Hacker News Reader

Comment by fnbr | Hacker News Reader

What's missing here is a discussion of compensation. I've seen at conferences how proud senior Google employees (ie Jeff Dean) are of the TPU work that their teams are doing, and I suspect that compensation is following accordingly. I doubt that Intel or similar companies are compensating their people at the same level. Whenever I see people talking about a talent war, I'm always suspicious that they're getting outspent and are unwilling to compete.

View on HN

At my university, the word on the street was to not waste your time with chip design or device physics, because most of that work was heavily outsourced and the employers didn't pay well. I thought that was a real shame, because our device physics professor was brilliant and his research sounded fascinating.

Sadly, the people I met from a very popular microchip manufacturer (they don't make CPUs though?) suggested that they had to work long hours and didn't like their jobs. They were actually.. surprisingly transparent about that. I appreciated it, of course!

I worked at a facility with a lot of these guys. They seems to work long hours, got laid off a lot, and the cars in the lot were not fancy at all.

The managers looked well rested and well paid though.

Also if you do make a device physics breakthrough then people will promptly ignore you. Doubly so if you're young.

Last I heard, you had big claims and have yet to test your ideas in physical hardware. People will probably be less likely to ignore you if you have something that you can demonstrate being able to manufacture reliably and that you aren't just exploiting some simulation artifact.

It's relatively hard to fool TCAD. SPICE simulations at least I can believe there can be scoping artifacts, but even those are generally in a certain pattern. In any case, I'd be okay with a "this is great if it works in hardware, but until then idc", but the response is more "you're a young idiot who shouldn't bother with memory cell innovation". A lot of this is superstition due to SRAM being finicky. I could go on about this, but let's take it offline in that case.

Reply on news.ycombinator.com