We got along better before all these managers arrived, could ship huge projects with 100 people and one ceo. And the ceo wrote code and didn't do any of the nonsense most managers fill their days with.
Then when you need to buy a house you have to leave.
Net migration to London from the rest of England and Wales is only positive for those aged 20 to 27
There's plenty of great experienced people outside London, but all the rich (business owners) are openly prejudice to everywhere else. https://www.bbc.com/news/videos/cwy57v07gn2o
The issue is actually that there aren't, let's say I go to derby, how many ML engineers will I find, or infrastructure coders, Biomed researchers or industrial chemists.
This sounds far more real to me than the original post. All the technical issues in the world don't bother me unduly, it really is the managers who make you hate work.
Money wise these corps are a system of their own, they pay enough to make you not quit. The more they pay, generally the more they suck.
Just need to wait till my 401k doubles one more time, my kids finish college, and the house is paid off.... just 10 more years
I believe the conclusion is correct in 2025, but the article in a way just perpetuates the 'misinformation', making it seem like finding if your code will compile to a dynamic branch or not is easier than it is.
The unfortunate truth with shaders is that they are compiled by the users machine at the point of use. So compiling it on just your machine isn't nearly good enough. NVIDIA pricing means large numbers of customers are running 10 year old hardware. Depending on target market you might even want the code to run on 10 year old integrated graphics.
Does 10 year old integrated graphics across the range of drivers people actually have running prefer conditional moves over more arithmetic ops.. probably, but I would want to keep both versions around and test on real user hardware if this shader was used a lot.
"it will almost certainly double in value in 7 years" implies 10% is an almost certain long term return. It is not.
10 years is a common rule of thumb to double your investment, this is a good guess since after adjusting for inflation, the real return of S&P 500 is about 6.8%. Which is about 10.5 years to double your investment.
(give or take S&P 500 being one of the best returns you will find historically)
I see an over-reliance on automated tests recently. Often suggesting just passing the CI tests is enough to approve a change. In ancient code it's just as important to limit the blast radius of your changes, and have a good reason for making them. Not changing something is the ultimate way to prevent a regression.
It's also important to realize claim denial isn't the same for all customers. When I was classed as an executive at a large blue chip company, my claims were approved quickly and easily at the company specific hotline. With the same provider, making the same claims, they are now partially denied initially and it requires long phone calls to get them to admit they should be covered and often it just isn't worth the hassle.
That game already exists. Intel and Amd had an odd relationship for years when Intel needed Amd so they could claim that competition existed. Cross licensing instruction sets and the like.
