Right, because everyone in IT knows how to write windows/linux device drivers and all have unlogged admin/root access.

... and the damn things head straight for me completely bypassing my wife so there's probably not as much female bias this gives the impression of. IMO, YMMV.

This is the most "correlation vs causation" quote ever. If you were in any kind of position of authority related to nuclear weapons, it would be Jurassic Park levels of hubris.

It's modded funny because OpenCL is all but dead for new projects. It got weighed down by industry infighting to the point that the big feature of OpenCL 3.0 in 2020 was undoing everything added to the spec after 2011.

So the idea of using OpenCL as a CUDA replacement, rather than something like ROCm or OneAPI, is funny. It's like rewriting C++ programs to use Pascal.

A motor doesnt have anywhere near the rotating mass of a useful kinetic energy storage flywheel. HTH.

Dunno about bikes, but theres little difference in driving feel (not sound) between an ICE car with an auto box and an EV. Unless you think shifting into D is some kind of at one with the machine experience.

I'm pretty sure new riders can learn more throttle = go faster. There was a big fuss in F1 back in the day about getting rid of manual shifters. How many of today's drivers would go back to them now?

Revolutions require a substantial proportion of a group/population to act together at the same time as one. A few people here or there trying it just get bulldozed.

Theyre working for a company founded by a thieving sociopath that treats its users as monetisable assets. Why did they think theyd get special treatment when push came to shove?

And more importantly - their money.

No, it's more about how teams work. Teams have a scope. They don't typically go beyond that scope. So if my team owns the Foo and Bar modules, I work on those. But if there's little important work on Foo and Bar, but a lot of important work to be done on Baz, it's generally organizationally difficult for us to work on Baz. Typically we need to be lent out by our manager and seconded to the other team. Which can be a lot of red tape and politics.

Now if you're imagining some alternate world where programmers an be moved at will- then we're already one big team instead of multiple small teams.

And no, a smaller team doesn't win every time. If it did, then then smallest team possible is teams of 1 and we'd all do that. There are sweet spots, which depend on the organization, the work to be done, and the importance of that work. For some that's bigger, for some smaller. I've definitely worked on teams that were both too small for the work, and that were too big.

They can, under some circumstances. If the scope of what they work on is too small to fill the team's feature set. Or if the work they would be doing is significantly less important than other work to be done, having them in one large team makes it easier to move to more important work and can get critical features built faster. In that case it may not be overall more work done, but it may move the important stuff quicker. If larger teams weren't useful on some level, we wouldn't have teams at all- we'd all be individuals.

They're one of the 4 biggest stock brokers in the US. If you aren't trolling, you're showing yourself to be really ignorant.

In the end- good engineers with sufficient experience and support will get stuff working with any methodology. Bad ones or ones insufficiently supported will fail with any methodology.

There are some things that agile works well for, but it's really limited to domains where you can quickly build something tangible for feedback and you have stakeholders willing and able to give frequent feedback. UIs are a good example. It's a horrible fit for anything that requires actual research, or that can't be shown to low technical knowledge customers frequently (in other words anything that actually needs weeks or months of backend work, algorithm writing, or infrastructure to be written).

The problem with that is the skills needed to manage and the skills needed to do real work (let's take programming as an example) are pretty distinct. Someone can have both, but they tend to have one or the other. Forcing those without the skills to do the practical work into doing it doesn't actually help the team, it just slows everyone down. And if they get on the critical path of any project you can be royally fucked.

There are a couple of ways to solve this problem:

1)Larger team sizes. This can work if the team owns enough to keep everyone busy, but it can lead to effectively being independent subteams calling themselves one team while being inconvenienced by each other.

2)Each manager managing multiple independent teams. This can work if it doesn't overload the manager. The biggest problem is when the manager decides one team is more important and doesn't support the other(s) enough. This works better the closer the teams are, as it requires the manager to know fewer sets of collaborators and politics