611–620 of 675 posts
Earlier quoted context omitted.
If enough people are doing it higher iq folks will bubble up in calibration
Which at the end of the day still doesn't test that you can do the job.
Earlier quoted context omitted.
Which at the end of the day still doesn't test that you can do the job.
Ok but it does tho - all other things being equal and you better believe it they do have other type of interviews not just lc
Earlier quoted context omitted.
This is a startup idea in fact, wait, someone is doing that: https://codesignal.com is one that I just learned yesterday, but it's not 'standardized' like what SAT does. In fact I believe software should have some qualification tests, e.g. general coding, database, cloud computing, etc. Like CPA for accountants. Each test should be valid for a few years in each category.
The hard part's not creating some kind of certification, it's getting desirable employers to accept it as a replacement for the most-painful parts of their interview processes. I suspect a lot of top companies don't want to make it easier to jump between them.
Similar to algorithms itself, interviewing is essentially searching for people that the company wants.
How a company interviews candidates define its effectiveness (hire the right person, less false positives, less false negatives) and cost (how much does the company spend on the interview process per hire).
Interviewing with LeetCode is acceptable effective: candidates.filter(leetCode) gives you a much smaller set of people good at algorithm brain teasers, and this set of people have acceptable approximation with the set of ideal candidates.
In other word, it's a lazy but effective enough. The majority of companies will only switch to alternatives when the cost is lower, or substantially more effective. It's broken from the candidates perspective, but not the companies'. Most companies will stick to "if it ain't broke don't fix it" unless we offer them 10x solutions - which is yet to be seen.
But I can also see it the other way around, we can undermine the effectiveness or cost for all companies: if we can develop better courses and bootcamps, letting more people hack LeetCode problems quickly. And then companies will naturally go for harder problems. In the end, the problems will be ridiculously hard that the companies can only filter people who memorizes LeetCode, and those set of people have virtually no correlation of good hires (good problem solving skills). And of course, it's also a profitable business to teach people this.
Earlier quoted context omitted.
The hard part's not creating some kind of certification, it's getting desirable employers to accept it as a replacement for the most-painful parts of their interview processes. I suspect a lot of top companies don't want to make it easier to jump between them.
The "most painful" part of any interview process is just the part that you happen to be the worst at. Personally, I don't find leetcode to be painful at all and would be much more enthusiastic about something that eliminates any part of the interview where you have to talk about yourself.
Earlier quoted context omitted.
> You can however assess how smart somebody is Why not just give candidates a standard IQ test then? They are probably more reliable than Leetcode...
Because people are under the impression that it's illegal. That and I think a lot of people will make its use be about race, so people avoid it.
But if we're saying Leetcode is simply testing for IQ and not really for software engineering ability, why not just test for IQ? IQ tests are designed in a way that after a certain threshold it's pretty hard to improve in them - they actually test some innate ability. Leetcode tests how well you are prepared for Leetcode, and perhaps how well you do under stress. It correlates only slightly with intelligence and even less slightly with programming ability.
I don't think it will stop unless 1) It's illegal. 2) It's ineffective 3) It's not economical for the company. To stop the trend you have to achieve one of the three. Similar to algorithms itself, interviewing is essentially searching for people that the company wants. How a company interviews candidates define its effectiveness (hire the right person, less false positives, less false negatives) and cost (how much do…
Earlier quoted context omitted.
> You can however assess how smart somebody is Why not just give candidates a standard IQ test then? They are probably more reliable than Leetcode...
IQ does not tell determination/hard-working-attitude at all, in fact many with high IQ ended up being nobody, as high IQ makes learning relatively easier, most got used to that and just let the 'grit' go. leetcode is the SAT for coding, not perfect, but at least it's close to fair play.
How is it fair, it favors young graduates and people with lots of time to prepare tremendously. Would it seem sensible to you that every time a doctor applied for a job, even if he has 20 years of experience, he'd need to compete with recent medical school graduates on some first year medical school exam?
I don't think it will stop unless 1) It's illegal. 2) It's ineffective 3) It's not economical for the company. To stop the trend you have to achieve one of the three. Similar to algorithms itself, interviewing is essentially searching for people that the company wants. How a company interviews candidates define its effectiveness (hire the right person, less false positives, less false negatives) and cost (how much do…
Between #2 and #3 you missed the most important thing there: it should be not only ineffective but ineffective enough to be measureablre at the top of C-level. Untill that it is just another pointless KPI alonh many other ones
> Most companies will stick to "if it ain't broke don't fix it" unless we offer them 10x solutions
It should also work if people undermine 1x solution to 0.1x solution. If there are a lot of bad hires at the company level, which they can't finish ordinary tasks other than LeetCode problems and keep screwing things - the team leads and managers will start to complain and eventually propagate to the top level.
Earlier quoted context omitted.
The tests aren't testing for competency at the job, and after a decade of experience writing software you have long ago realized that party tricks and cute algorithms are a fairly rare part of the job (generalizing here of course), so you stop thinking about them as much and get out of practice. When they do show up, you certainly don't have to do them in 10 minutes, and I think everyone would rather you didn't anywa…
>write a robust solution rather than a clever one but why must it be mutually exclusive? Are you implying all "robust" solutions, whatever that means, are dumb? Surely you put some thought in it to make it "robust"?
I think there might just be some vernacular nuance here though, maybe we can call it smart and robust, versus clever and opaque. Some problems are just difficult though, and if you get to work on that kind of problem regularly then that is pretty lucky.