A $182B Chip Maker: AMD's Labs – Full Documentary [video] (youtube.com)

The older I get (I’m not even 30 but nevertheless), the less invested I feel in so many of the problems in software engineering. It’s why I’ve gravitated toward embedded and computer architecture, the problems there feel more fundamental and necessary. I have a friend who jumped ship recently from the CS department to physics, and he tells me that spending so much time thinking about the theoretical underpinnings of the physical world just sort of ruined CS and software engineering for him. I imagine that physics can make almost anything else feel trivial, so maybe it’s an unfair example.

I know there are interesting problems in CS and software engineering but rarely do those of us working in industry get the chance to do anything truly novel. Embedded is simple, almost like using hand tools, but at least the products I contribute to are physical and solving real, hard problems. Just wish the money was better…

There are hard deadlines and many sleepless nights in semiconductor industry. After a decade I want to get out. There is no magic compensation or any other magic. Just customers waiting for your chip working according spec on promised date. The time pressure can be blurred working in a nice team with nice manager. But looking back it was all the shitshow. Better endless discussions about philosophy and the ability to recompile everything last minute or let the customer use old version for one week more than yearly world end cycle with hysterical group leader.

i'm not kidding, i believe Steve (tech jesus) has a bulldozer in his personal rig.

If you mean the people working at the site, probably something boring yet practical. I know someone who tests new, unreleased parts for a living and I once teased him if he's running on something new and fancy, and if I could have some too. His reply was that the market is competitive enough that they never sit on working parts, if it wasn't horribly broken it would already be for sale, and as he spends all day messing with broken hardware, he has no particular desire to do that once he returns home.

It's how journalists originally came about. You started writing about something you were interested in, and enough people wanted to read it that you could make a career of it.

It was large corporations that ruined journalism, and now those corporations are slowly dying.

They have also productized integrity, which is amazing to see (and worth supporting).

There are a lot of video game / hardware reviewers out there, but GamersNexus to me is the best. Enough that I have bought their merch - their coasters are so cool!

Ridiculously hard. It will be three years minimum for the first chip, and five until the second one (first is never good enough). Capital requirements run in the hundreds of millions. Meanwhile, you’re building the team from scratch, and you can’t really afford to “learn from your mistakes” because you only get two or three shots, so the team basically has to be world-class within eighteen months.

Depends on what you mean. If you want to be a fab, the answer is "hahahahahahaha how many billion dollars do you have?"

If you "just" want to send custom designs to an existing fab that's far easier.