Programming

Most performance optimisation can be done through improving algorithms and data structures, and knowing well the language you're using.

I don't think I've encountered any performance issue that wasn't either:

• Unavoidable because of what I'm doing
• Coming from an inefficient data structure or algorithm
• Coming from a bad use or pitfall of the language.

Of course there are specific cases such as low-power environments and such, but that's not what most people talk about when they talk about performance.

I also would like to point out my confusion with the obsession towards performance when it's not needed (such as "is languageA 10% faster than languageB?") but then everyone putting web technologies everywhere, from games to desktop software. It's starting to feel like performance is more of a cult than a pragmatic question at this point.

Performance optimization is hard because it’s fundamentally a brute-force task, and there’s nothing you can do about it.

I think that only takes effect after a certain degree of best practices, knowable practices and effects, and after educated guesses and theories with testing. That's a lot before you reach brute force territory. At least in typical and higher-level programming language development.

Youtube Video. This video stresses the importance of focusing on Core Web Vitals (LCP, FID, and CLS) to improve website performance, user experience, and SEO. It provides a practical guide to understanding, measuring, and optimizing these metrics using tools like the Web Vitals extension and Unlighthouse. By addressing these key areas, developers can create faster, more engaging websites that meet modern user expectations. The content is summarized by Transcriptly

Performance optimization is hard because it’s fundamentally a brute-force task, and there’s nothing you can do about it

There is, "common sense" seems to accelerate that bruteforce work. Also some developers seem to be better at performance bruteforce than others, some enjoy it, others hate it.

Simplicity also plays a very important role here, most software is built by adding more features and at some point you may be able to simplify things a lot and make run a lot faster if you just rewriting it with all the use cases in mind.