alt Hacker NewsI'll say invest totally in domain knowledge now. The value of knowing how to invert a binary tree from memory has dropped to approximately zero. Web development as we knew it for the past 20 years is completely dead as an entry level trade. The power is shifting to people with useful knowledge and expertise that isn't about twiddling bits.
Replies
What do you mean by “domain knowledge”? And how is it a competitive advantage?
I think this is true today, especially with complex domains, but I foresee a future where more and more walls fall. If you are in college now, go deep on a domain. If you are entering in 10 years, I have no idea.
The fact you are getting downvoted to oblivion shows how fucked HN has become.
Ain't nobody gonna hire a code monkey - you are being hired based on whether or not you can reason and enable workflows via tech.
If you're only name to grace is you can write pretty Python but cannot architect at scale or care to actually understand the bigger picture of what is being built and why, you will get offshored to someone who is also using Claude Code.
If I'm working on a fullstack for a cloud security product like Wiz, I'd rather hire an average developer who deeply understands the cloud security industry versus a NodeJS doc wiz who has zero empathy or interest in learning about cloud security. There are too many of the latter and not enough of the former in the American scene now, and especially on HN.
If HNers cry about how cut-throat the American market has become, they haven't seen it in China, India, or the CEE.
Are people still under the impression that testing candidates with coding challenges is in preparation of a job where real world problems are described like "invert the binary tree"?
There was never any value in simply the ability to invert a binary tree from memory. First, contrary to popular belief, this particular challenge is quite trivial, even easier imo than fizzbuzz. The value of testing candidates with easy problems is their usefulness in quickly filtering out potentially problematic coders, not necessarily to identify strong ones.
Second, another common take on coding challenges is that they're about memorization. Somewhat, but only to a point. Data structures and algorithms are a vocabulary. A big part of the challenge of using them "creatively" in real life is your ability to recognize that a particular subset of that vocabulary best matches a particular situation. In many novel contexts an LLM might be able to help you with implementation once the right algorithm has been identified, but only after you yourself have made that insightful connection.
Having said this I generally agree with the philosophy [0] that keeping things simple is enough 95+% of the time.
[0] https://news.ycombinator.com/item?id=47423647