Comment by JMTQp8lwXL | Hacker News Reader

Comment by JMTQp8lwXL | Hacker News Reader

Parent Full thread

While I would be personally enthralled to be operating a $35k/mo business, I can only wonder if having software engineering candidates optimize for solving leetcode style problems is a meaningful differentiator for employers, or a good enough proxy for identifying candidates who will succeed in a new role. Tech jobs are too lucrative for our criteria to simply be an open corpus of ~700 memorizable problem/solution pairs. It's going to eventually become ineffective. People will memorize and then flunk on the job.

What the industry is missing is a focus on the human element, which is woefully under-assessed for. Being a good teammate and knowing how to communicate with non-immediate team stakeholders (Product Folks / UX / Quality / Management etc) is crucial.

There is this trend of coding questions getting harder. Around '12 2-sum was a question that is on-site worthy, now being the first of leetcode people expects dynamic-programming type problems in on-sites.

One theory is that the employers (especially big companies) just want the top X% of candidate by some arbitrary metric, and assume that top X% has to be capable enough to pick up new skills as needed. As people practice more, the questions get harder.

Then of course we'd mostly likely to never use the knowledge tested in interviews in real life. But hey, the goal is not to test knowledge, but to differentiate the top X from a pool of a thousand.

> But hey, the goal is not to test knowledge, but to differentiate the top X from a pool of a thousand.

The top X leetcoders. Not the top X most impactful contributors to a company. There could be some overlap, but not complete.

I think there is still value in coding questions, although i agree the value may have diminished. They may be another case of 'as soon as it becomes a metric it's useless as a metric'.

Dynamic programming has always been a common interview problem for the past 15 years.

candidates optimize for solving leetcode style problems is a meaningful differentiator for employers

Isn't it the fault of employers though? They are the ones who ask puzzles, obscure algos etc, isn't it?

Reply on news.ycombinator.com