I've never seen VCs make a bigger mistake than generative AI.

They are funding many projects that are destined to be abject failures.

Even worse, they are totally missing the places where ML will actually have impact and generate enterprise value.

Let me explain:

To be clear-- innovations in large language models and ML broadly are real. We've made a decade of progress in the last month alone.

I've been running a large language model company for four years now, I am not a skeptic. These things will change the world.

but, the fevered tone mirrors the same mistakes that they made around Web3 or the "Creator" economy.

You will not generate large enterprise value by making everyone a "creator". Giving everyone in the world the ability to generate Picasso paintings is worth very little.

there are hundreds of millions of dollars being deployed towards glorified tech demos built on top of identical datasets. most, if not all of these, will fail.

where generative AI will change the world is in narrow, mostly boring domains. VCs are ignoring this.

Billions of hours of human potential every year are wasted on menial tasks. Data entry, form filling, basic knowledge work kind of stuff.

Large language models are uniquely good at these tasks. These range in scope and impact from fiverr gig work to hours of doctor's days.

"generative AI" will change the world by reducing the tedium of these jobs. giving technology leverage to thousands otherwise stuck doing elevated copy and paste.

there's a few dozen B+ outcomes waiting for whoever builds these narrow domain ML companies first.

giving technology leverage to data manipulation/input jobs is a massive economy unlock.

it turns massive fixed cost center for a business into variable cost with software margins. it also shortens time to usable data massively. its an easy enterprise sale.

so while your favorite creator economy VCs are funding new ways to generate dog pictures on your phone

there are billions of economic value to be generated by repurposing foundational models to solve basic data entry tasks.

ignore the hype cycle, build boring businesses.

TL;DR the greatest impact of "generative ML" will not be on art/creation, but by bringing technology leverage to billions of hours of boring data entry/manipulation jobs.

all the vc energy around generating art will burn astronomic amounts of money and generate zero value.

(this thread was largely written by gpt3)

rip textual seo soylent