Americans got 4.35 billion robocalls in July. One month. That works out to roughly 139 million a day according to the YouMail Robocall Index, and it does not even count the more than 600 million spam texts Robokiller says we absorb every single day on top of it. So before we talk about anything else, lets just sit with that. The phone in your pocket, the most personal device you own, is a wide open door that anyone on earth can walk through for free.
I am a real estate broker. My cell number has been on yard signs, websites, and public filings for seventeen years. I might be one of the most spam-called people in Texas (not a title I was going for). So this is not an abstract problem for me, it is a Tuesday.
Why the filters keep losing
Here is the thing nobody in the phone business wants to say out loud. Spam filtering is a guessing game, and the other side gets to study the guesses. Your carrier looks at a call and guesses whether it is junk based on volume, patterns, and the reputation of the number. The spammers respond by rotating numbers, spoofing caller ID, and rewriting scripts. The carriers adjust, the spammers adjust back. We have been running this exact loop for twenty years and the number is still four billion a month, so I would argue the loop does not work.
And AI just made the losing side stronger. The scripts are better, the voices are convincing, and a bot can now hold an actual conversation with your grandmother. The cost of running a scam at scale dropped to basically zero.
Even STIR/SHAKEN, the big industry fix, only verifies that a call really came from the number it claims to come from. Ok, fine, that helps with spoofing. But it says nothing about who, or what, is on the other end. We spent a decade verifying the phone number and forgot to verify the caller.
The actual problem underneath
Spam exists because sending is free and anonymous at infinite scale. One person with a laptop can place a million calls, and if three people bite, the math works. Nassim Taleb wrote a whole book about skin in the game, and spam is what you get when there is absolutely none.
So here is the question I keep chewing on. What would it take to end spam permanently? Not filter it better. End it.
You would have to break the economics, and the way you break the economics is presence. If every call or text required proof that a real live human was on the sending end at the moment it was sent, the million-call laptop dies instantly. A human can only be present for so many conversations a day. Scale is the entire business model of spam, and human presence is the one thing on earth that does not scale.
Would that be a big lift to build? Sure. Is it a more interesting problem than building filter number 4,001? It is not even close.
This is the first post in a short series on what I think is the defining problem of the next decade of the internet: knowing there is a human on the other end. Next week, what happens when three seconds of audio is enough to steal your mom’s voice.