Who the biased algorithms are written by are pretty irrelevant. The argument is they are biased against anyone who is not a white male, and that we should be aware of that.
The article has a hypothesis right in the headline: "algorithmic bias is a result of the insular, mostly white tech industry". So no, if it was written by an entirely Indian staff at an Indian outsourcing company, that would discredit the idea.
Canadians are criminally underpaid. I made significantly more than many full-time employees (I'd wager the majority) on more than half of my internships. Adjusted for rent and cost of living.
I do love London, Ontario though. Beautiful place even if it's too homogenous and quiet outside of around Western university.
Claiming racism and sexism will put off anyone in the center who may be convinced that there is a diversity problem. I would prefer the term "subconsciously biased against". It's not so much a concentrated group effort by everyone who hires to keep minorities down, but a subconscious effect caused by the media and constant anti-minority propaganda (only a few decades ago). Even minorities themselves are influenced by this (see: confidence issues).
It's not conducive to growth as a society or as individuals when you attack people who may otherwise be convinced to help you. The ones that are just blatantly racist are fair game (imo), though.
I mostly agree, but the media (and a slew of other industries) are pro-minority, not anti-minority, so if the media are undermining minority confidence, it's not necessarily intuitive. Further, even the implicit bias theory has a lot working against it; chiefly, it is utterly dependant on the principle tool for measuring implicit bias, the IAT, which fails to actually predict bias or even return consistent results (in short, it fails basic psychology tool standards). Implicit or explicit bias may well be the cause of these disparities, but we shouldn't punish skeptics as severely as we have been doing given the dearth of supporting evidence.
I can't see why you would be skeptical about bias given studies (not relying on IAT) showing that -- all else being equal -- people with minority names on their resume get significantly less callbacks? The same goes with subtle class cues and identification. If IAT is broken, it's a broken tool. Not every bias test uses it.
The idea that the media is pro-minority when we still have the media pissing themselves over kneeling in protest is silly. Also when Fox News is still watched by a significant portion of Americans. Also when the media constantly shows minorities in negative situations (yes, it's reporting, and information, but that creates biases as well). I mean it's easier to internalize a minority doing a violent crime as "terrible" than it is a white male embezzling funds or something. In addition, a few decades ago the story was not the same. The people who grew up then have developed a bias due to the media and the people who are raised by the people who grew up then will also internalize some of it. Discrimination is not something solved in a generation.
EDIT: Edited my post a bit because I completely misread your comment.
> Edited my post a bit because I completely misread your comment.
No worries. Sorry if I was hard to understand.
> people with minority names on their resume get significantly less callbacks?
Assuming this is a consistently-reproducible finding (I'm aware of at least one study that finds no disparity), how far can we extrapolate this finding? On this basis alone, can we conclude that pay is lower for employed minorities than otherwise equivalent whites/males, or that workplaces are more hostile? Is it sufficient to change policy? Can we extrapolate this to tech specifically? Should we fire people who openly question our extrapolations?
> Not every bias test uses it.
No, but most do, and it's considered to be the best bias test in social psychology (a very low bar, due to the difficult and complex nature of divining motives). If you throw away the racial-bias research predicated on the IAT and other unreliable tests, what remains is largely anecdote.
> The idea that the media is pro-minority when we still have the media pissing themselves over kneeling in protest is silly.
I don't see why; rightly or wrongly, most of the media is sympathetic to the BLM side of the issue. Media reporting is widely considered (by the left and right) to be socially left-leaning, and it's self-evident that the media goes to great lengths to show minorities in the best-possible light. Of course some outlets are exceptions--like Fox News--but exceptions don't disprove the rule. Even if the media are horrible racists (90% of journalists are Democrats, so we should expect them to favor a pro-minority narrative, but even assuming this isn't the case...), they are compelled by self-interest to at least appear pro-minority (in which case, they are still pro-minority for all relevant purposes).
With sincere respect, if you really think the media is anti-minority, I don't see how you and I can have a reasonable, agreeable, productive conversation about bias in tech, and maybe we should just let this thread die.
EDIT: Some edits made to improve clarity/organization.
> I can't see why you would be skeptical about bias given studies (not relying on IAT) showing that -- all else being equal -- people with minority names on their resume get significantly less callbacks
Studies like that generally tend not to replicate, including at least one specifically about race and names on resumes.
> The oddest turn is that of blaming the global rise in income inequality on SV, and then to lambaste SV for trying to come up with solutions to ameliorate that problem as it proceeds to its logical extreme.
People don't like new people getting money. A story as old as time.
I was interested if the parent moved and maintained his (adjusted) total comp and was happy with his decision, or was happy despite a loss in total comp. Obviously there are good companies and bad companies everywhere. :)
I had a loss in total comp, however, my wife and I are not your typical consumerist-minded people, we like to save and invest. "Make half a million, but live like we make 100k" has kind of been our motto. The move to PHX was a drop in total comp, a net increase in the amount we could save/invest given the COL, and a much larger increase in total happiness.
Perhaps as people who call themselves "engineers" we should move away from the idea that "gets the job done" is good enough? Especially when we have security breaches left right and center.
There is no obvious connection between the methodologies peddled by "Uncle Bob" or the author of the piece and security breaches. They are not claiming to be computer security experts. How exactly are unit tests or code reviews or similar rituals supposed to have prevented the numerous breaches by Julian Assange and his WikiLeaks associates, especially when most of the information appears to have come from insiders with authorization to access the data? (for example). In the Bradley/Chelsea Manning case the government has consistently implied that Manning had authorization to access the vast State department archive of diplomatic cables, as implausible at that seems. I don't see how unit tests or code reviews would have prevented John Podesta from clicking on an obvious phishing link. :-)
> Some people should not be in this industry -- not because they don't have the chops; but because they don't actually want to put in the time and effort to develop their skills. And it is an overwhelming task.
In what world is it an overwhelming task? Maybe I've landed only at amazing companies but nobody is working the full 8 hour workday. There is time to relax and read random slightly interesting technical things as a break. There are pushes to use slightly interesting technical things in your projects (regardless of how useful it is -- because we need to keep our resumes updated!) and most of the time management will allow it.
The hyperbole from software engineers on how hard this piss easy "profession" is, is just FUD and only serves to keep away people with low confidence issues. In other words, what you are doing is not only keeping potential candidates away, but greatly keeping underprivileged candidates away, and it'd be great if you stopped.
Software engineering given it's current state as the wild west is not a difficult profession in the slightest. It's not law, it's not finance, it's not medicine. We work (relatively) short hours for great benefits and have the ability to work from wherever we want on certain days. Your grades don't matter and you don't need to spend days learning anything other than what's tested in an interview setting or used in production.
Passion is overrated and stupid and it mostly comes from privileged people who had the time to fuck around on the computer all day (and execs wanting to exploit the workforce a bit more) -- not realizing that many people can do what they do given time and mentorship.
There are few other jobs that give you more bang for your buck in terms of effort if you're okay with ~200k/yr being your cap.
You might have had good luck with other engineers, I have definitely encountered people who could not seem to get their head around even basic programming concepts. It comes easy to some given the appropriate time/effort, but I have definitely encountered people who did not seem able to pick it up.
It is also a difficult field given that it is one of the few where you probably will have to drop a huge, comfortable skill set every 2-4 years and start over again with a new codebase, different frameworks, and different languages. That does not happen in finance, medicine and law. If it seems easy to you, try jumping into a well established but poorly written project, built using libraries/frameworks you are not super familiar with ;).
I definitely agree with the hours thing though, if you establish yourself at a company you can pull off some pretty amazing work/life balance things.
You are not starting over every 2-4 years. A lot of what you learned is reusable, those frameworks resemble each other great deal. Even with changing languages, it is hard only when you are changing paradigm (procedural to object to functional). Things like syntax take less effort to pick up and re-learn. Algorithms, debugging and structuring of code and similar meta skills remain. Moreover, you dont have to chase every fad. They come and go and it is perfectly ok to skip some of them. Focus on things that you need or that promise longevity. Nobody cares now that I skipped COM+ years ago, nobody cares that I skipped browser differences or ruby just a few years ago and nobody will care that I am ignoring angular and python now.
I did jumped into established projects using libraries/frameworks I was not at all familiar with and liked it. Thankfully companies here are willing to hire like that. It is period when they literally pay you for learning on the job. That is awesome, it is hard to ask for more.
I think it's an issue with teaching. I am teaching a few people in my spare time and the way highschool (& sometimes university) teaches programming is not conducive to learning or understanding. I've still yet to meet a kid who wants to learn that can't learn it given reasonable time and proper explanation. Given the Internet likes explaining basic concepts in forty different ways, I think as time goes on learning CS only becomes easier.
With good fundamentals I think picking up new frameworks and different languages is fairly easy. With the exception of Rust and Haskell, which I'm interested in but never got around to using well, the other languages all seem to come fairly easy after a month or two. There will always be idioms you do not know or fail to remember but that's why there are linters, code reviews, and senior engineers.
It's definitely daunting to jump into poorly written products, and I feel every company has at least a section of their code base that is like this, but that's what good on-boarding and ramping up is for. These are all process issues and not fundamental issues with computer science.
Very few companies I've been at are good at one of these things let alone all of these things. A lot of the time you _are_ slogging at it alone and sometimes you go alone for too long because of ego or lack of confidence. These are all organizational and process issues. Some of it is individual issues too that one must overcome, but making it seem like only the passionate succeed in software engineering like everyone is the top 0.01% of their craft is harmful to the profession as a whole.
My solution to most of these problems is to find a mentor at your company. Anyone who knows what they're doing. I find most people are super nice when you ask questions as long as you are willing to learn and aren't just trying to get them to do something that seems annoying (i.e. asking ops people why your build fails...). I've never met an engineer who _doesn't_ want to talk about how something is designed, the pitfalls, and the hacky work arounds. People like being useful to other people. Let them be useful.
Caveat: I've met a few "legacy" engineers who were hired early and have a negative influence on the direction of the architecture. In companies like these? You can just leave. Everyone wants a good software engineer and as long as you're willing to relocate, you'll land somewhere fine.
Maybe you're asking the wrong questions or are from an undesirable area. I mean the example question you're using is just trivia and not difficult to learn when a problem arises. I'm not sure how much more you want compared to "local variables, return addresses, parameters" tbh.
