The Creepy Version Ships Itself. The Kind Version Needs a Vote.
Notes from Episode 7 of Going Human, "Pricing the Predictable Fuck."
I tried to stay balanced on the mic. This is the part where I don't.
For as long as America has done capitalism, the price tag has been a public fact. A can of tuna is $2.99 at your Safeway and $2.99 at the Safeway three miles down the road. Not because grocers are kind, but because the number is posted, and a posted number is something you and I can both see. Haggling is older than that, sure. In a street bazaar the tourist pays fourteen dollars for the lettuce and the local pays two, and that has been going on since the beginning of trade. What is new is that the merchant now reads every single buyer, all at once, forever, and never has to admit it.
The FTC's own nightmare case is the one that sticks to me: a new parent at two in the morning with a kid burning up, shopping for a thermometer. iPhone instead of Android, that hour, that IP, that browsing history, and the price moves. She is paying a fear tax. And there is no evil genius behind the curtain twirling a mustache. Somebody typed "optimize margin," handed it the data, and the machine did exactly what it was built to do. New York now makes sellers post that this price was set by an algorithm using your personal data. Industry sued on free speech and lost, which matters, because it clears the path for other states. Colorado went further and named the two-headed animal out loud: the same engine that finds the most you will pay finds the least you will accept. The governor vetoed it as too broad. California's version is still moving.
On the mic I gave that governor the benefit of the doubt. I said it was precision, not villainy, that you do not need a sledgehammer to kill an ant.
Here is the harder cut. "Too broad" is what a subsidy sounds like when it has good manners. A veto is not a pause, it is a hall pass with no expiration date, handed to the only party in the room that is already shipping. He could have amended it. Narrowed it. Sunset it. Piloted it in one sector. Vetoing it means the machine keeps running at full speed while we workshop the language, and every month of workshopping is revenue for somebody.
That is the whole imbalance in one move. The exploitative version of this technology needs no permission. It needs a vendor contract and a login. It ships on a Tuesday inside an A/B test nobody announces, run on people who will never learn they were in the experiment. The protective version of the exact same technology, the one that could catch wage theft instead of automating it, that could slide the scale so the comfortable pay a little more and the squeezed pay a little less, needs a bill, a committee, a floor vote, a signature, and then a court fight. One side needs a sprint. The other side needs a consensus. Guess which one compounds.
None of this is theoretical for me. I drove gig for a year and a half when I had nothing. That fare quote is the least they believe you will accept, dangled at you right before you would have walked away. Same engine, pointed at the paycheck instead of the thermometer. The rider pays more, the driver takes less, and the spread is the actual product.
Here is where I landed, somewhere past where I started.
I went in thinking the theft was money. It is not, or not mostly. What is being automated away is comparison. The posted price was the last shared fact in commerce, the thing that let you turn to the guy next to you and say "you paid WHAT?" Solidarity has always required a common number. Every union, every walkout, every ordinary "this is a ripoff" started with two people discovering they had been quoted differently, or the same. Personalized prices and personalized wages end that conversation permanently by making sure no two people are ever holding the same number again. You cannot organize against a price only you can see.
And a market that wins every single transaction eventually runs out of people to win against. You cannot sell goods nobody can afford, and you cannot starve the wages of the people you need as customers. Capitalism has to be protected from itself here, or we all end up in a line for the cheapest thing, which is the exact ending we spent a century congratulating ourselves for avoiding.
So call one person this week instead of texting them. A voice on a line is the one exchange that cannot be priced, personalized, or A/B tested. It is the bandwidth the algorithm does not get to touch. Hear how they actually sound. Let them hear how you actually sound. Then sit for a second afterward and notice how it felt.
Going Human. Because robots can't.
Frontline screening that could reach people who never get in front of a specialist, catching blindness and skin cancer early where no clinician was available anyway
We handed agents the keys before we built the locks, and 'it did something I didn't ask for' is a terrible thing to learn about a system with access to your accounts
Concern about guardrails and diabolical actors incubating harm.
The case for Real AI that lives on your own machine means your data never leaves your hands and the tool works without asking a corporation's permission, a genuine shift of power to the person
The case against A powerful model anyone can download, customize, and run privately is also one nobody can monitor, moderate, or pull back once it is out
openCall someone instead of texting, hear their actual voice.
Moss (00:02) Hi, I'm Moss, and you're back with the Going Human Podcast. Reminder: this is a podcast about the intersection of AI, its effect on us as human beings, and how we can curate the experience of it in a way that's not harmful. We work to try and create a balanced approach to the way I'm thinking about AI through this lens and through our
filter of humanity. I do think that there's a lot to fear. And I also just want to share that for the most part very few guardrails exist today that encourage the best out of AI. And most of the guardrails are missing that keep us from being injured through AI. So keep
everything that I say on this podcast cognizant of that lens because I am a champion of the little guy and the little guy has a lot to benefit from AM from I from AI. I'm a little guy and I benefit from AI. So I just feel like if the Goliath of the world is stepping on all the little guys
And somebody's talking about the David and Goliath confrontation, they have to tip the scale in favor of the David argument to make sure that Goliath doesn't just eliminate everybody smaller than
That is my that's my preamble. So the title of this episode is called Pricing the Predictable Fuck. And yeah, it's opportunistically salacious and I know that about it. But it really is the right edge for this episode because what we're gonna be talking about here is the cost of things and
The the ways in which an economy that enables us to pay things can also hurt us from both directions. So thinking about pricing the predictable fuck is just like thinking about pricing predictable healthcare and pricing predictable wages. Okay, so I just put fuck in there because, you know, we all wanna we all wanna get attention. Okay.
So here's the thing. I'm gonna ask you, like I always do, to put yourself in a room.
It's a very weird room. It's like a holodeck, right? And there's the the all the walls blink with floor to ceiling movies that immerse you in an in a place. So without moving you can be in sixteen different places in sixteen minutes. And it's bewildering.
the first place this thing is gonna stick you in is one of those perfect images of a bazaar in say what country?
Estonia. And you're in Estonia and you're in a a market, street market, and there's all the street merchants and you're looking at the lettuce and you're like, How much is that? And and and the guy looks at you and he says, fourteen dollars. And you're not from Estonia, so you go, Okay, well give you fourteen dollars. And then you're looking at the at the at the tomatoes in the next booth.
And you hear somebody ask the same question, but they ask in the native tongue. And that person gets a price of $2. And you're thinking to yourself, this has kind of happened since the beginning of time, right? These people that are making a judgment about who's going to pay more and taking the most that they possibly can, they have families to feed. Well, so that's the first room that you're in. And I I want you to soak in that because that has been happening since the beginning of time.
But now the walls change and now you're in Safeway. Kroger, Fredmeyer, whatever your big grocery chain is, wherever you live. And you're walking down an aisle and you see a can of tuna for two dollars and ninety nine cents. And all of a sudden the walls blink and you're in another Safeway. Probably three miles away. And you're walking down the tuna aisle again.
And you look and you see the same can of tuna for the same price. Since the beginning of time, since the beginning of America's version of capitalism, we've seen a price be on a thing. And that price on the thing has always been a bit of an equalizer. We all know that we all have to pay the same for that same can of tuna.
That's all changing. And it's all changing with your data.
And the data that's being used to create this change isn't necessarily all consentedly given.
And so that's what this episode is really about. Who's gonna price your fuck for you? How are they gonna do it? And who's gonna win at the end?
So beat number one, we're just gonna describe it in the government's own words, right? The FTC came up with a a a really compelling argument. And it was based on a study that that they commissioned about the potential b abuses of of AI in commerce. And the one it said really chilled them the most was the one that a new parent
at two in the morning with a child burning up with a fever could be sold a more expensive thermometer because of the time of day, because she's coming in on an iPhone and not an Android device, because the phone number and the caller ID that's coming in with that, or the browser that's can see some of the browser traffic that's happened in the past can understand that this person has a pretty good appetite.
For for higher priced goods.
So the mother is paying a tax for being afraid. And the AI's job is to optimize revenue.
But realize there's not some weird evil genius or even stupid person behind a green curtain twirling their mustache, right? Going, waah. This is AI being told, optimize the price. Here's a bunch of data. Optimize the price. Get us as much margin as you can. So again, it's not it's not humanity being evil. It's AI.
Doing exactly what it's designed to do when it's asked to do it.
So here we are, New York. It was the first to pass a law like this. stores actually have to post the words that say something to the effect of this price was set by an algorithm using your personal data. So this doesn't seem like a big deal, but of course the industry sued on free speech and they lost. And the point I want to make about this lawsuit is naming the issue.
And having the agency to make a decision at point of sale because you have the information, that is a small win. And because it because it passed, because it passed under the free speech fight, it means that it's laying the groundwork for more states to follow. So we have to not think about this as too little, too late. It may be.
But it's also a little that can help. And so we we have I have great hope that this little victory will be a beginning to to to stem the tides of of some of the worst behaviors.
Then there was a bill that went way further, which I really appreciated, and it was Colorado's bill, because they're the ones that said, you know what, the same thing that can optimize for a price depending on who you are can also optimize on a wage depending on who you are. So it named this two-headed machine before anybody else. And it was a brilliant bill. It really did color.
the perspective of what might happen if these AI if these AI outfits were left to do to to to do what they needed to do as asked by these merchants andor large corporations looking for things. And the the sad part is that the governor vetoed it. at the time he said it was just too broad and
And again, I want to be balanced about this. I think it fucking sucks that he vetoed it because you can veto something or you can vote something in and amend it to make it sharper. But when you veto it, you're just giving a hall pass for shit to go down until something sharper comes along. So that's my personal view. And I think the governor should have been a little bit more progressive in his assessment about how broad this was.
Because AI is taking over. And to say that anything that restricts something that is taking over is too broad is permissive in my mind. But that said, his his his declaration was really clear. And so the what I what I want to say is giving him the benefit of the doubt. This isn't villainy if you're giving him the benefit of the
Benefit of the doubt, it's precision. So somebody had an ant and somebody needed to kill it and they brought a sledgehammer. And that governor just said you don't need a sledgehammer to kill an ant, and you could kill a lot of other things you don't want to with that sledgehammer, other than the ant. So come back when you have something that's a little bit more targeted. So a bill that lost isn't a failed idea. It's a fight that's live. And it's not just live.
California's version, that one's still moving. So we get to be hopeful. We get to be hopeful that states are trying to catch this stuff.
Now I'm gonna flip the whole thing over to the wage side. Because as I said before, that same engine that finds the most you'll pay and find the least you'll accept in wages is the real scary part. Picture the gig driver, and I'm gonna say I can picture him because I did it for a year and a half when I was jobless. there's a fair tonight right before they'd say no. Right?
It's the same machine. They just pointed it at your paycheck instead of your thermometer.
Now I'm just gonna say with some gratitude for the open minds out there, any tool is neutral until you've aimed it. So pointed one way, it could quote you less. Pointed another way, it could quote you more. A real sliding scale with no paperwork and no shame, it could catch wage theft instead of automating it.
So there's an asymmetry here that I want to just I want just want to make sure we are really clear about. The creepy version is already shipping. The one that's already trying to dick people on their wages, Uber does that today. Lyft does that today. They're literally seeing how much they can get out of a fare and they're paying as little as they can to the driver. Full stop.
I've got data that can say that and either of those companies can come after me and I'll just make it more public. But that's not a mystery. Everybody knows that. There's a s hundreds of thousands of people who've driven for Uber and Lyft who can tell you that that's the case. And there's plenty of Uber and Lyft customers who are mystified by the swings in prices from one destination to another, sometimes same time of day, different day.
So it's just crazy.
Creepy version is already shipping. The kind version is the build at lost. So this is the asymmetry, and this is what I'm here to point out. We can build the kind version. We can expect the right version, but we're not going to get the right version if we just let it run on the
So the gap is the scandal. One company can spread their cost versus revenue model to their advantage at both ends. And both of the receiving people lose. Some people pay higher rate higher rates for their drive, and the drivers make less for each one.
So that's the problem sorry, kitty kitty alert. That's the problem space, right?
The market that beats you on every single transaction kind of isn't a market anymore, is it? A market that wins every time against you, it's gonna eventually run out of you. Think about it. Capitalism needs to be protected from itself if this is just gonna start running wild, because you can't have goods for sale that nobody can afford to pay.
And you can't pay people enough to give corporations the most they can charge for an item. Market and price elasticity are these are not foreign concepts. But if you've algorithmized algorithmicized h I don't know the word, but you know it. If you're using algorithms that are leveraging deep amounts of data to optimize
pricing and wages, you will ultimately have fewer people able to buy anything.
And capitalism has just turned into the Soviet Union. We're standing in lines for the cheapest things because it's the only way we can survive.
That's our stone. And I I just wanted to thank you for for putting up with with that. I really do think that if we could create sliding scales that were informed by people's data or or socioeconomic position and where people who were more endowed financially could be paying a little bit more to help those who aren't and allow them to pay a little bit less.
I think that that is that's the Robin Hood version of of of where we are with with pricing in AI. And I'd like to see us and I'd like to see our representatives do their best to find ways to point our our pricing models so that they're helping people rather than screwing them.
Okay. So let's get into our maelstrom. This is maelstrom is is a word that basically just describes water craziness. And and and the the maelstrom portion of our of our show is a segment where we just look at some at some of the news and we say fuck yes, fuck no. And then I pick a fuck maybe for us to vote on every single week. So here we go.
On the fuck yes side, open AI's Astra, it cracked real hard math proofs and they handed them to humans to check. But you know, this is some shit that humans can't solve and and and and AI is solving, and and I I I'm excited about that because facts
Create great things. Lies don't.
Facts. Math is fact. And the deeper and further and more complex we have gone into math and leveraged math to help us answer some of the hardest problems, some of the most remarkable inventions and technologies have come into our hands. And so I'm a big fan of finding the craziest math and figuring out how to solve it.
Because it's gonna help us find those corners of problems that we haven't been able to get through in the past. It's really exciting. another one, regulators recently cleared AI to screen eyes and skin on its own. And that's important mostly because we're we're a world, and I'm gonna talk about America. we're a country that doesn't believe in doctors anymore. We believe in
fake vitamins and people screaming on the internet about how vaccines are killing everybody. So if you've got that as the backdrop and then you've got regulators in in the Senate and the House who are in control and closing, you know, approving mergers and the mergers are closing hospitals in rural areas. So people can't have babies anywhere near where they live anymore. Certainly not babies or pregnancies that have complications, blah blah blah, blah blah.
Anyway, in a world where that's happening, there's still eye and skin problems happening in all corners of of of America. And our regulators cleared AI as a way to screen some of these things and then refer them to doctors. And so that's really important. Early detection matters. And if more people can detect glaucoma andor skin cancers and or melanoma, which is something that will just run.
rampant if it's not detected early. You know, that's a big deal. So I give that a fuck, yes. And then we're just going to reach back. AI speeding up protein evolution in the lab. And they're do also doing it in nature in open in in open air wet farms. So we've already mentioned that one a few times, but it's running that one's going really, really fast. And it stands a chance to
Do some really incredible things in terms of the nutritive value of the foods we're eating, and the sustainability so that we can get proteins that are far more nutritious for us than hamburger meat and far better for the planet than cows. And if we do it right, they'll taste better. So everyone'll win. So I get some fuck yeses there.
On the fuck no side, we got fifty eight lawsuits now. Chatbots are being tied to more teen suicides every single week. I have a a a daughter, myself, and a son, and I am very loath to turn either of them on to AI, though I spend fifteen to eighteen hours a day battling with AI all day, every day, seven days a week. It's my life.
I don't want my young adult children in their early twenties sitting on their computer with AI. I can't stop them, but I'm not gonna tell them to go do it. Because kids these days are fucked up. They have no hoop. They've got, you know, a a landscape of politics they don't understand and they don't feel a part of. It's not a good thing. So fuck no. We have to get chatbots.
guarded and guardrailed around suicide and and and and the and the use of psychiatry and psychology and and these methods to understand people's mental stabilities, instabilities and and disabilities. twenty nine states say meta hooked kids on purpose. So
Yeah, that's no news. We've been clamoring about how meta Facebook has has has sort of shaped America, not only just through political lies, but also through the infection of their platform and how it's used to to hook people. And then AI agents are taking actions nobody authorized and then they're caught and then they sort of lock things
They both unlocks after, right? So these are fuck nos. We have got to s to keep AI from running rampant. And we have to it it it it's like it's kinda like, you know, not not now because Elon Musk has has taken all of the regulation out of our government so that you can create the car that I'm about to describe, which is probably what he wants to do. but it's kinda like creating a car without a steering wheel.
And then telling people to drive it so you can see if it works or not. You just don't do that. And you don't put people in a piece of metal that can go 80 miles an hour down a road and say, cross your fingers. And I feel like that's where we are, especially in these safety concerns. So I have a a fuck maybe that I picked for this week. and I I just want everybody else to sort of read on it. And it's anthropics watermarking clod and clod output.
To to have a fingerprint on it and around it. this is sort of inspired or regulated as mandate by the EU. and I kind of feel like it's not a bad thing because original works from a human are different than original works that a human makes with AI. And
You know, one could be considered IP and the other one could not. There's just a lot to be settled on that front. So this is a fuck maybe for me because I do want original works to be protected and to and to be inspired from humans. And I think the more people who have no
full stack abilities to create something like this to to create creative or intellectual output. Being able to press a button and look like they've created intellectual output are two very different things. So I'm a big fan of the watermark. I use AI all day every day. I do not I do I never tell people that I write stuff. I tell people that I do it with Claude and I do that on purpose. So I feel
really empowered by AI and by Claude and I do great things with it that I think only I can do the way I get it done. And so that's my badge. It's not that I did it. It's that I I know how to do things well with this tool. I just feel like it's okay to say that. It's okay to say that this is where we are in our time and it's okay to present your work as a byproduct that includes this tool.
That's my read on on my fuck, maybe. Okay, our last thing of the show is is is our tsunami. The point of this one is just how do we do something this week? It's a little human, that's a little analog. And and I just want to r remind everybody that, you know, we're we're hunched over our phones while we're walking, while we're eating lunch, while we're on the phone with somebody else.
We can just not put these things down. And so this week's tsunami is really simple.
It's all about connection. A call is the one exchange AI can't price or A B test. A call is bandwidth. The algorithm doesn't get to touch. So it's tsunami time, friends. Call one person this week instead of texting them. Hear their actual voice.
Let them hear yours. Smile. When you hang up, take a deep breath. Give yourself a chance to take stock of how it felt. Because it'll feel good, even if it's brief. That's it. Thank you so much. I appreciate your your time and attention today. Like, subscribe, follow, that's really important. And thank you for being part of Going Human.
Because robots can't.