Almost certainly not at the scale of the consumer gaming industry, however!

Google is making millions of TPUs per year. Nvidia ships more gaming GPUs, but it's not like multiple orders of magnitude off.

I'm willing to bet TPUs wouldn't be nearly as successful or sophisticated without the decades of GPU design and manufacturing that came before them.

Current manufacturing numbers are a small part of the story of the overall lineage.

It's pretty interesting that consumer GPUs started to really be a thing in the early 90s and the first Bitcoin GPU miner was around 2011. That's only 20 years. That caused a GPU and asic gold rush. The major breakthroughs around LLMs started to snowball in the academic scene right around that time. It's been a crazy and relatively quick ride in the grand scheme of things. Even this silicone shortage will pass and we'll look back on this time as quaint.

You are missing his point. They very likely don't start building TPUs if there were no GPUs.

I'm not missing the point. If you recall your computer architecture class there are many vector processing architectures out there. Long before there was nvidia the world's largest and most expensive computers were vector processors. It's inaccurate to say "gaming built SIMD".

We don't need to know the exact boundaries of what's acceptable to recognize obviously harmful behavior and make efforts to stop it on a societal level.

This is the classic "perfect is the enemy of the good" type scenario.

Let's make imperfect progress if that is what we're currently capable of.

The drug so popular no one thinks of it as a drug any more.

Criticism without a better solution is only so valuable.

How would you do this instead, and why?

I went through the exact same thing as you, and I needed some time to explore different ways of living. I tried being a drone pilot, a kayak guide, and a paddleboard instructor and learned a lot in the process.

After those forays I designed and built a trailered coffee from scratch and now I run it on a public park that overlooks the ocean.

I am more fulfilled than ever, I can pay my bills, and I get to do WAY more "real" engineering than the bureaucracy of my past life at FAANG ever allowed for.

Thanks for your positive story. How were your finances before you got off the rat race? Were you quite comfortable for a risky situation or you just risked it and it worked out?

I was in an extremely good spot financially when I quit (years of runway) due to living well below my paycheck for years of working at FAANG.

I was terrified to do so, but the only alternative I could see at the time was killing myself if I had to continue as is (lots of mental illness prior to and greatly exacerbated by my time in tech).

It is the biggest leap of faith I have ever taken, and I do not have a big appetite for financial risk.

Corded yard tools will always suck because they can only pull <15A on a standard US outlet. They just don't have the juice.

It's always better to go with batteries for electric outdoor stuff for that (and other) reasons.

What kind of batteries are you running that can support the equivalent of 15A AC for hours?

I have both a battery leafblower and a corded one. The corded is far more powerful and of course does not run out. The battery one is quick and convenient for small cleanups but only gets about 10 minutes from a full charge, then it's back to the charger for hours.

Recently I cleaned up a large roof full of leaves, took about an hour with the powerful corded leafblower. That would've taken weeks with the battery one given the small power and the ~10 minute runtime.

I mostly use the battery one since it's easier and most jobs I do are tiny. But it is no substitute for the corded one.

You cycle multiple packs to run for hours.

Your battery blower sounds like it's just not very good.

I can move piles of wet leaves easily with my makita blower that uses two 18v batteries. It's a pretty old model too.

Batteries can surge power and not risk fire hazard like AC over a long extension cord. Manufacturers know this and have to intentionally limit draw way below the 15A ceiling so a 100ft 14AWG cord doesn't trip breakers or burn houses down.

You don't normally need that much power continuously, you need surge capacity.

It seems very likely to me that the sensations experienced during exercise are highly variable among individuals.

I say this because my experience is very different from yours: I get a very perceptible "high" once I get into the rhythm of a good workout. Think mild euphoria, mood lift, and general feeling of "rightness" in my body once it's been well wrung.

This only happens if I'm in decent shape, though. If I've fallen out of shape it's a slog.

Edit: I can't remember the podcast, but I recall some discussion of emerging clinical evidence in exercise response variability along many dimensions that may help explain the disconnect.

A friend told me he was addicted to running because he literally got high from it. I said, running hurt for me. He said it used to for him too. I asked how long until it stopped hurting. He said 2 YEARS!!!!

There's no way I'm going to run for 2 years on the hope that one day it will stop hurting and get enjoyable.

I've done weight training for about 5 years now. It's not fun. I definitely get no "high" from it. But I do like the results.

There is no universally "correct" granularity.

You could easily scoff the same way about some number of API endpoints, class methods, config options, etc, and it still wouldn't be meaningful without context.

It's ok to split or lump as the team sees fit.

> There is no universally "correct" granularity.

There may not be a universally correct granularity, but that doesn't mean clearly incorrect ones don't exist. 50+ services is almost always too many, except for orgs with hundreds or thousands of engineers.

Credentials should only be provided at the application root, which is going to be a different root for a test harness.

Mockito shouldn't change whether or not this is possible; the code shouldn't have the prod creds (or any external resource references) hard coded in the compiled bytecode.

I totally agree, I’m being tongue in cheek, but given how poor some codebases can be, the more precautions the better ie compilation failures on non-mocked functions.

It's useful to get "glue" code out of the way while building, but to the point in the article it all becomes very difficult to debug and maintain once there are problems in the that layer.

Spring Boot and other similar frameworks come to mind; by forcing huge amounts of indirection you lose a lot of visibility of your call stack because the convenient "glue" code is now orchestrating everything at runtime, but that code isn't yours, and it isn't easily inspected or fixed.