I am calling it: AI agents are at the peak of inflated expectations

I am calling it: AI agents are at the peak of inflated expectations

My evidence: Bret Taylor’s Sierra.ai raised a $110m round

Of course, I believe AI agents and copilots are a HUGE category. Heck, I’m building one (heysam.ai). But, this round implies a crazy expected return

First, my justification for this round:

it’s the price investors paid for Bret, not for Sierra.

And here’s my bear case:

OpenAI already lets you create custom GPTs. It’s likely gonna integrate with enterprise data sources: Box, MSFT, Zendesk, Salesforce etc. soon making it super easy for enterprises to create agents themselves

How can Sierra win against OpenAI? I don’t think the moat is the LLM or data; the moat is the UX they create to solve specific enterprise problems. As a horizontal platform, OpenAI will likely not build great problem specific UX.

But, even a small startup can copy your UX. How can Sierra win against small startups built on OpenAI with the same UX? Well, they have the distribution advantage of Bret Taylor.

But, Zendesk, Intercom, & other big incumbents have a bigger distribution advantage than Sierra. How can Sierra win against those incumbents?

You see my point? The market is big and Sierra will be big, but there’s a very real ceiling to Sierra’s success. And that makes me question the implied expected return of $110m round.

Having said all of that, would any investor say no to the opportunity to back Bret? Nope.