Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Signaling in tech is some fucked up shit (2016)

daiyi.co

Post reply on HN
Print view View original

101–110 of 205 posts

Re: Signaling in tech is some fucked up shit (2016)

#101

Post by d0m »

It's just probabilistic.. someone who's really into SICP and clojure is showing signs of being interested in more than your typical minimum 9-5 job requirements. It's not "THE" sign, just one of many. SICP teaches many things, such as how to write compilers, how to think about abstractions or the benefits of immutability. Is it so weird to think that someone's willingness to get better and learn more is more likely to get more job offers?!

Here's one way to look at it, considering everything else equal, if you were an employer, would you pick someone who knows X (where X is jquery or whatever requirements you need) or someone who knows X + Y (where Y is some interesting technology that may eventually turns out to be useful)?

Re: Signaling in tech is some fucked up shit (2016)

#103

Post by randomdata »

SCdF wrote: post #3

> when I told other software people I’m a web developer, I got treated like shit. I was a lower class of coder because I wasn’t “solving cool problems” because making web sites is easy. Of all the stupid arrogant awfulness in tech that has crossed my path this one is definitely high on the list of the most baffling. Web dev is the hardest and most intimidating dev work I've ever done, by a country mile. (Whether that…

The web is 'easy' in the sense that it is well (at least in quantity) documented. If there is something you need to know, it is unlikely that you will have much trouble finding it. Justified or not, I am unsurprised that people with more esoteric knowledge, that is not easily found through a Google search, want to try and boast about it. That's pretty standard human behaviour.

Re: Signaling in tech is some fucked up shit (2016)

#104

Post by throwawaymath »

scarejunba wrote: post #83
throwawaymath wrote: post #37

Earlier quoted context omitted.

That's not a statistic, nor is it quantifiable. I'm not taking issue with your experience, I'm taking issue with the fact that you've implied your claim has quantitative rigor to it, because I'm extremely skeptical that's the case. For what it's worth, I'm challenging you as someone who is neither a JavaScript developer nor a Clojure developer. I'm of the opinion that biases like the one the author laments are perpet…

Don't get hung up on the language. People sometimes say "statistically speaking" to mean "Based on my experience over a large number of tests". It doesn't mean "I applied a test of statistical significance to arrive at my conclusion". You're then meant to combine their stated experience with their stated credentials and your assessment of their trustworthiness to arrive at some conclusion. This imprecision may annoy…

Sure, I understand what you're saying. But I'll continue to challenge the use of language like that. In my opinion, saying something is true "statistically speaking" is more grievous than saying something like, "I'm 99% sure." Both are artifacts of the English language, but I think the former carries much more quantification behind it, implicitly speaking. In that sense it's not just about informal use of language.

This is because saying something is true "statistically" is a claim made in both scientific and casual contexts. It's mentally taxing to figure out which context is in play - someone could have come into this thread with actual statistics and said the same phrase, verbatim. Therefore (as I see it), it's realistically plausible that someone could read the original comment and come away believing recruiters have done actual statistical analysis a la TripleByte to arrive at their conclusion.

Re: Signaling in tech is some fucked up shit (2016)

#105

Post by dang »

machiavelli1024 wrote: post #26
iwangulenko wrote: post #20

As someone running a technical recruitment agency, I can assure you that it is true that someone having built stuff in Clojure is more likely to be considered as a good engineer than someone who did JQuery all the time. Recruiters use an intuitive approach to conditional probability. Given only the info alone (if someone did JQuery or Clojure) exposes you to two bell curves, with the X-axis being "programming ability…

Yep. Funny that HackerNews seem to be unable to wrap their head around the fact that prejudice is just applied Bayesian statistics.

Please don't take HN threads on generic ideological tangents.

https://news.ycombinator.com/newsguidelines.html

Re: Signaling in tech is some fucked up shit (2016)

#106

Post by tmh88j »

EngineerBetter wrote: post #29

Maybe it's my age, and a youth of grunge, metal, hardcore and punk, but whatever happened to doing what you love and not giving a shit what naysayers think?

>but whatever happened to doing what you love and not giving a shit what naysayers think?

I'm in my 30's and have used a lot of different languages and have worked in a lot of different aspects of software development. As long as I'm being compensated at my skill level, I couldn't care less what language I'm using. I'm currently a web dev and the company I work for primarily uses PHP. To make a long story short, a good friend of mine works for a huge university and if I hang out with him during the week, I'm inevitably going to be around some college kids. The typical icebreaker conversations of "what are you studying" often times leads to chats with CS students. The amount of trash talking I get from kids who haven't even graduated and have no professional experience once they find out I primarily use PHP is astounding. Golang and Python seem to be cool amongst college students, so once they find out I also have more experience with those languages than they do with programming in any language, I suddenly become cool and worth their time.

Re: Signaling in tech is some fucked up shit (2016)

#107

Post by troupe »

__blockcipher__ wrote: post #93
wellpast wrote: post #91

Earlier quoted context omitted.

Your question assumes that the programming talent bell curve for men and women are different, which I wouldn’t expect to be the case.

Seems like an interesting assumption. I’m mot convinced there is a difference, but neither am I confident that there isn’t one.

If there was a difference, are you allowed to note it or will you get fired for it? Programming skill is hard to measure. Chess is an interesting area that is a bit easier to quantify.

Re: Signaling in tech is some fucked up shit (2016)

#108

Post by wellpast »

__blockcipher__ wrote: post #93
wellpast wrote: post #91

Earlier quoted context omitted.

Your question assumes that the programming talent bell curve for men and women are different, which I wouldn’t expect to be the case.

Seems like an interesting assumption. I’m mot convinced there is a difference, but neither am I confident that there isn’t one.

Yeah I have no instincts in this because my experience has always been so fewer women than men on the engineering team. So this is totally anecdotal. But the few women I’ve worked with have always been in the upper part of the talent scale, while there has been much variance among the men. I.e., I’ve encountered many low talent (and high talent) men but not one low talent woman. Maybe there are enough hurdles for women (even if the only hurdle is that there are so few _other_ women) that you kinda have to be interested & reasonably good at the job to stick around. If this is true the bell curves may be better for women applicants. However amy anecdotal experience could be simply wrong and it somehow is even or worse for women or something other like overlapping curves of different shapes.

Re: Signaling in tech is some fucked up shit (2016)

#109

Post by Ologn »

> I instantly get elevated to a respect-worthy status for free...Coming upon the Correct Signal by accident

I have worked at companies with open positions, and I have had to sort through a lot of resumes. I also get an informal Bayesian probability heuristic in seeing the resume and then seeing the person who wrote the resume. People who are on the ball, who are the standard deviation above the rest, know what to put on their resumes which looks impressive. People who are a standard deviation below the rest not only don't know how to put together an impressive resume that stands out, they don't even know how to put together a median, average resume.

If I was hiring programmers with little experience, and everyone has the same skillset - Javascript, ES7(8,9...), V8, Javascript framework du jour (React, Angular, Vue, Node.js) - they all look the same. If someone has Clojure on their resume as well, it sends a certain signal out. You still want to talk to that person, and verify that they are a standard deviation above the rest, but it sets you out from the pack. If we don't go by such signals from resumes and the like, I don't know where to go from. The OP is correct though, once you send up a signal you are a standard deviation above, then you have to follow through and you start feeling compelled to be able to produce at a level a standard deviation above.

Re: Signaling in tech is some fucked up shit (2016)

#110

Post by alangpierce »

algeom wrote: post #80
BurningFrog wrote: post #33

Earlier quoted context omitted.

The problem is that you're both right. Putting up a simple web site that does something basic, that you then walk away from is easy, and is what a lot of people who call themselves web developers do. Building something like Gmail or Facebook OTOH are endeavors only slightly below the moon landing in complexity and “solving cool problems”. And it is equally legitimately "web development"...

Can you explain what you mean when you compare Gmail and Facebook to the moon landing? On its face that comment looks way out of whack to someone who is outside the tech industry.

Obviously software was totally different back then, and Apollo 11 was much more than a software project, but FWIW the Apollo guidance computer software was 130416 lines of assembly[1][2], while Facebook is probably at least 100 million lines of code at this point[3]. Also note that Facebook uses languages that are much more expressive than assembly, so "1000x as much code" probably indicates much more than 1000x as much logic being expressed. The reason Facebook and Gmail seem relatively simple is because they're designed to feel simple so they can be used by regular people, but the full implementation details are very, very complex.

[1] https://qz.com/726338/the-code-that-took-america-to-the-moon...

[2] https://github.com/chrislgarry/Apollo-11

[3] https://www.quora.com/How-many-lines-of-code-is-Facebook

Post reply on HN