Because of basic economics. The opportunity size of AI for NVidia is unlike anything we have ever seen. Of course they pivoted.

They didn't need to do any of that by the way.

I'm gonna need someone smarter than me to show me the numbers on that. Fornite by itself is insanely profitable.

That's like saying playing baseball must be profitable because of how much money A-Rod made. The returns are skewed.

A game can be massively popular but many many games fail to hit the mark. Many do not see success and many do not even ship.

Ok, but Fortnite is a massively popular success, even as its popularity slips. Fortnite's run so far could have sustained Epic for years, even without other revenue they get from things like Unreal Engine. Games as a whole may be a risky venture, but we're talking about Epic here; the mystery is not how to succeed in games, but how a company that had an earth-shattering run of success in games is now in such a position.

Just because it's popular, doesn't mean it's financially successful. Take a look at YouTube. They lost money hand over fist for decades.

This. You need to fire your CFO immediately if you don't have billions of dollars in cash after the run you just had on Fortnite.

They should've setup an endowment fund, could've been self sustainable by now.

The solution is to have the quote manually prepared, entered into the system of record, and then automate the outbound phone call to let the customer know, and agree to the work based on the accurate quote. That's the savings.

It seems to me that calling the customer to get them to agree on a quote is the most important contact?

It’s like telling a salesman to just enter data into a CRM and trust their livelihood to an AI closer. See how that goes over.

That call is really important. “Do I really need to spend all this right now? Can I just get by with x instead of x,y,z?”

For now I’m not sure it will be efficient to provide models enough context, for one thing, to do it reliably.

So you've saved 30 seconds per customer for a total of 3-10 customers calling per day.

Why? Just put it all in the context window.

Yeah, just stuff the entire website and pricing table into the context window.

Documentation rarely reflects how anything is actually done, referred to by good business analysts as 'shadow functions'.

LLMs are already good enough to read corporate email and document shadow functions and hierarchies.

Corporate email documents even less.

No, current LLMs are already good enough to read the subtexts from documents, email, call transcripts where available. They're extremely good at identifying unwritten business practices, relationships, data flows, etc.

As a email client.

It was. I did some I.T. consulting in the Fortune 500 and a few ran Notes on the AS/400.