LeetCode is a joke in late 2024 | Hacker News Reader

LeetCode is a joke in late 2024

3 pointsakaTrickster8 comments

Frankly, all of the small questions can be solved by an LLM in 3 seconds, and the difficult ones can be specified in formal english and solved similarly by Claude / ChatGPT40mini in an absurd amount of time. Solving them seems like trivia, why do we even value leetcode still?

View on HN

The problem with leetcode is that it selects people who practiced leetcode and rejects people who did not but would potentially be better engineers. It's the problem of any metrics: people will optimize for the metric.

If you try to answer with an LLM, you'll have to copy-paste the question (or write it), wait for an answer, read the answer, understand it, correct it (because most likely it won't be 100% correct), and answer to your interviewer, all while pretending you are not doing that.

You may as well Google the question and try to find an already-existing answer while pretending you are not doing that, right?

My feeling is that the interviewer will quickly realise what you are doing.

Do you get better as a programmer by solving leetcode problems or coding on things that interests you?

Leetcode was never valuable, except as a "We need a measure. This is a measure. Therefore, we use it." kind of thing.

There is no such thing as an impartial yet meaningful measure that can be equally applied across all candidates. You should be discussing your business, your issues, and the candidate's ideas (with some small practical examples) for how they can move things forward in the area they're applying for.

Someone's leetcode rating isn't going to move your business needle.

Well, their leetcode rating says something about their ability to solve novel abstract problems using computer science tooling. All other things being equal, the candidate better at that is the better candidate.

Meanwhile, it is not a complete assessment. You can do all the other things you've stated and leetcode will still add value.

It depends on how its used. For example, if you have to repeat your performance in a live interview, you'll struggle to do so. Similarly, if the employer uses a confirmatory set which the LLMs get wrong (e.g. adventofcode does that), you will struggle.

The point of leetcode et al is to measure how able you are to solve problems which you have not seen before. It is not a comprehensive evaluation but a useful part of one.

The point is to pick a candidate from a big bunch, even if it's just through narrow trivia that tests recall and speed of thinking. So what if computers can solve leetcode problems? The companies that use leetcode never cared to switch to anything else.

Problem-solving ability is important. You depend on artificial intelligence so much, and I'm afraid that people like you will be replaced soon. In short, if you want a decent job, your skills and knowledge should show the interviewers what makes you outstanding. By stating "I can write flawless prompts and GPT will solve it for me," a new role may be emerging in the market known as a prompts engineer. For most jobs, they are not hiring based on writing prompts.

valued like a museum of pre-23' era:)

Reply on news.ycombinator.com