Was reading the Neuron Daily’s article about Microsoft imposing chat rate limiting on Bing, and they wrote:

From now on, Bing will be confined to:

  • 50 chat turns per day (turn = question + reply).
  • 5 chat turns per session.

According to Microsoft’s data, only 1% of conversations exceeded 50 messages, so most users won’t be affected.

I don’t know what Bing chat’s current active users are, but according to CNBC, as of a few days ago, “more than a million people have signed up to test the chatbot.”

As you may know, 1% of 1M+ is 10,000+ people.

That figure may not be the same as their current active users. Who knows. What I do know is that product teams use this same thinking all the time, “Oh, it’s only 1%,” because that is easier psychologically to manage than to say, “Hm, how can we screw over 10,000+ people without it seeming like a big deal?”

“1% of users” sounds like not a lot. 10,000 actual living breathing human beings with lives that are impacted by whatever product choice you decide to make, starts to sound like a lot. And 1M users for a web service run by a large company is not a lot. ChatGPT supposedly has surpassed 100M users. 1% of that becomes suddenly 1M people.

Don’t get me wrong, product teams have the often unhappy task of having to compromise with reality. You can’t build a product without designing limits, and sometimes (often) you have to test the limits in order to find them.

So I totally get it. But let’s at the same time let go of this thinking of “most users won’t be affected,” (not to pick on this newsletter author quoted above – it’s just a type-example of something I’ve seen run rampant in product thinking) because it only serves to diminish the importance of impacts on actual humans. Okay in a chatbot setting, this might not mean life or death (or it might, depending on context), but in others it very well might.

And yeah, you want your product to be broadly useful, but as a “power user” myself, I tend to want products that don’t say ‘fuck you’ to outliers and edge cases, but that have enough vision and bravery to dig in on what’s driving those avant-garde users, and give them tools to do more, and go even farther.

Related writing & research