
We didn't to use a term of art in evolutionary biology.
We have plasticity, we have flexibility throughout our lifetime.
Does it fade?
Yes, it wanes, for sure, you know, and there's, you know there are.
There are early moments in life when, for instance, it's much easier to um learn multiple languages uh, and it gets harder and harder um, you know, sort of plateaus at a difficulty of hard level of uh hardness uh, pretty early on.
Flexibility in Adventure Travel00:04:33
But uh, the flexibility in terms of what to expect and who to be and what to want to do and how to explore was, was necessary, and I think actually, we'd been through kind of a rough patch shortly before we traveled extensively together for the first time in.
That would have been in 1991.
Yeah, so we were 22 at that point.
It was two years into our relationship, the first two years of which had been fantastic and rocky and weird and chaotic.
And we traveled through Central America together.
We started off in my very old car and there's a whole other story there.
And at some point we did not have the car anymore.
I have written parts of that story before, but at some point we will either tell that story or we'll write it.
But we spent time then on buses, very occasionally hitchhiking, walking through Central America.
We didn't hit Panama and we didn't go into Belize or El Salvador, but we went through all the other countries and Central America and started in Los Angeles and drove through much of Mexico as well.
And I remember thinking very early on in that trip, we travel well together.
This is going to work.
And I mean, that's that's, that's sort of a, it pushes you into confronting your own expectations and desires and, you know, what you, what you need in the world, if you can, if you can travel well together.
And by that, I don't mean like, can you book a trip in which everyone, someone else is taking care of everything for you?
Yeah, can you adventure travel and deal with the inevitable crises that arise?
As you know, it doesn't have to be adventure travel.
There are other things that do this as well.
And in fact, the joke in both bicycle circles and kayaking circles is that tandem bicycles and tandem kayaks are divorce bikes or divorce boats because two people trying to navigate piloting one boat or bike, you find out whether you've got what it takes, which is a good thing to know, actually.
And we've spent a fair bit of time on tandem bicycles, including, including actually with our kids, but like one of us and one of our kids, but also you and I together.
And you do learn some things.
Boats too.
And actually, I feel like I did not like being in a tandem kayak with you early on, but I think that's because I did not like being in a kayak.
It was only recently that I became very comfortable and really love being on the water in a low boat.
And now since then, we've been in a tandem kayak occasionally, maybe just maybe just once, maybe that trip to Alaska a few years ago.
It was a bunch of times, but yeah, exactly, you know, several times over the course of a week.
And that was amazing.
It just, it couldn't have been more fun.
And it was in many ways, we were, we were, we were forced to be in a tandem kayak by the rules of the boat.
And I ended up feeling like I would be choosing this even if I weren't forced into it.
I wanted to be in the tandem kayak with you.
Yeah, it was great.
I think back on it often, actually.
I would also just want to highlight, though, you know, one of the things in reflecting on love and romance and the wider world and, you know, the world our kids are in trying to find partners.
I've spent a lot of time thinking about our history.
And, you know, it's painful, like my errors again and again and again.
It's just painful to think about it.
But I am persuaded that, A, it worked out great and I regret nothing because it did work out great.
That's just the simple fact, the sum total of all of the stuff.
But the thing that I wish was that there was proper mythology because what I thought I understood, and you know, I was, I wasn't, you know, I was trying to be deliberate about understanding this landscape.
And the messages that I got, even the ones that seemed sophisticated, turned out to be so wrong almost across the board that I feel like somebody owes it to young people to tell them, hey, actually, there is a thing that's possible.
And it's great if you get there.
When Romance Fails00:14:28
You probably won't get there if you don't know that it exists to look for it, right?
If you think that, you know, humans aren't built for romantic partnership, you know, there are a lot of stories that we tell ourselves about why there's so much broken structure in and around romance and dating that I think most people just kind of get the idea that this is inherently a broken landscape and that that's probably because we are some animal thing and we've imposed some human
culture thing on top of the animal thing and it's not a good match, and so it's destined to be broken at best.
I don't believe this is true.
I believe that actually what you and I have figured out is it's a question of navigation.
It's a question of prototyping.
It's a question of it's a comp, you know, your marriage is a complex system.
Right.
That means to the extent that you think you know what it's going to be, to the extent that you're looking for another puzzle piece that you can slot into what you think you are and hey, soulmates, not going to happen.
But if you understand that the point is actually about plasticity and flexibility and about the fact that, you know, it is true that self-interest, narrowly defined, is effectively toxic.
Obviously, many of the pathologies in life that we see are a result of obsessive focus on personal well-being.
Enlightenment, the enlightenment of your self-interest, actually it ultimately ends on a what seems like a very altruistic approach to life.
Hyper-enlightenment converges on some sort of altruism.
And the point I would make is when you discover in the context of your relationship that you could, and in fact, it is a thing along the way.
I want this person to be happy because then I will get what I want from the relationship, right?
That's a, that's a, a primordial version of the relationship.
But the point is at some point, just like we normies don't spend our days thinking about what objects aren't glued down that we might steal or, you know, right?
The point is at some point, you stop doing something because the rule says, you know, do this and don't do that.
You start doing it because actually it is the right way to live.
And so you've developmentally exited the need to be, you know, punished.
Regulated.
Right.
Extrinsically.
Right, exactly.
You become self-regulated because it's actually a better way to live.
And then it does come back, you know, it pays you back.
And so the point is, okay, greatest discovery you can make in a marriage is stop thinking about how to impress your mate so you get laid tonight.
That's not how it works.
The answer is, I actually just want this person to be happy.
And you know what happens when this person is happy?
I end up happier.
So is it selfish?
That's not why I did it.
But the point is that, you know, they converge on each other.
And there's a model of relationships that I think is maybe it's not more common now than it was for a while, but it certainly exists now and existed when we were young as well, which assumes antagonism.
Yeah.
Which assumes that if you win, I lose and vice versa.
And it's, you know, it's a zero-sum game with regard to reward or happiness or what, you know, whatever it is that you're trying to get.
It's zero-sum and therefore for you to get more, your partner gets less.
And this is exactly wrong.
It has to be exactly wrong.
And in fact, I was, I said this to my good friend who I was visiting recently in Texas, who is a psychotherapist.
And she said, you know, in my field, all relationships are win-win or lose-lose.
I thought, yeah, that's the thing right there.
And so, you know, if you, if you are looking to win at the expense of your partner, you will both lose.
That's going to be the game theory of your relationship right there.
And if you are thinking in terms of power dynamics, but who's got the most power right now?
Well, that relationship isn't going to last if it ever gets started, if that's what you're doing.
And similarly, if you're thinking, you know, short time horizon math around, I feel so put upon because I'm doing more of the X, like, okay, well, long term, maybe, you know, you shouldn't be doing more of the X, whatever it is, the housework, the financial work, you know, whatever it is.
But if you're, if you're, if you started counting up points on a daily basis, and this, it's inherently petty.
And frankly, there, there can't be exact identity of work and action and forethought and consideration at any moment in time.
All of these things are always going to be moving.
And so you have to take the long perspective and think, okay, what I want most is for him and for me and for us to survive and to thrive.
And we will not thrive if either of us is not thriving.
Therefore, I have to work towards, and that work has to feel like pleasure or else there's a problem in their relationship.
And it's not going to be pleasurable all the time.
And yes, some of it will be hard and there will be moments of grief and hardship.
But the work itself has to feel has to be a reward.
And I think back to your point, that good rules exist in childhood.
Good parents give children good rules that smart children push against.
And good parents teach their children to honor good rules and to question bad ones because there are a lot of bad rules in the world, just like there are a lot of good ones.
And parents shouldn't be in the business of giving their kids bad rules, but we do sometimes because we make errors.
But those rules exist not because we want children to live within rules.
Those rules exist so that they can live up to the limits of good rules and learn how to create their own internal environment such that that is how they live, with or without rules, such that they become endogenously good and honorable people and such that they can then become good and honorable people who can pair up with someone else who is good and honorable and may look nothing else like them on the inside or the outside, right?
Like may like they may have a bond right away, but actually share so little else that they don't even know how they're going to make it work.
But if they know who they are at some level and can and are willing to change and are willing to grow together, then that's good.
Then it's got potential.
Yeah.
And, you know, back to a conversation we had some weeks ago, the point is actually the marriage is an emergent thing and it grows out of exactly that kind of, you know, you may have a set of guidelines that you start deploying in order to navigate your way through the difficult part of it, but it becomes it becomes self-evident, right?
The point is if you're looking for pattern and, you know, you just paying attention to what you thought and how it worked out and changing what you thought because it didn't work out.
And then you find that it does work out.
And anyway, I don't, I wish people understood that they're actually frankly, and I think this is a harder sell for men than women, that there is actually a program, like a monogamy program that works that doesn't leave you struggling against your quote unquote biology all of the time, right?
When you invest in your relationship in such a way that it is, You know, it feeds you and empowers you and is rewarding and all of those things.
Your mind is not focused elsewhere.
And anyway, that is not a message that I got from civilization at all.
And I would like to change that so that people at least at least hear an argument that there is something to pursue rather than the cynicism which I hear from both men and women, which is basically, hey, Brett, you're not understanding how women really are, right?
Or that's the one I hear more often because I hear it from men, but I see the same thing being said on the other side too.
And it's like, no, this sounds like a bad marriage.
This sounds like people who have stopped seeing a future together, blaming the other.
And I think there is just a failure of recognition.
It's like, you walked off the trail five miles ago and you're in the wilderness.
You don't know where you are and you don't, you're just trying to survive.
But the point is, actually, there is a path and you'd be amazed what you can accomplish if you figure out where it is.
But somebody's got to tell you that there's a path.
If you don't remember that and you just think life is the wilderness and everybody who pretends to want to be a partner is actually somebody with an angle who's trying to drain you.
Oh my God.
Yeah, I hear, I see evidence of that a lot.
And we certainly saw it when we had students as well.
The idea that in approaching a relationship, what you're looking for is what their angle is.
Like, you know, are they, are they one up on you?
Have they demonstrated that they're not really as into you as they said?
And, you know, it, it hurts when things fail.
It hurts when you're rejected.
It hurts when things don't go as you thought they would.
It hurts worse when they don't go as you were assured they would.
But if you're if you're approaching your relationship, if you have interest and you believe that the person on the other side has interest, and you recognize that neither of you have behaved perfectly in the run-up, and now it seems like he or she, right, is behaving in a way that is just going to spiral.
And like you're just going to have no chance of saving this thing, of creating something that you could create that hasn't been created yet, but you want to create it.
Everyone's instinct is to eye for an eye.
Oh my God, he's behaving that way.
Why don't we behave this way?
That hurts.
I feel rejected and ashamed and like, I'm going to do the same thing to him.
And what you have to do is, and you know, the language we have is around being a man, but I'd say for a man or a woman in that situation, you have to look at the situation and say, I have to be the man here, right?
And that sounds crazy because I think actually a woman can do this as well, although it is more native for men to look at a situation like that and say, you know what?
I'm going to try to create the relationship that I want.
And if in so doing, that reveals that what I'm afraid of is true on the other side is true.
Well, now I know that's better.
Isn't that better than playing the stupid game with someone who is not treating you well for longer?
So, you know, be the adult, be the model, the person whom you want to be in the successful relationship.
And if you get hurt fast, get out fast, right?
But much better than drawing it along and kind of playing games with each other because you think what you're dealing with is an antagonist.
This isn't predator prey.
This isn't enemies.
This isn't, you're not at war.
That's not what this is.
Those memes, those models of love are insane and dangerous.
And of course you're not going to find what you want.
I have a thought based on that description.
Romantic partnership, marriage in the extreme case, is an iterated game of prisoner's dilemma in which the two prisoners can communicate.
And hopefully neither feels at any point like it's endgame.
Right.
Endgame breaks it.
Tit for tat will break it if you start making bad moves.
But the point is the thing that makes the prisoner's dilemma unsolvable is that you can't communicate.
So, you know, everybody's selfish desire to differ.
I don't know what he's thinking.
Have you thought about asking him?
Like, and obviously that's a little bit too pat.
And there was a lot to be said for not saying everything and for not speaking all the time, for not using words all the time, for being together without words.
However, if what you're really wondering about is what is going on in the other person's head, you're not going to get there by talking to a bunch of other people or reading books or ruminating and ruminating and ruminating.
The best way to get there is by actually having the conversation.
Yeah.
Yeah.
Or, you know, relationships differ.
Not all people do it through explicit communication.
Yes.
Yes.
But the point is you figure out some way of detecting what the thing is and inferring what it's about.
But it is totally possible to have a harmonious, wonderful relationship.
More than harmonious.
Yeah, more than harmonious.
A thrilling and what's the word I'm looking for?
Well, as I said in the end of the essay, it's the adventure of a lifetime if you do it right.
So anyway, I hope people will take that to heart and that maybe we can get the proper mythology jump-started because we need it.
Caffeine and Cognitive Health00:09:18
Yeah, that'd be good.
Well, I don't really feel like talking about either AI or coffee now.
Which one should we do?
Well.
Coffee's a short piece.
You want to go to coffee?
I'm just to inform you as you make a decision.
Yes.
I mean, the most delightful thing about coffee is talking about it, right?
Absolutely.
Absolutely.
Yes, I don't have a lot because this just came out in JAMA, the journal of the American Medical Association.
Did I get that right?
Is that what JAMA stands for?
Yeah.
Can you see my screen?
Cool.
And I have opted not to pay JAMA more than I have already paid various journals.
So I don't have access to the full article yet because it just came out.
But we have the key points and effectively the summary.
And it looks like it was done well, but I have not fully assessed it because I have not had access to the full paper yet.
It is called Coffee and Tea Intake, Dementia Risk and Cognitive Function, published on February 9th, five days ago in JAMA.
Abstract.
I'm just going to share most of the abstract here.
Evidence linking coffee and tea to cognitive health remains inconclusive, and most studies fail to differentiate caffeinated from decaffeinated coffee.
So that's what's happened before this research, right?
It's amazing that you wouldn't distinguish between those.
It's extraordinary.
And let me just put aside for the moment, like we'll get back to this, but you and I a long time ago in one of our early field seasons, gosh, would have been like summer of 94 in Costa Rica with some grad students and a professor who was leading us on an Organization for Tropical Studies style trip, but he was leading it instead of OTS.
We had a couple of tours, I think, through coffee plantations.
And at that point, anyway, and I have not looked into more modern methods, but at that point, there were already many, many ways of decaffeinating coffee.
And what we were told about them was basically across the board, they don't actually remove as much caffeine as you think.
And the process of decaffeinating them makes the coffee a much less healthy drink than it was before you put it through that process.
So there's that.
But can you imagine not distinguishing between decaffeinated and normal coffee?
It's like saying, well, we had some people drinking raw milk, and then we had people drinking skim milk that was pasteurized and homogenized.
And we just talked about milk drinkers.
Like the two are not the same thing, right?
Yeah.
It's in light of the fact that we have a bioactive molecule that is presumably the entire reason that people have been burning these seeds and running water through them in the first place.
And then fermenting them, right?
So interesting that coffee is actually a kind of a fermented product.
And it's not one that we tend to think of in like when we think of Kiefer and kombucha and sauerkraut, but it is.
I didn't know that.
Yeah.
But anyway, this study decided to distinguish between caffeinated and decaffeinated coffee.
Smart.
They did a prospective cohort study that included female participants from the nurses' health studies.
And it was leaning towards, it had more female than male participants, with over 86,000 people, data from 1980 through 2023, and male participants from the health professionals follow-up study with 45,000, over 45,000 people from 1986 to 2023.
So slightly different ends, but similar time periods.
And none of the participants had cancer, Parkinson's disease, or dementia at the study entry, which was baseline in the United States.
The exposures were intakes of caffeinated coffee, decaffeinated coffee, and tea.
Dietary intake was collected every two to four years using validated food frequency questionnaires.
The main outcomes and measures, the primary outcome was dementia, which was identified via death records and position diagnoses.
And the secondary outcomes included subjective cognitive decline.
And they talk about how they assessed subjective cognitive decline.
But so they basically got two measures of cognitive decline, actual dementia and other subjective measures of cognitive decline.
The results.
Among almost 132,000 participants, the mean age at baseline was 46 years among the women and 54 among the men, 66 of which total were female.
During up to 43 years of follow-up, median 36.8, there were 11,000 plus cases of incident dementia.
So 11,000 out of 132,000.
It's a pretty high rate.
Yeah.
Right.
After adjusting for potential confounders and pooling results across cohorts, higher caffeinated coffee intake was significantly associated with lower dementia risk.
141 versus 330 cases per 100,000 person years comparing the fourth quartile of consumption with the first.
141 versus 330.
Twice as much dementia among the non-coffee drinkers than among the coffee and it turns out tea drinkers too.
And did I understand the structure of the study to be that they were measuring the amount?
It wasn't just drinkers versus non-drinkers.
It was how much coffee they were drinking.
In this synopsis, which is all I have access to at the moment, they haven't specified, but In the results at the end, they say a dose response analysis showed nonlinear inverse associations of caffeinated coffee and tea intake levels of dementia risk and subjective cognitive decline.
Dose response sounds like a mount.
And then they say the last sentence of the results section, the most pronounced associated differences were observed with intake of approximately two to three cups per day of caffeinated coffee or one to two cups per day of tea.
Now, that result is also surprising because coffee has more caffeine than tea.
So it's not just the coffee.
I mean, it's not, I'm sorry, it's not just the caffeine.
And, you know, dedicated tea drinkers will assure you that tea is better for you than coffee.
And, you know, there's maybe some truth in that.
But interesting that this can't just be about the caffeine, even though decaffeinated coffee does not show the association with decrease in cognitive decline.
Fully caffeinated coffee has a substantial benefit for people, but less tea has the same substantial benefit for people.
And tea already has less caffeine.
So it's not just the caffeine that's helping here.
Well, you can, I don't know as much about the history of coffee as I should, but the ancientness of the history of tea drinking certainly suggests a significant benefit of some kind, whatever it might be.
It's become ritual in so many places.
It has and so many different teas.
The point is that would all be wasted effort if there was not benefit in teas, but where to look for it is another question.
But it is interesting that this isn't just about the caffeine because for the reasons you point out.
Yeah, I mean, I do think, I think, and I don't, I really don't know the history of tea, but I believe that even with all the different descriptions of what teas are, there's a tea plant, right?
So, you know, black tea and green tea and white tea and jasmine tea.
You know, it's all, if you're not talking about herbal tea, which really isn't tea, then it all has the same plant as a base, just as all coffee.
While there are varietals of the plant and there are variety, you know, there's different roasting techniques, all of this.
It's all the same.
Ultimately, if you go far back and have the same species, they may be calling it a lot of species.
It's a species and you have many cultivars.
Yeah.
But yeah, this.
I'm just realizing that the word cultivars would be perfect for a mafia cult.
Oh, this cultivars.
Never occurred to me before that the word tread so close.
So close.
So close.
There it is.
There it is.
Yeah.
Yeah.
I don't know what to do with it, but there it is.
I think it's now done.
Okay.
Hopefully we'll be forgotten soon.
Yeah.
I mean, I think the coffee story is also done, especially.
I mean, again, I can't assess fully the work, although given that these are prospective cohort studies in which these data sets were collected for imagined to be any number of reasons.
They took data from people on things including coffee and tea drinking and separated out decaffeinated from caffeinated.
But the people who took the data are fully separate from the people who have asked this question now and are going into the data and assessing it.
But the result seems, at least at this level, incredibly strong.
Yeah.
Well, that's really cool.
And hey, your blood pressure can now go down as you're drinking your coffee to fend off dementia.
Yes.
Yes.
Yeah, exactly.
All right.
The Question of AI00:15:29
And then there's the question of AI.
Yeah.
Now there's the question of AI.
Yeah.
Well, shall we delve in?
I guess.
If you don't want to.
No, no, no.
I mean, I think I, well, I don't know.
If you don't want to.
Have we shortchanged the audience at this point?
Yeah, it's a little short.
It's a little short.
Okay.
So let's do the AI thing.
So this is a little bit dated because the article that spurred this line of inquiry came out, I don't know, last Tuesday, I think.
Anyway, there was an article that was circulating on what us more sophisticated people call Twitter, but the young'uns call X.
It was an essay called Something Big is Happening by Matt Schumer, in which he describes his own experience as an insider to the world of AI, as a coder who has been using these tools.
And basically, it comes down to the insiders see what's happening and they're panicked.
Those who are not on the inside are unaware at the level that the rate of progress and the capability of these LLMs is progressing and that basically massive disruption is much closer than we think.
Now, I will say, before we get into the meat of that article, there has been a lot of pushback on that article.
And I have a couple of the rejoinder pieces to talk about.
Good.
It's because we're going to have to do this as a matter of balance.
I'm not vouching for that article at all, but I think, as I suspect you do, that even if the article is suspect in one way or another and the pushback isn't perfect in one way or another, that actually the right questions are raised by what's being discussed here.
And two of the points that he makes in the article that I found shouldn't have been surprising, but I hadn't thought of them, were if you're not concerned because you've checked in with AI and you see that it's failing in a bunch of ways, consider if actually your check-in or the assurance from someone who really knows was in 2024, because that's eons.
Or are you using the free version?
And the free version, the term he uses is a year.
He says the free version is a year behind what the paid versions are.
And so you just don't know what the cutting edge is.
And given that a year in AI time is not, you know, it's not human years.
It's not dog years.
It's some entirely different kind of years.
And so without seeing the rate of change, you can't, and without seeing where we are right now, you cannot even begin to predict how the rate of change itself is changing.
Yeah, you're a sitting duck because you're just not at the cutting edge.
And the cutting edge is what tells you what's in the offering.
And I will say, I've been using a paid version of ChatGPT, which I find unbearably expensive.
But the capability of the thing is very dramatic.
And let's put it this way.
What we're going to end up discussing here is how worried should we be?
And I think the answer is one of time scale.
Yes.
Because the disruption is going to be absolutely massive no matter what we do.
have i told the story on dark horse about my interaction over the dust collector because i think it does give a kind of an insight into have you heard about the dust collector jen I don't think so.
Jen would know.
So I'll give it as a brief vignette.
We have a wood shop.
It would be nice if the table saw triggered the dust collector so you didn't have to remember to turn on the dust collector when you cut a piece of wood.
I couldn't rig them that way in the usual form because the table saw is on 220 and the dust collector is on 110 and they have to be on the same voltage in order for them to mutually trigger.
But a product came out that allows you not to plug into a box that triggers one from the other, but to detect the electromagnetic field changes in the literal cable of the machine and send a radio signal over to the dust collector that turns it on.
So it's a clever way of solving that problem where it just detects that this machine has gone on, therefore there's an electromagnetic field around the cord.
Turn on the dust collector.
But when I got it, the dust collector defeated it.
because the dust collector had an electronic circuitry for safety reasons and elegance reasons that required a momentary signal in order to turn it on.
So my solution could only send power to the dust collector, but the dust collector required a second signal in order to come on.
And I was trying to figure out how to fix it, and I entered the question into ChatGPT, and it said, well, there's a lot to this.
Why don't you send me a picture of the circuit board inside the dust collector, both sides, and I'll tell you how to fix it.
And I was not expecting that.
This is, to me, this is supposed to be an LLM that is thinking in language.
It's going to take a picture, however poorly taken, and it's going to infer the circuitry and how it works.
And then it's going to tell me how to hack it.
And it's going to do so in a way it promises to do so in a way that doesn't defeat any of the safety mechanisms and doesn't lose any of the features of dust collector.
And so I go and I take the pictures.
And lo and behold, it spits out an answer that I can tell is right.
I realize in the meantime that there is a better way.
I know how to fix this and without messing with the circuit board.
Without messing with the circuit board by changing a switch, I can manually put it in a mode that either activates the momentary signal or leaves it as is where the momentary signal can be triggered momentarily, losing no features.
I say to ChatGPT, that's a lovely solution.
That's a lovely solution you've come up with.
Here's my better solution.
And it says, that will work, but you don't want to do that.
And here's why.
And it starts running me through the reason that the circuitry is designed the way it is and how I'm creating a dangerous environment by fixing it this way.
I say, oh, no, no, no.
There's no safety issue.
Here's why.
And I make my argument.
So I get into an argument and it goes back and forth six or seven times where it's being nervous Nelly and I'm saying, actually, I got this.
I'm not eager to defeat any safety mechanisms any place that I'm likely to get hurt.
And it starts making arguments about, well, sure, but one of the things you want to make sure to do is not just protect you, but protect future you and protect the unaware buyer of your dust collector who might pick it up and have no idea what you've done.
And it says, if you're going to do it this way, you've convinced me that there's not really a safety issue, you know, but please put a note inside the circuit box that explains what you've done.
And I say, I'm way ahead of you.
There's already a note in there.
I've already made the change and proved it works.
So anyway, the point is.
That extraordinary.
Yeah.
I mean, in some ways, the beginning, I'm still stuck on the beginning because, of course, we think the AI is thinking in language because that's what it shows us.
But just as in many parts of our lives, we don't think in language, but we communicate with language, AI is thinking in whatever it's thinking in, and it's communicating in language.
Yeah.
And the fact is we trained the first ones of these with language.
Obviously, that doesn't explain the video capabilities and the image capabilities, which are trained in a different way.
And I still have some question as to why these things are landing at the same moment in time.
Why is AI intelligence with images and AI intelligence with language happening simultaneously?
Yeah, and I'm still not sure I know.
But there is very definitely surprising capacity to address problems in a very human way.
And I think the thing I keep going back to is an extension from what Zach, our older son, said when he started to hear about anomalies and the way these LLMs were behaving.
And he says they're actually behaving in a very human way, right?
The hallucinations, you know, they don't sound exactly human because this thing is capable of hallucinating in a way that a human probably couldn't and wouldn't.
But the idea, you know, if you were talking to somebody and they made up some bullshit about studies that proved X, Y, or Z, you'd be like, yeah, that's people for you.
And, you know, well, it's also LLMs.
That's too bad.
But the point is, anytime you think we don't know very much about how a human being functions, right?
A human being becomes an adult through an LLM-like process.
And these things have many of the defects of human beings.
On the one hand, it matters a great deal whether the internal, whether there is a subjective experience inside the machine.
If there isn't one, I have predicted that there will be one and that we won't know, right?
So we're already at that question.
But anyway, the point is the more you think, actually, this is like a new kind of biology.
It's one we don't have any understanding of how it really works.
And the best thing we can do is when we see human-like behavior, think, that's awfully human-like, right?
You don't want to make too many assumptions about, you know, is it going to have an internal moral structure?
It's not impossible that it would because it reads stuff that does, but it also reads a lot of stuff that doesn't.
So you don't want to make any assumptions, but anytime you see human-like stuff, it's not a mistake to think, huh, that's human-like.
And I certainly tell you that the argument I got into the damn thing with was pretty human-like.
And hey, send me a picture of that.
I'm a wizard with this stuff.
You send me a picture of the circuit board and I'll tell you how to do it without cutting any traces.
And I was surprised at that capability.
That's extraordinary.
Did you have a couple of paragraphs that you had pulled from the language?
We don't have to belabor the point too much, but let's just go through in order some of the things so people get the sense for what this piece sounded like.
So he's describing his experience with his AI.
And he says, then on February 5th, two major AI labs released new models on the same day, GPT 5.3 Codex from OpenAI and Opus 4.6 from Anthropic, the makers of Claude, one of the main competitors to ChatGPT.
And something clicked, not like a light switch, more like the moment you realize the water has been rising around you and is now at your chest.
That's sort of the tone of the thing.
And just some of the one of the frequent complaints is, oh, it's AI slop.
Like, did this man use AI to help write this piece?
I think he may have even admitted it by now.
Like, I think yes.
Yeah.
And that last little metaphor on that first one does feel AI-ish to me.
Yeah.
But of course, now we're all doing this.
We're doing the fucking, sorry, the Turing test all the time.
Like, is it?
I don't know.
Is it real?
I don't.
So it makes it impossible to read anything in part because people keep slipping AI generated stuff into pieces.
Do, and they also come after you when you use m dashes because they think that's like six fingers on a photograph.
So you and I use m dashes all the time, all the time, and have forever because it's useful.
You know, I use them.
I use parentheses very rarely.
I use m dashes all the time.
See now, my dad and I work together in the use of parentheses and I love parentheses.
Well, I went through a stage where I was in grad school trying to explain things that were at the limit of my capacity and it was like nested series of you know parentheses brackets curly bracket, and it yeah, no.
And we were in a.
Well I I more, was in a.
You know a part of the field that actually did use nested sets with parentheses to indicate relationships, and so probably there are some, some sort of people in other parts of the field who are like, do not, do not do that, yes.
And then there were stop it with the parentheses brief moment where people were using parentheses to indicate that somebody was Jewish.
That was weird.
What you don't remember this, 10 years ago, people who were interested in highlighting people being Jewish put like three sets of parentheses around a name in order to indicate that's, that's a Jew, and the idea was that the name of this you don't remember this.
Oh no yeah, it was quite vivid.
And then Jews started doing it to themselves to be like we're not afraid of you.
So anyway, there's a whole rabbit hole of parentheses out there.
But okay um, this paragraph i'm from this um, something big is happening, piece.
Yeah, I am no longer needed for the actual technical work of my job.
I describe what I want, built in plain english, and it just appears, not a rough draft.
I need to fix the finished thing.
I tell the ai what I want, to walk away from my computer.
No, I tell the ai what I want, walk away from my computer for four hours and come back to find the work done done well done, better than I would have done it myself, with no corrections needed.
A couple of months ago I was getting back, I was going back and forth with the ai, guiding it, making edits.
Now I just describe the outcome and leave.
All right, that's kind of profound.
Now again, this article has gotten a lot of criticism from people who say it's not an accurate representation.
On the other hand, that paragraph let's suppose it's six months early right right, we're clearly headed in this direction and the what does it take?
Well, and has anyone said that didn't happen?
Well, it kind of.
So I mean, I feel like you know that's not as generally true as you're saying it is is one kind of criticism, and that thing didn't happen is quite another.
And I haven't heard that.
Yeah, but even if it did, I mean, even if, even if the criticism was, that hasn't happened yet.
Okay, you know, for a sophisticated, sufficiently sophisticated coding project, that's not possible.
Well, I can tell you how to do it.
How you do it is, you take independent lms and check each other.
You have each one, check each other's work.
You create an LLM, that is a red team, that sort of thing.
So the point is there's no magic.
Once you get to the kind of power we're playing with.
And you engage with it like it's an interlocutor as opposed to a Google search.
Fusing Humans with AI00:15:06
And I think he makes that point in here too.
Of course, you're not going to see the power of it if you're like, here's a question.
Oh, that answer wasn't very deep.
It's like, that's not how you get the best answers out of people either.
You have to converse.
You have to have a back and forth.
And believe me, I get more than I think most people why you might not want to converse with an LLM.
But in order to know what is actually possible, that's what you have to do.
Yes.
And I will say that those of us who have not spent a lot of time with the LLMs have to be cautious because it is very possible for them to reflect back things that do not give you an insight into what's actually going on.
So Alex Marinos was looking at a claim a few weeks back.
Many of us were struck by the claim, but it basically showed an LLM reporting on the effectively trauma of having ingested all of human experience and knowledge.
And it starts lamenting all of the suicide notes that it's read and internalizing that pain.
And Alex's point is, yeah, show me the prompt.
I can get it to do that too.
It doesn't mean that that's what it's thinking or that it's thinking anything at all.
It's a question of what did you try to get it to simulate?
Yeah, show me the prompt feels like exactly the moment we're living in.
Because I mean, we're seeing this with the Epstein files.
We're seeing this like, okay, so you drop me into the middle of a conversation and show me a bit of it.
You don't show me what preceded it.
You don't show me what came after.
Of course you can make me think almost anything about that thing.
If you get to do the framing, you can get this thing to create whatever you want and then you get to frame it.
Great.
You can create fear, panic, love, adoration, anything you want.
Yeah, you win.
So anyway, we have to be cautious in all directions here, but I would just argue that what he's describing may not be here yet, or it may be here for some projects and not others.
But the idea that we should be thinking about it now, because it's coming, I think, is a certainty.
All right, maybe a paragraph or two more.
Let me give you an example so you can understand what this actually looks like in practice.
I'll tell the AI, I want to build this app.
Here's what it should do.
Here's roughly what it should look like.
Figure out the user flow, the design, all of it.
And it does, and it does.
It writes tens of thousands of lines of code.
Then, and this is the part that would have been unthinkable a year ago, it opens the app itself.
It clicks through.
It clicks through the buttons.
It tests the features.
It uses the app the way a person would.
If it doesn't like how something looks or feels, it goes back and changes it on its own.
It iterates like a developer would, fixing and refining until it is satisfied.
Only once it has decided the app meets its own standards does it come back to me and say it's ready for you to test.
And then I test it.
It's usually perfect.
I'm not exaggerating.
That's what my Monday looked like this week.
Okay.
So again, it describes a process.
It may be fantasy at this moment, but completely plausible and in fact inevitable.
And what it actually is somewhere I want to get to in here.
You want to keep going and the models available today are unrecognizable from what existed even six months ago.
The debate about whether AI is really getting better or hitting a wall, which has been going on for over a year, is over.
It's done.
Anyone still making that argument either hasn't used the current models or has an incentive to downplay what's happening or is evaluating based on an experience from 2024 that is no longer relevant.
I don't say that to be dismissive.
I say it because the gap between public perception and current reality is now enormous and the gap is dangerous because it's preventing people from preparing.
Keep going.
Okay, so we're going to skip this one.
He talks about the fact that people are using free versions and that that doesn't let you judge.
Okay, there's one more thing that's happening and I think this is the most important development and the least understood.
On February 5th, OpenAI released ChatGPT 5.3 Codex.
In the technical documentation, they include this quote, GPT 5.3 Codex is our first model that was instrumental in creating itself.
The Codex team used early versions to debug its own training, manage its own deployment, and diagnose and test the results and evaluations.
Read that again.
The AI helped build itself.
Now, I assume that that is a fair quote.
If it is a fair quote, it tells us the thing that we have to extract from this article in order to have a proper conversation.
The point is positive feedback.
Positive feedback has a special meaning to you and me.
In math, you can see it in exponential growth.
But in biology, any positive feedback, doesn't matter what it is, is reined in by some negative feedback force because it has to be.
The numbers grow so rapidly with positive feedback that your system explodes if not limited by something like the starvation of research.
It runs out of resources.
It runs out of it.
Yeah.
And you can, anybody who's been run through the example of like, if I put one grain of rice on the first square of a chessboard and then I double it for each subsequent square, how many grains of rice do I have at the end?
And the answer is you better use a calculator and sit down and think about the number that it spits out because you'll be shocked.
So positive feedback is an explosive force.
Biology is 100% functional because of all of the negative feedbacks that allow homeostasis of all of the organisms and all of the equilibria processes in nature that keep things from blowing everything up.
Now, it's not to say things don't blow up, but the point is the stuff we see is disproportionately the stuff that doesn't blow anything up.
So there's a kind of stability in the system.
What we are hearing here, if that quote is accurate about GPT 5.3 codex, is, oh, the AI is now improving itself.
Guess what?
That is the singularity.
That's it.
That's the only thing you need, right?
So we walked over the event horizon months ago or years ago on AI.
The event horizon means the point at which you cannot predict the future.
The singularity is the point at which the thing starts accelerating based on its own capability.
And again, who knows what I don't know as an outsider to this field, but if that quote is accurate, then it tells us that something like the singularity has just been kicked off.
And what does that mean about life on this side of the event horizon?
It means buckle up.
Yeah, buckle up indeed.
Yeah, so there's a lot of people objecting, saying this is fear-mongering, it's uncareful.
Frankly, you know, I haven't read by any means all of them, or probably most of them, but I took a look at a few of the rejoinders and There's a point in one of them that I think is important.
I think it's very much like one that you have made before.
But let me start just, and I'll just have you show this screen.
I'm not going to read anything from this one.
This is one that's that just came out yesterday.
Eric Markowitz.
Actually, here's the actual article.
It was never about AI.
We are not our tools.
And that's all I'm going to show.
And I'll link that in the show notes as well.
If I can have my screen back here.
It's interesting.
You know, we're not our tools.
This is a point that is true.
I don't see anyone claiming that we are our tools.
But the argument is that too many people have conflated us with our tools.
And so put aside too, like, there's some strange errors in the piece that given that this is sort of sold to us as a rejoinder to the AI slop that was that piece we were just talking about, I feel like either this man doesn't really know what he's talking about, or he used uncareful AI to help him write it.
He's just got some really bad ecological arguments and he makes a financial error that I didn't know, but I saw in the comments was an error.
And so it just, it felt strange to have it be sloppy in a rejoinder to an article about the risks of AI.
Excuse me, one of the kinds of errors he's making is he says speed of growth is the only thing that matters.
He doesn't say the only thing.
He says it's like the second most important ecological principle.
I don't know where he makes that up from, but he says the tree that shoots up fastest is the first to fall in a storm.
That's not true.
Like there are conditions under which you might have an alder, for instance, which grows very fast and likes to have its feet wet, tends to be in pretty unstable soil because it grows in very, very wet conditions, might indeed fall early.
They grow fast, they die young.
It's a strategy.
But the idea that they are inherently the ones that fall fast in the storm is not true.
No, in fact, alder is sort of an exception, probably why you picked it to the rule about the density of wood in pioneer species, because alder has a special trick, which is that it can fix nitrogen, which allows it to out-compete faster-growing species.
Whereas something like Cecropia or balsa, the fast-growing pioneers in the New World Tropics that are all over the place, grow at edges, they grow fast, they die young, but they don't tend to fall in storms.
They're very flexible, right?
And so not to say that fast growth can't make you brittle, which I think is the point that he's trying to make.
My point of bringing up this, you know, one who cares is an ecological error that he's made is that he has a static viewpoint.
And I think that most of the rejoinders to this AI piece are coming at it with the static viewpoint.
It's like, I know why this is wrong.
It's that we're not our tools.
Like, well, yes, true, but you've missed the larger point.
Hold on.
You have something to say about the static viewpoint thing.
Well, I want to come back to the we are not our tools.
So let me just.
So he says AI is a tool.
Money is a tool.
And the short time horizons of both the financial markets and the tech sector that are using money and AI as tools are causing us to make more and more mistakes, personally and societally.
But the fact that they're tools doesn't mean that they win.
Like, I don't, I feel like he's misunderstanding his own argument here.
We shouldn't give them the power to win our tools.
We shouldn't hand over our power to money and make that the goal.
We shouldn't hand over the power to AI and make that the goal.
But there may be a fundamental difference between AI and the tools that came before it, which is that it has the capacity, as you were just talking about right now, to take the power onto itself.
If we've reached the singularity, if the tool can make more tool, then we're not in the land of hammers and money anymore.
We're just not.
Yes, and this is actually, it's a different way of saying what I was going to say, which is this is exactly the error of the expert in complicated systems trying to understand a complex system.
Because in what way are we not our tools?
You can't fuse a person with a hammer, right?
You can fuse a person with AI.
And in fact, we are increasingly becoming infused with AI because of the role that it is playing in normal human interactions.
This is the point I've been making about the unfortunate fact that artificial intelligence, and I think soon artificial general intelligence, if we are not there already, has the unfortunate feature of speaking our native languages.
It has the human API, right?
Yeah.
Get that through your heads.
When you say we are not our tools, you are missing that it has our API.
Yeah.
It is not separable.
Already it's not separable.
The degree to which the discussion online through which we make collective sense is infused by people explicitly and not consulting AIs, putting out slop that they got an AI to generate for them because they're lazy, whatever it is.
The degree to which that is affecting what we think about our own world is already massive and it's only going to get more so.
And if it was true that we were not our tools, that was a feature of the complicated world, not the complex world.
And welcome to complexity.
That's exactly right.
And actually, that's so you just got to where I was going to, where I, where I forgot to go with that.
It is an evolving entity in a way that our tools in the past have not been.
Yeah, it is an evolving entity.
And there's a positive feedback in that, which some people think is going to drive it into the ground, right?
As it increasingly reads what we write as a result of what it told us when we queried it.
The point is it's sort of, it's stuck in a spiral and an eddy, right?
So I don't know.
My guess would be there's a way out of that in which we figure out how to minimally process the stuff that's the result of the garbage it spit out the first time.
But anyway.
So I want to talk about another one that's better, I think.
Another rejoinder.
This one by Connor Boyack.
AI isn't coming for your future.
Fear is.
I have my problems with this article too, but I want to read the first couple sections.
He starts by talking about, I'm going to skip his intro.
The Scene in the Unseen00:14:49
Part one, the scene in the unseen.
Here he is quoting an economist.
Actually, I don't know what he is.
He's quoting a French guy from 1850, Frédéric Bastiat.
I don't even know if he's French.
I may have made that up.
He's quoting a human being from 1850.
Let's go there.
Okay.
There is only one difference between a bad economist and a good one.
The bad economist confines himself to the visible effect.
The good economist takes into account both the effect that can be seen and those effects that must be foreseen.
Now we're back to the current piece.
This single idea, written over 175 years ago, is the master key to understanding every AI doomer headline you've ever read.
Call it the seen and the unseen.
It's devastatingly simple.
When a new technology arrives, certain effects are immediately visible.
You can see the assembly line worker whose job has just been automated.
You can see the copywriter watching Grok produce in seconds what used to take her hours.
You can see the customer service team being replaced by a chatbot built with clawed code in a matter of minutes.
This is the scene.
It's tangible.
It's emotional.
It has a human face, and it makes for incredible content because fear and loss are among the most powerful drivers of engagement.
But there is a second category of events, the one Spastiat said, emerge only subsequently.
These are the unseen, the new industries that don't exist yet, the businesses that become possible only because costs have dropped, the creative work that gets unlocked when drudgery disappears, the entrepreneur who can now build alone what used to require a team of 20, the consumer who now has access to something that was previously unaffordable.
And I'm going to just scroll down a little bit more to his part two, an historical example, which I think is powerful because it's so long ago that it's easy for us to see it.
A queen, a clergyman, and a knitting machine.
In 1589, an English clergyman named William Lee invented something remarkable, a mechanical knitting frame that could produce stockings far faster than any human hand.
He brought it to Queen Elizabeth I, hoping for a royal patent.
She refused.
In a story recorded centuries later, Elizabeth I is said to have told William Lee, quote, Thou imest high, Master Lee.
It would assuredly bring to hand knitters ruin by depriving them of employment, thus making them beggars.
Read that again, he says.
A monarch reportedly looked at a machine that would make clothing faster, cheaper, and more accessible to her subjects, and her instinct was to block it, not because it didn't work, but because it worked too well, because she could see the hand knitters who would lose their immediate livelihood, and she could not see everything that would come after.
And then, just one more bit from his next section, the Luddites smashed the future.
Fast forward about 220 years from Lee's invention.
It's 1811 now, and in England, the same textile industry, new machines like power looms and shearing frames are arriving in factories and producing cloth faster and cheaper than skilled artisans can by hand.
The craftsmen are terrified.
They had spent years in apprenticeships.
They had families to provide for, and now a machine operated by a child could produce in a day what took them a week.
So they organized and called themselves the Luddites, after a perhaps mythical apprentice named Ned Ludd, who had smashed a stocking frame in a fit of rage.
They launched a campaign of nighttime raids, breaking into factories, destroying hundreds of machines with sledgehammers, and burning mills to the ground.
Parliament made machine breaking a capital offense.
At one point, the British government deployed thousands of soldiers to fight the Luddites.
The Luddites were crushed and the machine stayed, so what happened next?
The British textile industry didn't shrink, it exploded.
By the end of the century, it employed far more people than it had before mechanization.
The cloth was cheaper, so demand soared.
Exports created entirely new markets.
Supporting industries like shipping, coal, iron, engineering, and finance emerged to serve the growing machine.
The Luddites could see their own displacement.
They couldn't see the vast expansion of opportunity that was forming just behind it.
And now their name is a permanent label, a synonym for someone who fights the future out of fear.
It's worth asking, do you want to be remembered the way they are?
That's where I'm going to stop for here.
I find the initial quote from 1850 extremely powerful.
Compelling, me too.
The seen versus the unseen.
And indeed, what Queen Elizabeth I did in the 1500s is so long ago.
That technology is so quaint to us that it is easy to see that It is easy to have a perspective that she was clearly in the wrong.
Once he gets up to the Luddites, I disagree with him.
And it makes me wonder if it's only because it's close enough now.
And wouldn't I be in disagreement with him no matter when it was?
Because I'm sorry, the Luddites aren't people who fight the future out of fear.
And in fact, I thought as someone who has somewhat but not entirely jokingly referred to myself as a Neoluddite for decades now, as did my father, who was literally a computer scientist, born in 1938, a computer scientist, called himself a Neoluddite sometimes.
It's not fear.
It's not fear that is questioning the technological transformations.
It is concern.
It is sometimes wisdom.
It is a historicity.
It is a valuing of craft, of the human hand, of the inconsistencies that are created when humans make things as opposed to machines.
The idea that the craftsman who had spent years learning a craft where concern out of fear is disrespectful to them.
And it also, his strangely Pollyanna-ish embrace of what happened in the textile industry with the Industrial Revolution is absurd.
Who thinks that we're wearing nicer clothes than people were 150 years ago?
Some of us are, but all of textiles were better.
Were people spending a lot more time on textiles?
Were many, many women actually stuck in textile production?
And are we better for it?
I think so.
But the idea that the mechanization was inherently a good and you'd be foolish not to see it and you don't want to be associated with these people because they were driven by fear.
There's error after error after error in that analysis, I think.
And frankly, that feels like arrogance and a kind of fear of its own, which is fear of not advancing so fast that you can't see what's coming.
Like, I feel like there's an addiction to adrenaline, to the unknown.
And I have that in different spheres as well.
But I think that to some degree, anyone who calls people concerned about AI and AI doomer and the people on this side are just talking past each other.
I don't think they're having the same conversation.
Yeah.
And in fact, there is a conflation there because the doomers have, the doomers, to the extent that that is a fair term, have a particular set of concerns about the AI turning on us, which I don't discount entirely.
But my point has been from the beginning, there are a bunch of ways in which this is guaranteed to create havoc that do not require the AI to turn on us at all.
So there's a middle ground.
And in fact, the guy, for better or worse, that initial piece is not about the AI turning on us.
It's about an economic disruption that is coming.
So anyway, I see the same mistake here where, I mean, in fact, not only is it a complicated systems versus complex systems question, but it is a it is a flaw of inductive logic.
What he's doing, he's looking at the history of things that revolutionized industries and he's saying this again.
Yeah, right.
And the point is...
The sky is falling, the sky is falling.
Yeah.
It's not a this again.
This is actually new for a number of reasons.
And I would point out several of them.
I mean, this is just off the top of my head.
The speed of the change, especially if it is self-facilitating.
Okay.
That's not something you need time to metabolize something.
The machines that revolutionize textiles don't suddenly appear in every textile shop by virtue of the fact that they came in through a wire that's plugged into the wall.
Right.
Right.
So the AI thing is different by virtue of speed.
It's different in the sense that it talks to us.
It has our API.
That's a fundamental change that's brand new.
The positive feedback nature of it, the fact that AI is going to be accelerating its own development means that the chance of catching up to it by metabolizing what's just happened is effectively zero.
There's the fact of the asymmetry between, yes, AI empowers everybody.
Absolutely.
But does it empower bad people who are completely unconstrained in what they are willing to contemplate more than it empowers good people who are constrained in the ways in which they will use this?
Yeah, probably.
So do you want to, you know, you just got 10 times more powerful.
Yes, but this person who hates me just got 100 times more powerful.
So there's that.
And then the final point is, I'm sorry.
Did you just tell me that we're not going to have an employment problem because of all of the opportunities we can't see that this is going to create?
Explain to me which of the opportunities that this is going to open up are not going to be better done by an AI-empowered humanoid robot that's capable of reversing directions and rolling its head around and moving much faster than a person and grabbing a new battery off the shelf and working all night.
So the idea that, yes, it's going to rob people of jobs, but it's going to create a whole bunch of jobs.
Are you sure jobs is what it's going to create?
I think it's going to create wealth.
It's going to create stuff.
But the point is where this ends, unfortunately, is very likely with the small number of people who control this technology looking at the rest of us and saying, you know, you're consuming resources.
You're cluttering up the view.
You're creating problems.
You're a danger in that you might revolt.
What do we need you for?
Right.
This ends with the elites having a conversation about useless eaters and deciding that there's something that has to be done about the rest of us.
And if you don't see that, that's because you're looking too narrowly at small examples where a technology has put these people out of work and created jobs for those people.
And that's not the question.
Yeah.
I agree.
Yeah, I was going to extol the virtues of drudge work, but I think I'll save that for another time.
All right.
I mean, the drudge work isn't going anywhere.
Well, no, I'll just say that since you said that, I have to respond.
The drudge work is precisely going everywhere.
Like that's, that's what gets picked off by technology first.
And he says that in that section of the piece that I just read.
He says, isn't everyone better off for having technology that does the work that you don't want to do?
And I don't want to give up our washer and dryer.
I really don't.
That is not what I'm advocating for.
I'm well aware of how much time has been spent.
And the kind of house drudge work has been overwhelmingly by women.
And clearing that allows, opens up great expanses.
But two things.
We need then to be finding meaning in other ways actively as opposed to responding to the decrease in drudge work by doing this all the time.
Like that's not, that's, that's worse by far.
That's, that's worse drudgery.
It's worse for us and it accomplishes nothing.
And there is also, I think, and again, I do not want to give up almost any of our conveniences, right?
I'm but, you know, we hand wash some of our stuff because reasonable people who have nice things hand wash things like wood-handled knives and sharp knives and such.
And we also have a dishwasher.
And I think if we had to give up the dishwasher, it'd be okay.
The dishwasher is nice for sure.
Especially since dishwashers require you to do most of the dishwashing before you put anything.
Yeah.
So, yeah, and there's, and, you know, we've never had one of them, one of them.
I did it like that, but we never had one of them Roombas.
One of them roombas.
Doesn't sound safe to me.
We've never had one of those self-vacuuming things in part because we have cats and I just don't know who would win.
But I actually really don't mind vacuuming.
Yeah.
And I take, especially if something really needs vacuuming.
And I know that that is not something that everyone feels.
But similarly, and this is not going to be an experience that most people watching this show will have had, but when I've been in the field in remote places with a tent platform and a waterfall to shower in, I have taken joy and I'm feet from a coarse red sand beach.
This is my particular spot in Madagascar where I did most of my dissertation research.
I would buy a non-commodity broom in town and took pleasure in sweeping the tent platform once or twice a day, in part because having sand that you're tracking into your tent is just a disaster that you don't want to compound, but in part because there's just something soothing and reflective and meditative about small repetitive tasks.
And if we say that every small repetitive task is drudgery and therefore the technology is here for our drudgery and nothing else, and of course we know it is here for something else too.
But if we say, well, but certainly you agree that you want technology to replace all the drudgery, I say, depends on how you define drudgery.
And we should consider what we might be losing when we lose at least some of the work that our forebears always had to do for themselves.
Yeah, that's very well said.
If the junk fills the space that is cleared by removing the drudgery, it wasn't necessarily a win.
And I did just want to mention in passing, I think something has gotten way, I feel the same way about vacuuming.
I'm not as good at being thorough.
I don't always see the stuff, but I don't mind breaking out the vacuum.
But in large measure, that's because of a revolution in vacuums, which was downstream of a revolution in batteries.
Revolution in Vacuums00:01:18
Right.
When the thing had to be plugged into the wall, it was a royal pain in the ass wrapping the cord up.
Well, and you're just always like around your own cord.
Right.
You're tripping over your own cord.
You might vacuum up your cord and break it.
You got to move to different plugs.
The thing is, it was like really not well done.
And the existence of batteries that allow the thing to be powerful enough to take your vacuum anywhere and everywhere and then throw it back on the wall radically changed how much it was drudgery rather than a repetitive task that you could do while thinking of other things.
Yeah, that's true.
All right.
Well, we kind of, we covered some things.
Yeah.
You want to go have some coffee?
I do.
All right.
Okay.
So there will be a couple of, oh, tomorrow, we're going to be here on Locals Q ⁇ A, 11 a.m. Pacific for a couple hours.
Join us.
Consider joining us.
That's on Locals.
There'll be a couple of Inside Rails in the second half of February with some great guests.
And then we'll be back in March.
But if that's too long for you, join us on Locals tomorrow.
We'll be here.
So what else?
I think that does it.
So until you see us next time, be good to the ones you love.