066 - Bad Business Models

"Reliably Bad & Overpriced" is one of the most successful business models in the world.

Fast Food chains are an excellent example of this, as they provide reliably bad food (but also reliably not the worst), at a 20-40% markup relative to authentic local places.

This business model thrives thanks to many of the same cognitive biases that lead investors to make reliably bad (but again, usually not the worst) investment decisions. This can be considered as the 20% mark, where 80% of decisions would have been better, and the remainder would have been worse.

Fundamentally, most people would rather invest in someone they know, even if they know going in that it is a bad idea. The meaning underlying the phrase "The devil you know" is such a strong influence that many would rather take the reliably bad option than a 50/50 coin flip on the unknown. This is called the "Neglect of Probability" bias, and other cognitive biases such as Normalcy Bias and Survivorship Bias can also play key roles.

As noted by a few famous researchers in the field, even when the one(s) in charge want their people to take reasonable risks, such as CEOs encouraging research and innovation, most subordinates are far too risk-averse to make a rational decision. They'll routinely refuse a 50/50 chance with a 10x or greater payoff, crippling their companies in the process.

We see the same thing in AI now, both in the offerings and on the investment side. The offers that receive funding are most often "Reliably Bad & Overpriced", and the realistic returns on investment are frequently nonexistent. Worse still, many of the AI offerings are steadily drifting deeper and deeper into fraud. The growing number of Tech executives who've committed obvious perjury and have yet to face prison time is one metric highlighting this drift.

These dynamics work to the detriment of all parties, with both near and long-term consequences. Human stupidity is one market force that should never be underestimated. At the moment, it appears to be the dominant factor.