Billionaire Michael Saylor Says You Now Have to Put Typos in Your Messages to Prove a Human Wrote Them: ‘If You Put Errors In It, I Believe You Typed It’

In a recent interview on Diary of a CEO, billionaire Bitcoin (BTCUSD) advocate Michael Saylor made an observation about email that has nothing to do with money and is harder to shake than most of what he said about markets. “Now when you send a message to someone, if you want to prove that it came from you, you have to actually put errors in it,” he said. “Right? It's like if you put errors in it, I believe you typed it. The AI can draft the thing as though it had a PhD in English, and it had 20 years' experience as a copy editor.”

Saylor, the executive chairman of Strategy (MSTR), said it on the Diary of a CEO podcast in an interview published on Aug. 6. It was an aside rather than an argument, which is probably why nobody outside a couple of transcript aggregators picked it up.

It lands because the signals people used to read authorship from have quietly stopped working. A colleague's clipped sentences, a friend's habitual comma splice, the boss who never capitalizes anything: those were fingerprints. Fluency used to cost effort and therefore meant something. It now costs nothing, and the polish that once implied care can just as easily imply that nobody was there at all.

Saylor then took it somewhere harder. “We're reaching this point where lack of effectiveness is just laziness, right?” he said. “Like, if you wrote something which wasn't perfect, it's cuz you're lazy, not because you're not perfect.”

Sponsored Content: The Trillion-Dollar Energy Opportunity Investors May Be Missing

Although if typos really were proof of humanity, the software built to detect machine writing would be looking for them. It is not. The two things AI detectors actually measure are perplexity, meaning how predictable each next word is, and burstiness, meaning how much sentence length and structure vary. Errors do raise perplexity, which is why the intuition survives. But detection vendors report that models built since 2024 explicitly normalize for typos, slang, and grammatical noise, because they were trained on human writing that contains all three. The trick Saylor describes works on a reader. It does not appear to work on a machine.

In the interview, the thought was a stepping stone to somebody else's argument. “The AIs create perfect documents. They do perfect research. The robots will do any amount of work,” Saylor said. “And I think Elon makes this point, which is we're about to trip over an age of abundance.” Bartlett had read some of Musk's own remarks aloud a few minutes earlier, so the framing in that stretch of tape is the host's rendering of Musk rather than Musk speaking.

Saylor does not fully buy it. Elsewhere in the same conversation, he said of Musk's abundance argument, “He's half right. I agree with part of what he says, that is, consumer goods, consumables, utilitarian goods will become abundant, but there are always going to be scarce desirable goods that will not become abundant. And I think he overstates the case. Money will still be valuable.” Musk has made the abundance case in stronger terms elsewhere, including the suggestion that money itself stops mattering.

The practical residue is small and worth having anyway. A too-perfect email from someone who has never written one before is now weak evidence of nothing in particular, and the honest reading is that you cannot tell. If Saylor is right that the only remaining proof of authorship is imperfection, the awkward corollary is the one he did not say out loud: anyone who wants to fake being human already knows to add the typo.

On the date of publication, Caleb Naysmith did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.