Good morning Players!

I have a confession to make. I love AI drama.

I’m obsessed with reading Tyler Cowen’s blog right now because he is sparring with every person on earth with a Substack right now.

Even though I don’t spend all day crafting prompts with ChatGPT, I feel enriched by the AI boom if only because I get to binge on some top tier erudite tea.

Today’s piece won’t seem like it’s about AI at first, but it is.

Everything is these days.

The Cobra Effect

Back in Colonial India, British colonists grew increasingly concerned by the presence of cobras near the city of Delhi.

Logically, they instituted a bounty for every dead cobra. At first, this strategy was successful and the number of venomous snakes dropped as locals killed them off for the reward.

Eventually, though, some budding entrepreneurs started to breed cobras specifically to kill them for extra income.

When the British became aware of this, they scrapped the policy. Now out of business, the breeders set their cobras free boosting the wild population of the snakes to higher levels than ever.

This is what’s known as unintended consequences. More precisely, a perverse incentive.

Perverse incentives

The more you look for perverse incentives, the more you find them everywhere.

• Railroad workers who were paid by the mile started building curving railroads to maximize profit.

• Corporations that got paid in carbon credits to remove the most dangerous chemicals like CFCs from the air started producing more CFCs in order to then remove them.

• Doctors who put stents in healthy patients because they received remunerations to do so.

• An online store operator who found that negative reviews on other sites boosted his SEO started going around and trashing his own product.

• A newsletter writer who was told to not use Wikipedia in middle school then uses the perverse incentive Wikipedia page to find examples of perverse incentives

Many of these examples are net-negative for society. They are examples of short term profit making in the face of long term progress. As you read examples after examples of the system being gamed and incentives being perverted, it’s easy to become pessimistic about humanity and the motivators that govern us.

But I see it in a different light. It’s a bit of a double-negative, but what about the idea of a reverse perverse incentive? Incentives that when they are perverted, produce positive-sum results in the long run.

Reverse perverse incentives

The story of the cobras reminds me of my time in school. My generation had one fact drilled into our heads by teachers: do not use Wikipedia. The danger being, of course, that Wikipedia was an unreliable source.

From 8th grade onwards I can recall having the fear of god put into us if Wikipedia showed up on our bibliography.

But everyone would still use Wikipedia because it’s fucking awesome for research. By explicitly telling us not to, more students ended up using it and found that going to the original sources cited by the Wikipedia pages was fantastic for filling out a bibliography for a research paper.

In short, we became cobra hunters. But instead of turning into a lying, no-good, plagiarizer, it turns out that using Wikipedia only became me more equipped to handle the real world upin leaving school.

My first job out of college at Morning Brew required me to quickly get up to speed on complex topic. Wikipedia became my best friend and I was glad for my days in school that made me adept at using it to its full extend

My relationship with Wikipedia is illustrative of a concept I call reverse perverse incentives. That is, incentives that push people towards bad behavior in the short term, that lead to better outcomes in the long term.

We need cobra breeders right now

The impetus for today’s piece came from seeing this recent David Sacks tweet.

What Wikipedia was for us back in the day, is what ChatGPT is for students now. The difference being that ChatGPT is infinitely more powerful than Wikipedia.

As teachers try desperately to rein in AI usage in school, it will only have the opposite effect. In this case, the “perverse” incentive is that students will become experts at using the most powerful technological advancement since the printing press.

This is perhaps the most high-stakes set of misaligned incentives ever recorded. If current and future generations of American students are incentivized to find creative solutions for using AI to escape plagiarism detection or cheating allegations, that could result in untold productivity boosts.

Goldman Sachs already forecasts AI will boost global GDP by 7% over the next decade. Just wait until Jimmy from 8th grade figures out how to get ChatGPT to feed him answers to his oral Spanish exam through a hidden earpiece…then solves hearing loss when he grows up.

As educators try to actively thwart the youths from selling cobras, we’re unintentionally raising the greatest generation of cobra breeders the world has ever seen.

Final thoughts

This piece, like so much of the current discourse, is ultimately about AI.

Right now, were on the crux of an AI boom that everyone from Elon Musk to Mrs. Smith in 9th-grade English is trying to suppress.

As tech influencers write open letters, and governments get their ducks in a row, I hope that we incentivize the correct or—and maybe this is even better—the incorrect things.

I believe that when it comes to AI, perverse incentives could end up saving humanity…or dooming it.