What kind of work was he doing for them to try to sell to him? Same with the job interview?

> What kind of work was he doing for them to try to sell to him? Same with the job interview?

My understanding is that the oil and gas industry is pretty heavy needs for scientific computation and data visualization. That's how they discover new places to drill.

Oil and gas industries are data crunching monsters. While I thought it was impressive I got to work with < terabyte datasets I’ve met so many data scientists crunching petabyte seismic datasets that could give backblaze a run for their money.

I’m convinced if SV hadn’t happened in SV it would have happened in Dallas or Houston.

Some blips of that alternate timeline:

Schlumberger buying Fairchild, which had FLAIR (Fairchild Laboratory for Artificial Intelligence Research), then opening SPAR (Schlumberger Palo Alto Research) and moving FLAIR into that. Also long forgotten CPU-Architecture called

https://en.wikipedia.org/wiki/Clipper_architecture

later sold to https://en.wikipedia.org/wiki/Intergraph

https://www.slb.com/who-we-are/our-history/1980s

Closing of SPAR, and moving parts of that to Austin.

https://www.latimes.com/archives/la-xpm-1988-10-18-fi-4498-s...

https://en.wikipedia.org/wiki/Jay_Martin_Tenenbaum

Some leftovers, via acquisition of Applicon and Bravo3!VLSI:

https://www.staticfreesoft.com/electricHistory.html

Steven M. Rubin, Computer Aids for VLSI Design https://www.rulabinsky.com/cavd/

> I’m convinced if SV hadn’t happened in SV it would have happened in Dallas or Houston.

Part of what attracted investors and innovators to the bay were the non-compete laws and counter culture of the time.

What year are we talking? I get that is a pretty big use case now, but back then (mid 1980s?) was it also the case?

When I was a kid in the early 90s, a nearby university had a Cray 1. We went on field trip to see it. IIRC, they got it because some oil company had gotten a better supercomputer and didn't need it anymore, so it was probably originally purchased sometime in the 80s or earlier.

Also a quick search found this from 1985: https://journals.sagepub.com/doi/abs/10.1177/003754978504400...:

> Supercomputers are becoming a useful and important tool in the finding and developing of oil and gas reserves. Applications of supercomputers in the petroleum industry involve two important aspects: enormous computational power and massive data management. Vector computers are being used in petroleum engineering to simulate the flow of oil and gas in a reservoir, the faster performance of the vector machines making many, heretofore, unmanageable calculations possible. In exploration for oil and gas, supercomputers are being used to store, classify, and interpret huge amounts of geophysical seismic data.

Funny, I just found that link myself. I guess it makes sense, there was data and there was a need to comb through it. Seems like little to do with graphics directly, and more to do with being able to manage large amounts of data.

"Seems like little to do with graphics directly, and more to do with being able to manage large amounts of data"

And since we are visual beings, managing large amounts of data is best done in a graphical way. But it is not easy to do that right. You can create super beautiful looking, but total missleading visualisations.

My father worked in the sugar industry, his first computer came with an operator :)

When Fairchild got out of the computing business (84/85 ish) it was the huge oilfield services company Schlumberger that bought the AI lab run by Marty Tenenbaum. IIRC S had some Crays and a Connection Machine. Schlumberger Palo Alto was across the street (other side of Foothill Expwy) from PARC. IIRC they had a lot of interesting visualization work; I turned them down (despite their excellent research staff) for PARC because I wanted to work on theory of computation.

Why was oil and gas a target for sales for them? I'm sure the industry has a ton of money, but back then what was the use case?

They do a lot of seismic work to see what's underground (set off explosions, map the results) and need to crunch and visualize a lot of data.

Same reason high-powered graphics boards are used today for cryptocurrency - high throughput calculations.