On: Forget moonshots, investors want profit now

[waihtis]

As proven by all the AI wrapper startups raising enormous seed rounds

[JumpCrisscross → waihtis]

> all the AI wrapper startups raising enormous seed rounds

Aside from Cohere, don't they mostly have massive and rapidly-growing revenues?

[guappa → JumpCrisscross]

A massive revenue that is dwarfed by massive expenses isn't "profit"

[ein0p → guappa]

Depending on the trajectory of both P&L it could still be worthwhile

[waihtis → ein0p]

So moonshots it is then. I think we just demonstrated the typical investor though process here quite well!

[JumpCrisscross → waihtis]

> moonshots it is then

Moonshots involve fundamental technology risk. LLMs are no longer unproven technology--the problems are in scaling and product-market fit.

[waihtis → JumpCrisscross]

True, but only partially. We really don't know (or at least investors don't understand) what LLMs are really good for. Some of the use cases put out there now will wither away.

[ein0p → waihtis]

Moonshot is when those trajectories are not yet known and won’t be known for quite some time. A good example is nuclear fusion. This is something else: risky but clearly tractable innovation.

[onion2k → JumpCrisscross]

Rapidly-growing revenue that will all get spent on OpenAI API requests once OpenAI shifts its focus from growth to profit.

[selcuka → onion2k]

> once OpenAI shifts its focus from growth to profit.

In the meantime there will be a zillion competitors to OpenAI using open engines for basic tasks, and it will become a commodity product.