I'm very sorry to report that Y Combinator founder Paul Graham has decided to dig in his heels about the correlation "between founders having very strong foreign accents and their companies doing badly."
Rather than listen to feedback about how patterns can turn to fallacies if they're not questioned, how harder-to-understand immigrants fuel innovation, and how unacknowledged biases can prevent you from seeing good ideas, Graham has decided to keep cloaking his calcified thinking under the banner of "empirical evidence."
Although Graham already had an opportunity to clarify his refrain about strong foreign accents on Hacker News, on Twitter, and even on villainous Valleywag—an opportunity he won't extend to immigrants with strong accents, mind you—he begins by explaining that his aversion to strong foreign accents is merely a communication issue.
Here's the thing, Paul. We all know what you meant. We're trying to tell you that you are not seeing the bias behind that. It might be a lot easier to understand a 24-year-old Stanford grad saying, "Uber for unicorns," but that doesn't mean the company isn't going to do badly as well. If you think I'm being facetious, take a look at the X for Y companies that got funded in Y Combinator's last class. Me-too ideas abound.
Eventually, Graham's essay circles back to good ole "empiricism":
We have a lot of empirical evidence that there's a threshold beyond which the difficulty of understanding the CEO harms a company's prospects. And while we don't know exactly how, I'm pretty sure the problem is not merely that investors have trouble understanding the company's Demo Day presentation. Demo Day presentations are only 2 minutes and 30 seconds. With a presentation that short, you can just memorize it at the level of individual phonemes. Most batches we have groups that do this.
This data set is from the 564 companies accepted into Y Combinator, and is therefore statistically insignificant. Another way you know it's insignificant? Because Graham cannot describe A SINGLE "BAD" CORRELATION attached to "strong foreign accents." Surely with YC's scant number of home runs, there are more pressing observations about what type of founders and ideas find success.
[Empirical evidence] is a good (though debatable) approach in scientific experiments, but when used to make judgments about peoples, especially if you don’t have a clear explanations for your own beliefs, then it is worrisome. When an empirical observation is passed out as wisdom, all it does is set up a dangerous precedent and such thinking only reinforces certain canned beliefs and propagates fallacies as truths and maxims.
Another case of Graham being blind to his own bias: he says that the reason he hasn't just let the controversy blow over is because his intention is "to help founders." If I can also have a chance to clarify: having the "Bertrand Russell of Silicon Valley" harp on strong accents above all else isn't helping anyone.
To contact the author of this post, please email firstname.lastname@example.org.
[Image via Getty]