Welcome back.

Have you thought about subscribing? It's free.
seths.blog/subscribe

The ghost in the machine

“The computer wants you to click this button.”

“It thinks you asked for something else.”

“He’s mad at you.”

Thousands of generations ago, we evolved our way into a magnificent hack. It turns out that we can more safely navigate the world by imagining that other people have a little voice in their heads just as we have one in ours.

By projecting the narrative voice to others, we avoided fights that could be fatal. It’s a powerful shorthand that allows us to use limited brain processing power to interact in complicated cultural situations.

It worked so well, we began applying it to dogs, to lizards and even to the weather. It’s a great place to find the origins of bad decisions and superstitions.

The truth, of course, is that your cat doesn’t have a voice in her head. But we still act like she does. And that cloud doesn’t really have an angry face in it, a bug we see so often that we even gave it a name. Pareidolia is proof that the mistake is almost universal.

And now, AI chat is putting the common sense of this to the test. We know exactly what the code base is, and yet within minutes, most normal humans are happily chatting away, bringing the very emotions to the computer that we’d bring to another person. We rarely do this with elevators or door handles, but once a device gets much more complicated than that, we start to imagine the ghost inside the machine.

If it’s working, keep at it.

The problems arise when the hack stops working. When we start making up stories about the narrative intent of complex systems. Sooner or later, we end up with conspiracies, misunderstandings about public health and opportunities missed in the financial markets.

Emergent behaviors (like the economy and computers and the natural world) aren’t conscious.

It’s hard to say, “I know I’m making up a human-centric story to explain systemic phenomena, but it’s a shortcut I use… do you think the shortcut is helpful here?”

Shields up

Years and years ago, I helped the Weekly World News make a book.

While their periodical was weekly, it certainly wasn’t news. They were just four people in a small office in Florida. They gleefully made stuff up every week. They had a few filing cabinets of stock photos, and they invented stories featuring UFOs, aliens, “scientists” (in quotation marks) and various other diversions for folks trapped in the checkout supermarket line.

And now, of course, we are all trapped in that line. And now, the algorithms are pushing spineless profit-seekers to bombard us with junk, junk that shows up on the home page of search engines, in our social media feeds and in our email.

Adblockers are one of the most popular innovations of the last few years. What I want is a junkblocker. A big button on my browser that says “shields up.” And just imagine if it was set to on by default.

No celebrity gossip. No conspiracy theories. No weight loss breakthroughs. It would automatically block fist fights, trolling, urgent but unimportant breaking news, insights about the royal family, discussions of whatever happened to a star from thirty years ago, aliens, UFOs, MLMs, the latest pump-and-dump schemes, things that are true but irrelevant, things that are relevant but didn’t actually happen and stories designed to demean, degrade or intentionally inflict distress with little recourse available.

When you put it that way, who doesn’t want a button like that?

Somehow, we survived as a culture for centuries without exposing ourselves to thousands of profit-driven manipulations dumped on our living room carpet all day, every day.

No wonder we’re exhausted after a day online.

The gap between impossible and normal

It keeps getting shorter and shorter.

This video couldn’t have been made, at any price, 18 months ago. 18 weeks ago, it would have required a thousand hours of work.

Now, here it is. This impossible is going to happen faster and faster and faster.

Is it possible to care at scale?

After 25 years, I stopped using a certain credit card for business. It was easily millions of dollars worth of transactions over that period. Did anyone at the company notice? Did anyone care?

I still remember losing a client in 1987. Small organizations pay attention and care very much about each and every customer. Verizon and AT&T, on the other hand, don’t even know that you and I exist.

Small family farms have significantly higher yields than neighboring farms that are much bigger. That’s because the individual farmer cares about every single stalk and frond, and the person with a lot of land is more focused on what they think of as the big picture.

But it’s pretty clear that if you add up enough small things, you get to the big one.

Caring at scale can’t be done by the CEO or a VP. But what these folks can do is create a culture that cares. They can hire people who are predisposed to care. They can pay attention to the people who care and measure things that matter instead of chasing the short term.

Large organizations have significant structural advantages. But the real impacts happen when they act like small ones.

Your own billboard

Large sections of Los Angeles are studded with billboards for minor TV shows. These billboards exist nowhere else, even though there are televisions globally.

Obviously, there’s ego at work here, but it’s sort of productive.

First, there’s the ego of the producers/networks. They like showing their peers what they’re up to, and it probably makes it easier to recruit the talent that lives nearby. If you’re in the famous business, being more famous, even locally, is a boost.

And then there’s the ego of the stars. After all, if they see the billboard, it’s as if everyone sees it.

Social media is simply a smaller scale digital example of this very tendency.

And getting your billboard right–and doing work that makes it easier to get your billboard right–might be one of the single best side effects of useful social media.

But, like billboards in LA, it’s best to not take them too seriously.

Simple techniques for complex projects

Warm up the machines that take a long time first.

Stress test the go/no go parts of the project as early as possible.

If the cost is low, replace dependent processes with parallel ones.

Do the difficult parts when energy is high and the budget hasn’t been depleted.

Ship before you run out of time or money.

Invest in slack buffers for any critical dependent components.

Budgets are a tool, not a weapon.

Thrash early, then lock down decisions and don’t change them.

Assign each task to the least expensive/least busy people able to get the job done. (This is probably not the project manager).

Prioritize feedback from people with taste, skill and influence, not proximity or volume.

Identify go/no go points based on irreversible actions or unrecoverable costs.

Anticipate the difficult and high-risk moments in advance and reserve resources for the ones you can’t anticipate.

Heroism is more fun but less reliable than good planning.

Conspicuous (non) consumption

One way to show status is by demonstrating how many resources you have. A bespoke suit, a huge graduation party, a fancy building… A bully who physically intimidates or an angry driver who cuts you off in traffic are each working to show their status and strength.

But it’s also possible to demonstrate security and confidence by doing precisely the opposite. The billionaire in a t-shirt. The person who holds the door open and lets you go first in line… these are also demonstrations of status.

The interesting question isn’t whether someone has status. It’s whether they’re gutsy enough to demonstrate it by making things better for others.

Is it a t-shirt brand?

Not all projects become t-shirt brands, nor should they.

The risk is in thinking you’re building one when you’re not. T-shirt worthy brands are a very small subset of the whole.

The question is: Would your customers want to wear your logo on a t-shirt?

Why?

If you’re creating identity, possibility, connection and giving folks status, it’s easy to see how you could build a t-shirt brand in just about any field. Sports teams do it for a living. Google had a t-shirt brand for a long time, and so does Penguin Magic and even Festool. I’m not sure, though, that many people want a t-shirt from BMO bank, Marriott or International Paper. Netflix might be, Roku isn’t. Of course, no t-shirt brand is for everyone, that’s part of the point.

If you’re simply providing a good service at a good price, perhaps you don’t need to go to all those meetings and waste so much time and money on “branding.”

Why would someone want to wear your name around town? What’s in it for them? Go build that and the t-shirts will take care of themselves.

Allocating scarcity

If we’re lucky, we invent something that’s going to be in high demand. Reservations at a hot restaurant. Limited edition trading cards. Concert tickets…

How to decide who gets them?

One attractive option is “first-come-first-served.” It feels fair, after all. The theory is that people who really want what you have will spend time (waste time) in line to show their commitment. But of course, this is a tax, and an uneven one at that, since some people value their time more than others.

Another is to simply auction off the scarce items. The good news is that the value of the scarce item won’t be squandered on time wasting, but will go to the company. But this might feel unfair, as it rewards people with more assets, as so many things do. On the other hand, it’s pretty clear that people allocate resources differently than we might expect.

The third method, the fairest of all, is to have a lottery. Invite your best customers, or charge a commitment fee, and then randomly allocate the loot. The good news is that you won’t alienate customers who feel as though it’s their fault that they didn’t wait in line long enough, or spend enough.

Each decision has effects. And it’s up to the producer to decide which emotions they want to be responsible for creating.

Revisiting stamps for email

I started agitating for this in 1997 and wrote about it in 2006. The problem with the magical medium of email is that it’s an open API. Anyone with a computer can plug into it, without anyone’s consent.

This creates an asymmetric attention problem. The selfish, short-term-thinking sender benefits by emailing as many people as possible, and the recipients suffer.

This doesn’t happen with traditional mail, because there’s a cost to sending it.

With GPT arriving, expect that spam is going to increase 100x, and that it will be eerily personalized, invasive and persistent. That it will be really difficult to believe that an email isn’t junk, because there’s going to be so much junk, and it’s going to be harder to filter.

And yet, email is powerful, and convenient and we’ve been using it for our entire careers. Is it doomed?

Some apps are showing up that are trying to create a paywall for email. An unknown sender has to make a donation to charity (the recipient specifies the amount) to reach your inbox. People have tried this off and on for decades, but it’s hard. There are two problems with this being widely adopted.

The first is that it creates an attention obligation on the part of the recipient. It’s socially awkward to sell access to your inbox and then ignore the email.

The second is that there isn’t much of a network effect, and while a few people might adopt it, the problems with email don’t improve unless it’s widespread and persistent.

Here’s an alternative:

A simple plugin for gmail (and then, eventually other providers) that tags the email you send and receive.

Senders who send more than 50 emails a day need to buy “stamps”, perhaps for a penny each. The money goes into escrow.

Recipients can easily mark an email as unwanted. They can also upvote an email, which will send a signal that allows their peers to be sure they don’t ignore what they just got.

If enough people mark your emails as unwanted, you lose your escrow, it goes to a worthy cause. If it’s legit, the escrow remains and you don’t have to buy more stamps.

If a sender doesn’t use the system, they’re not going to be able to reach any of the people who do. So not many people have to be early adopters before it becomes widespread–if you want to reach most people (and you don’t know which people have it and which don’t) you’re going to need to turn on the tagging. It’s a tiny cost to pay for attention in a world where attention is scarce.

Normal people won’t have to pay anything, and email will get better for them as senders and receivers. And businesses that mean well and do well ought to be happy to pay.

If too many senders view the penny stamp as a chance to spam people (and lose the penny) then just increase the cost of the stamp to a nickel, etc. Pretty soon, algorithmic spamming is simply not going to pay off.

Giving anonymous people and organizations the chance to steal your attention all day, at scale, seems like a worse idea every day.