|
DACK FAYDEN posted:As an outsider, do those exist? As someone on a H1B visa, yes.
|
#
¿
Jun 17, 2016 06:53
|
|
|
|
| # ¿ May 10, 2024 22:26 |
|
|
Analytic Engine posted:Are those numbers only base salary? They seem low for big companies. Yeah, base salary only -- but you can filter by both year and location, which makes a big difference. But it's not like you're going to find accurate numbers on equity anywhere, especially for senior devs, including Glassdoor.
|
|
#
¿
Jun 17, 2016 19:35
|
|
|
Analytic Engine posted:This is a helpful addendum to the H1B and Glassdoor salary chat because it includes equity and bonuses for big companies: There are some clear data quality issues there. Senior SWEs at Microsoft don't get that little base pay, and E5s at FB don't get that little equity, for two.
|
|
#
¿
Jun 20, 2016 02:42
|
|
|
Arsenic Lupin posted:Given that Uber knows, via GPS, exactly where his phone is at all times he's got the app open, this would be an easily falsifiable lie, and a competent lawyer would advise against it. We probably shouldn't rule this out, since mysteriously he is claiming the exact same round $200 in 'employment related expenses' every week. e: Notably, only data from 3 weeks is being provided as well -- two during the holiday season, despite their claim that the driver has continued to drive for Uber for nearly 10 months (Sept. 2015 to July 2016 when the lawsuit was filed). It's a virtual certainty that those weeks are being cherry-picked. blah_blah fucked around with this message at 21:17 on Sep 5, 2016 |
|
#
¿
Sep 5, 2016 21:05
|
|
|
Arsenic Lupin posted:I am not a lawyer. Assuming he wins on the "I'm an employee, not a contractor" point, it would surprise me a great deal if the minimum-wage laws were written to say "You can pay somebody less than minimum wage one week, as long as you pay more than minimum wage another week." That would gut a lot of the purpose of minimum wage. I dunno about the expenses; they've presumably been averaged, as you imply. There would be fixed costs for car payments, depreciation, and insurance. But the comparison is fundamentally silly. The flat $200 expenses per period independent of hours worked mean that he could similarly work 2 hours during a week, make no money, and claim a loss of $100 per hour. The plaintiff here is calculating their time worked to (roughly) the nearest 30 seconds, their income to the nearest cent, and their expenses to the nearest hundred dollars (and not even for the same time period in question). Oh, and they are also only giving you data for 3 weeks out of roughly 40 that they've worked. That's probably not something you should be taking seriously. Lead out in cuffs posted:The article says that Uber reimburses those, so maybe that's a flat rate for the reimbursement? The calculation of $/hour is clearly (total compensation - expenses)/(hours worked), if you want to double-check.
|
|
#
¿
Sep 5, 2016 23:02
|
|
|
archangelwar posted:Edit: and of course they were cherry picked, but the claim states the average for the entire course of employment was $6/hour. As before, using mysteriously round numbers.
|
|
#
¿
Sep 6, 2016 02:53
|
|
|
archangelwar posted:As expenses are going to be estimates unless every expense is meticulous (and given that not every automotive expense is 100% applied and would thus be an estimated reduction in overall expense), rounding is common and rounding to round numerals or 5 increments is even more common. Unless his estimation deviates from reality by a ridiculous amount, it would have only little impact in the ultimate average which was still significantly below minimum wage. And at this point you are just assuming everything is a lie. $200/week in deductible expenses to run a taxi service does not sound unreasonable at all. While it is certainly possible that he has massively inflated the numbers, do you have anything at all to offer as a compelling reason why I should distrust the number other than the idea that the roundness upsets you? Presumably he should know the exact amount spent on gas/insurance/repairs (the three examples given in the document) over that timeframe, for starters. Expenses are one of the absolute most relevant things here (since no one is seriously arguing that Uber drivers make less than minimum wage before expenses), and really not all that difficult to account for. So it's difficult for me to take it seriously without a significantly more detailed financial picture than is given in the filing. It's also notable that the low wages/hour seem to arise largely from claims to be working ~60 hours/week (nearly 80 in one of the weeks where more detail is given). It would be interesting to know how that is being accounted for. Note: I don't have a hard time at all believing that Uber can pay significantly below (effective) minimum wage, especially if you take out a lease to buy your car (and it's super scummy that they encourage this), if your hours are primarily off-peak or in low-traffic areas, or if your car is low-efficiency/non-electric or expensive.
|
|
#
¿
Sep 6, 2016 03:57
|
|
|
archangelwar posted:and depreciation is inherently estimation, Vehicles have a cost, a mileage lifespan, and you know roughly how many miles you drive for Uber. It might be 'estimation' but it's certainly not rocket science, nor would it be that inaccurate. Similarly for gas costs (assuming you aren't driving e.g. a Prius, as seemingly half of Uber drivers are doing nowadays). Insurance is a little bit trickier -- but in many cases the nominal cost (but not the expected one) is just going to be zero due to people driving under their own personal insurance. archangelwar posted:If your assumption is that he is flat lying by significant amounts just state that, stop trying to justify it, as you are offering nothing other than wild assumptions to support your statements. Clearly I think that the expenses number is inflated, and the number of hours worked likely is as well. These estimates (including depreciation, which isn't mentioned in that part of the filing) still put Uber pay after expenses as 50-115% more than the $6/hour figure quoted there. You seem to have uncritically accepted that $200/week is about correct for expenses. Note that using the assumptions in the Buzzfeed article above, 250 working days/year at 75 miles/day (probably above the median, but in the ballpark of correct c.f. this link) comes out to around 5.5k of expenses per year, which makes roughly a $2/hour difference. Note that several of these expenses are also tax-deductible as well, making the true expense even lower -- for example, taking the standard mileage deduction alone reduces your taxable income by roughly $10k in the above, which saves $1.5-2.5k depending on the tax bracket of the particular individual.
|
|
#
¿
Sep 6, 2016 05:38
|
|
|
I think the thing that mystifies (or maybe impresses, in a perverse way) me the most is that Theranos had a large number of smart, highly-paid, reasonably intelligent people working for them, and it doesn't seem that they ever pieced together that the product they were working on every day was complete snake oil. Obviously their culture of paranoia and compartmentalization contributed to this, but still you'd think that something like that would have been hard to conceal.
|
|
#
¿
Sep 6, 2016 22:08
|
|
|
silence_kit posted:Are you an idiot? Do you think that the academic publishing companies charge exorbitant fees because it is expensive to host .pdfs on a website and have volunteers do all of the work? There's no way in hell that academic publishing companies are operating anywhere close to at cost. Yes, he's an idiot. The overhead of academic journals is massively overstated and the quality of people who generate said overhead is shocking low. The value added in math/CS/physics by journals + arXiv versus arXiv alone is close to zero, and the actual cost of the value added (to academic institutions, coming from essentially volunteer work) is probably in the neighborhood of thousands of $ per person-hour worked. It's a massive inefficiency.
|
|
#
¿
Sep 11, 2016 06:11
|
|
|
Slanderer posted:You have clearly never looked at an open-access journal, because otherwise you wouldn't say something so loving stupid There are some high-quality purely electronic, open access journals in math, like EJC and EJP. Annals of Mathematics is the single best mathematical journal and was open-access until 2008. More recently people have been working on open-access journals which are basically just arXiv overlays. The best example of this is Timothy Gowers' Discrete Analysis, a journal with an extremely high-quality editorial board.
|
|
#
¿
Sep 11, 2016 20:21
|
|
|
e_angst posted:Interesting, a math sperg thinks that what works in his field applies universally in all other cases and all other disciplines. How unexpected! hobbesmaster posted:Yeah, usually it's physicists that do that. I named several fields besides math, but in any event it's hardly a controversial opinion that academics cut out the middleman (Elsevier had revenues of ~3B USD in 2015, Springer-Verlag ~1.5B) when they are the ones doing most of the paper writing, formatting, reviewing, and so on. I've published in a Springer journal before. My arXiv preprint shows up higher in Google rankings, is arguably better formatted, and -- notably -- doesn't cost $39.95 USD to read. Peer review and open-access are absolutely not incompatible or even related (all of the journals I listed are peer-reviewed!). In any event, what he said was Slanderer posted:You have clearly never looked at an open-access journal, because otherwise you wouldn't say something so loving stupid I hardly think it's being a 'math sperg' to point out that there are numerous healthy, high-quality open access journals, and entire fields that are working very hard to prevent (often publicly funded) university libraries from getting gouged. Academic CS is even less dependent on traditional publishers than math is.
|
|
#
¿
Sep 11, 2016 23:47
|
|
|
Discendo Vox posted:What would an ethical, legal Uber look like? It would look like a taxi company. Yeah I'm sure that a bunch of rent-seeking medallion owners is exactly what we'd arrive at if we started trying to reason out an ethical system from first principles.
|
|
#
¿
Sep 14, 2016 00:33
|
|
|
Tiny Brontosaurus posted:Google's still a fan of the "devote several unpaid days to a gauntlet of interviews with us" thing though, which should be next on the chopping block. Pretty sure their process is just screen + onsite, same with basically everyone else at this point (except the companies that are doing take-home exercises, which are significantly more bullshit in my opinion). My girlfriend just interviewed for a bunch of accounting jobs and each individual application required a bunch of exploratory conversations which culminated in an eventual onsite (and in some cases followups there as well). It's probably less 'performatively arduous' than getting asked a lot of coding/math questions at a whiteboard, but in terms of actual time that you'd have to take off from a job it's significantly less. I think a lot of other white collar occupations are in the same boat ('easier' interviews, but longer time taken for interview process). Interviewing at most decent tech companies takes precisely 1 day of PTO, which is literally the minimum that you would expect for a company to do their due diligence on you as a candidate. If you want to argue that the technical interviews allow for a fair amount of individual bias to creep in towards underrepresented groups then that's one thing, but I don't think that that's really the argument you're trying to make here (and asking questions that have a pretty clear 'right' answer is a countermeasure against those biases, though they obviously still exist).
|
|
#
¿
Sep 21, 2016 20:40
|
|
|
Tiny Brontosaurus posted:No, a full-day interview is insane and the fact that it's been so successfully normalized to you depresses me. Full-day means ~3 hours in this context (including lunch), not 8. But whether it's 2 hours or 12 you probably will need to take a day off anyways. hobbesmaster posted:But nobody gets the "correct" answers up on whiteboards, and if they do its because you've memorized that question and they'll go on a to a different one. This is super wrong.
|
|
#
¿
Sep 21, 2016 20:59
|
|
|
Tiny Brontosaurus posted:Paid interviews are great, but if you have the power, advocate for interview sessions outside standard business hours. A Saturday recruiting session would open doors wide open for a lot of people you're completely missing out on right now. So your proposal to make the interview process suck less for prospective employees is to make it suck more -- by a precisely equal amount -- for existing employees?
|
|
#
¿
Sep 21, 2016 21:39
|
|
|
Tiny Brontosaurus posted:Yes, exactly. If someone you're recruiting doesn't have some manner of portfolio of past work for you to review then you're filling an entry-level position and should get over yourself. Face-to-face interviews are about personality and gauging the candidate's interest in the company. Roughly 1/3rd of people in my role are Ph.Ds, and a shockingly high percentage of people I interview (despite graduating from top schools, having graduate degrees, working at other reasonably high-quality companies, etc), can't do basic probability/statistics, or simple data manipulation/transformation/analysis tasks in the interviews I do (you know, things I use several times in my line of work every single day). I don't care what their portfolio of past work is if their interviews go like this (and I'm sure their theses and graduate work have lots of very complex and impressive math that they don't actually understand that well). I'm talking things like not knowing the relationship between the density of a random variable and its expectation/variance, or not being able to spot the most obvious case of sampling bias ever constructed. This is the nature of Fizz buzz and other similar programming questions. WrenP-Complete posted:Fwiw, I've hired doctors, nurses, social workers and addiction counselors (so not all licensed professionals) with about 3 hours of total interview time. Perhaps this jobs are less technical but they involve a huge amount of trust because they have a duty of care. He's moving the goalposts. Most on-site tech interview processes are about that length. The ones that are longer are ones that have some off- or on-line component, like a timed automated test or a take-home assignment. But those are probably more friendly to candidates with less time flexibility. Tiny Brontosaurus posted:100 recruits or 1 recruiter, hmm... I thought you tech ubermenschen were supposed to be good at math? There also exist such things as overtime pay and a non M-F work schedule. You had me questioning whether I came by my math Ph.D legitimately with that jibe, but unless you've found some magical method by which a single interviewer can batch screen hundreds of recruits at once, this is a pretty uninteresting proposal. Also, people who are not looking for a M-F work schedule are generally not clamoring to work weekends (you know, the days all their friends and families also have off) in their place.
|
|
#
¿
Sep 21, 2016 21:59
|
|
|
Tiny Brontosaurus posted:I'm a woman, dummy, and I'm not moving poo poo. If you haven't personally heard of the kind of day-camp interviews I'm talking about then bully for you, they're still bad. Sorry for the misgendering. Nevertheless, I'm pretty sure I'm more in the loop about what typical onsite practices are in tech than you are. 5 30 min interviews is the de facto standard and has been for a long time. I'm sure you can find some startup that makes you do day-camp (I interviewed at one of them, a long time ago) but this isn't really representative. Tiny Brontosaurus posted:Yes I have, you loving idiot. I've been talking about it for several pages now, even. You want to whine about moving goalposts, why are you suddenly talking about screening when I've been talking about interviewing? You don't need to have people on-site to screen them. And your bitching about hypothetical work schedules is so loving dumb I don't even know where to begin, please remove yourself from human society. I meant interviewing, not screening -- my bad, though it's not like the two are essentially different in terms of time requirements. The point remains that just getting dozens of people in just creates a commensurate need for employees to work weekends, unless you have some proposal that allows you to interview them in parallel while getting the same signal.
|
|
#
¿
Sep 21, 2016 22:15
|
|
|
hobbesmaster posted:Scrappy little startups like google. "You'll usually meet with four Googlers—some potential teammates and some cross-functional—for about 30 to 45 minutes each." I'll let you do the math on how that compares with 'about 3 hours of total interview time'.
|
|
#
¿
Sep 21, 2016 22:19
|
|
|
Tiny Brontosaurus posted:I remain so goddamn depressed that you have accepted awful, time-wasting, bloated interview processes as so immutably How It's Done that you can only get sarcastic and sneer about "magic" when someone says there are other options. Yes, denizen from the Dilbert Universe, you can recruit people in other ways. Really truly. There is a world beyond the gray flannel of your cube. In turn, I find it depressing that you literally seem unable to conceive of the possibility that there may be a gap between someone's portfolio/resume and the skills they actually have, and I also find it pretty arrogant that you think that you are able to condense assessment of complex quantitative/technical topics into a few minutes (which is what you would have to do, if indeed it were possible to eliminate as many ineffciencies from the process as you seem to think exist). I guarantee you have no idea how frequent it is to get really bad candidates who also have really impressive resumes. I'm also fairly confident that many of these candidates would make it past the interview processes you propose to replace the existing ones. Slanderer posted:I had no idea stuff like this used to be in their interviews. This was a (small) part of the required physics curriculum at my school--the prof. gave it a particular name, but the takeaway was "back of the napkin math". These are called Fermi problems. I don't think they are great interview questions but IMO this back-of-the-envelope type math comes up a lot in what I do, and it's good to be able to estimate what the impact of your work will be. People who are particularly bad at this often spend a lot of time building out, and working on features that affect like 0.01% of our userbase.
|
|
#
¿
Sep 21, 2016 22:37
|
|
|
Tiny Brontosaurus posted:I literally do this for a living you smug douchebag. Maybe you're poo poo at recruiting because you're poo poo at human interaction. If you want actual specific solutions to your precious snowflake recruiting problems you can pay me for them. I mean, there's also lots of software engineers that do their jobs at a living and are bad at them as well. Nothing you've written here gives me any confidence that you aren't the analogous thing in your field. Have you also considered disrupting the educational system? Think of all those 3 hour, end-of-semester exams that could easily be replaced with an affidavit that the student indeed read their textbook or attended all of their classes. H.P. Hovercraft posted:how do these candidates have impressive resumes while simultaneously lacking critical professional knowledge that you would need to screen for via complex technical assessments? The archetype for candidates of this type is something like: Masters or Ph.D from [good to decent school] in [STEM field]. Worked as data scientist/analyst/engineer/SWE/whatever at [mid-tier company] or [startup I've never heard of]. Under that bullet, they say [implemented machine learning model in production], [raised company core metric by X%], [demonstrated x,y,z cross-functional skills]. Maybe they also took some Coursera courses or finished in the top x% on a Kaggle competition (both things where you can copy other people's code/assignments). Then they can barely write working SQL/Python and don't know the first thing about a linear regression or how to interpret/calibrate it.
|
|
#
¿
Sep 21, 2016 22:57
|
|
|
Tiny Brontosaurus posted:Also lol forever that Mr. King Genius Recruiter couldn't possibly look at a person's actual work product, their portfolio, and asses their work quality from that. Too Smart to Learn Things: The Tragic Tale of blah_blah As you may be aware, most people are unable to share the work they do for their previous employers in any significant way. Even if they were able to, it's very difficult to tell how much of the work is theirs, given collaboration and all that. For example, with Kaggle competitions, top competitors share enough of their basic work on each problem that you can probably get to top 25% or maybe even top 10% on the leaderboard just by copying that basic work and submitting it as your own. So it's meaningless as a quality metric. In rare cases candidates have very high quality public facing work (on their blog or GitHub or elsewhere). If I was interviewing the guy who wrote this I'd hire him on the spot, but alas this is not the norm (nor do I want it to become the norm that everyone has to produce unpaid analytical work on their own time on public datasets to get data science jobs in the future).
|
|
#
¿
Sep 21, 2016 23:19
|
|
|
Tiny Brontosaurus posted:In any case, while you have that impairment that prevents your brain from remembering what you were arguing about, I remember that the issue at hand here is expecting candidates to run a gauntlet of 1:1 interviews with multiple people. If judging quality is so incredibly impossible, what is it the fourth interviewer can catch that the first three didn't? The different interviewers/interviews test different skills. For SWE, it's more like algorithms/data structures + system design. For my role, it's some combination of algorithms/data structures + databases/data manipulation + math/stats + analysis of a particular data problem. So yeah, since the fourth interviewer is asking questions about a completely different topic than the first three, they are picking up different signals. You really are just guessing how all of this works, aren't you? Tiny Brontosaurus posted:That's not unique to tech. A comedy writer can't point at a late-night monologue and say "one of those jokes was mine." It is a challenge but gauntlet interviews are a poor tool for solving it. I'm getting annoyed at the smuggo up there who keeps sputtering about how it's The Only Way and telling me I must suck at my job because I know more than one way to tackle a problem. If there's one person here talking about The Only Way, it's probably the person who's making blanket statements about technical interviews being unreasonable for technical employees and 2+ hours being an unfathomably long amount of time to assess skills that have been cultivated for the last 10+ years, for jobs that routinely pay 200-500k a year. The particular irony of the situation is that you're basically doing here what every obnoxious tech 'disruptor' does when they go into a field that they have no domain knowledge of.
|
|
#
¿
Sep 22, 2016 00:05
|
|
|
Liquid Communism posted:H1B reform isn't because there's a lack of candidates. That's not it either. The real problem is that a disproportionate number of H-1B visas are eaten up by a handful of Indian consulting companies that pay their employees peanuts. This means that companies who pay their employees 2x or 3x what the likes of Tata and Infosys do, and who hire much more highly skilled and highly qualified employees, are completely unable to employ their desired candidates in the USA (with no guarantees of future employability either). Some sort of auction for visas (w.r.t employee compensation, not visa cost), or a significantly higher pay floor, would go a long ways towards making the system more efficient.
|
|
#
¿
Sep 22, 2016 05:43
|
|
|
The Real Foogla posted:wait doesnt that proove his point? they want to reform it to increase supply, driving down those prices No, they want to reform it so that their prospective employees, who are much more highly skilled and will be compensated 2-3x the amount that the candidates that Tata and Infosys are putting forward, aren't put on equal footing in the selection process. Overall 'supply' is not the problem. The problem is that the system puts unskilled, low-paid consultancy workers on an equal footing (or higher, given the experience of those H-1B shops at working within the existing system) with world-class, highly-paid FTE employees at top companies. That's why it needs 'reform'.
|
|
#
¿
Sep 25, 2016 01:11
|
|
|
nm posted:She was essentially very lucky to find the one professor who didn't want to sleep with her. You'd have thought Berkeley would be more progressive, you'd have thought wrong. Pretty sure that Berkeley is noted to be worse than average with regards to this sort of thing (they've had recent scandals in both their law and physics departments).
|
|
#
¿
Oct 1, 2016 05:50
|
|
|
Neo Rasa posted:Wasn't there a statistic a few months ago where they found that something crazy like 70% of SF's homeless camp inhabitants were residents until like a year and half ago? You might be conflating that with this stat (http://www.sfgate.com/bayarea/article/What-San-Franciscans-know-about-homeless-isn-t-7224018.php): SFGate posted:The most recent homeless count, conducted in January 2015, found 6,686 homeless people in the city. Seventy-one percent of people reported living in San Francisco when they became homeless, up from 61 percent in 2013. Just 10 percent said they were living outside California when they became homeless, and the remaining 19 percent were living in the state but not in the city. Most homeless people in SF are originally from SF or California, but that doesn't mean that they became homeless within the last 18 months. Homelessness has actually decreased from historical highs (http://sfist.com/2016/06/27/san_francisco_homeless_history_1982.php): "SFIst posted:Also, as the discussion of homelessness continues this week, people should bear in mind the falsehood that homelessness is a larger and more terrible problem than it has ever been in San Francisco. This New York Times piece from 1998 quotes a figure from the Coalition on Homelessness saying that 16,000 people were living on the streets of SF at the time — nearly 10,000 more than were counted in the most recent homeless census, though experts argue about that count and say that there may be as many as 10,000 homeless here currently.
|
|
#
¿
Oct 13, 2016 03:47
|
|
|
Dr. Fishopolis posted:He wasn't conflating anything, and the number is 71%, up 10% from 2013. 1) I said 'might' to cover my bases 2) Those articles quote literally the exact same sources as the ones I linked did. 3) None of those articles have any data on how long current homeless people in SF have been homeless for, which is what he was talking about. Dr. Fishopolis posted:edit: wait we're literally talking about the same data. You're really reaching to minimize this stat for some reason. No, the data doesn't show how long they've been homeless, but if there's a 10% increase in the homeless population that were former residents over two years, that's significant and illustrates a clear problem. I'm not minimizing anything. That statistic sounded completely implausible so I checked and indeed, it doesn't seem to be true. It has to also be contextualized by the overall homeless numbers in San Francisco. The '10% increase' that you refer to could just as easily be accomplished by a decrease in the inflow of homeless into San Francisco (if I had to guess based on the inadequate data presented here, it's also the most likely explanation to be true, given the long-term decline of homelessness in SF). Only an idiot would argue that there isn't a 'clear problem' with homelessness in San Francisco, and it's certainly not the case I'm trying to make. blah_blah fucked around with this message at 04:36 on Oct 13, 2016 |
|
#
¿
Oct 13, 2016 04:02
|
|
|
SurgicalOntologist posted:Even though it's now clear that there was no insider info involved, the perception remains. My understanding was that certain employees of company A had access to data from bettors at company A (prevalence of individual player picks, etc) and used this information to (successfully) wager in the analogous contests on company B's platform. Is this understanding incorrect (it quite possibly is) or are you arguing that this isn't insider info? One of the biggest problems with DFS relative to e.g. online poker in terms of a competitive ecosystem is that poker is a much more complex game to automate near-optimal play in (only HU limit holdem was anywhere close to being solved during the poker boom, which was very much a niche game), and that individual players can scale their play volume much more easily in DFS than online poker (efficiently multi-tabling more than a handful of tables was very difficult for all but a relatively small number of elite players, whereas in DFS it is relatively simple for very big winners to scale their volume to the appropriate amount given their bankroll/risk-tolerance/etc). blah_blah fucked around with this message at 01:14 on Nov 7, 2016 |
|
#
¿
Nov 7, 2016 01:08
|
|
|
cheese posted:Uber fares are definitely not that much cheaper than taxi fares. An Uber/Lyft to SFO is ~$25 for me. A cab is $47 + tip.
|
|
#
¿
Dec 3, 2016 08:12
|
|
|
the talent deficit posted:relying solely on uber to get around is a miserable experience that i can't imagine would be any better without human drivers. it's alright for infrequent use like getting home from the bar or not having to walk to the train station when it's pouring rain but it's never ever gonna replace mass transportation and private vehicle ownership I've done it for 3 years in SF and it's pretty great -- here. It's both surprisingly cheap (less than $2k/year for me) and incredibly convenient in a city that is geographically small and where parking is extremely pricey and difficult to find. It would not be pretty great in the South Bay, to say nothing of any city that is either substantially larger or has a substantially lower supply of drivers (basically everywhere). This is one of the fundamental problems with a lot of SF startups -- they are great at solving 'problems' faced by people living in SF with a lot of disposable income with a high propensity to tradeoff money for time. This does not describe most people in most places.
|
|
#
¿
Dec 4, 2016 00:24
|
|
|
the talent deficit posted:you're not replacing car ownership with uber you just live in an area where you can get away with rarely using a car The last part of that sentence is certainly true, but I don't think the first is. I would not be willing to only get around using public transit and I wouldn't consider cabs an adequate substitute, so I probably would own a car if Uber/Lyft weren't an option. Given the cost of parking/insurance/etc around here, this is a minimum 4-5k expense per year, not counting the cost of the car itself and maintenance on the car. the talent deficit posted:move to the suburbs, get married and have a couple kids and now you are looking at an order of magnitude increase in your uber bill. god forbid you live more than a handful of miles from your job or school or walmart. that's the reality for the majority of the population I don't disagree with any of this.
|
|
#
¿
Dec 4, 2016 03:30
|
|
|
Arsenic Lupin posted:Zoe Lofgren is introducing one that prioritizes visa allocation by salary (!!!!) This is pretty good if you look at exactly how the current system fails. Add an exemption for postdocs/academics and one or two other occupations, and you're probably there.
|
|
#
¿
Jan 27, 2017 00:22
|
|
|
the talent deficit posted:it's bad if you think h1b's should be used to bring in high achievers in diverse fields to drive economic expansion and it's bad if your primary concern is wage suppression of american workers but it's good if you're facebook, oracle, google, etc and you need bodies for your code mill the talent deficit posted:they're still cheap relative to who they are replacing. a developer at google or fb can make $200-300k pretty easily. they can bid $150k on h1bs to put to work optimizing ads, starving out industries that can't afford to meet that $150k and put a bunch of high earning americans out of work I don't think this is quite the case. The highest paying H1B jobs are healthcare and law, not tech, so it is not a foregone conclusion that an auction based system will just result in 100% of H1Bs going to the tech industry. With regards to your second point, H1Bs currently make market rates right now at top tech companies. In reality, anyone who can get a job at one of those is likely to be able to get a job at similarly selective companies, resulting in a reasonably efficient market. A developer at e.g. Google making $300k a year is generally acknowledged (avoiding the argument as to whether that is actually true or not) to be generating way more value to the company than his total compensation, meaning that the incentive to replace him with someone making $150k isn't all that high. Conversely, severely restricting access to the labor market for the likes of Infosys and Tata, whose entire model is built around undercutting American wages and eliminating the need for individual companies to hire their own tech workers, is going to be unequivocally good for American programmer wages overall. You can certainly argue that this isn't the best implementation of the system and that perhaps individual disciplines should have sub-quotas (I already mentioned academia as one), but the status quo is over 70% of H1Bs going to STEM disciplines and over 50% going to tech. There's a lot of gains to be had from just improving efficiency in those categories.
|
|
#
¿
Jan 27, 2017 05:24
|
|
|
Subjunctive posted:That sounds a bit ridiculous. YC's portfolio isn't big enough to constitute a meaningful market. It is pretty helpful if you are some sort of business-oriented SaaS solution in particular -- Zenefits is a good example of this. For more consumer-oriented products the network isn't as valuable.
|
|
#
¿
Jan 29, 2017 00:55
|
|
|
the talent deficit posted:an auction sounds good but it would just mean google/fb/oracle/salesforce replace highly paid americans with foreigners making the new auction price/h1b min These companies (at least some of them -- no idea about say Oracle) are paying H1Bs equally today so I'm not sure why you'd expect this to change under a new system. Under the current one they could also lowball their existing H1B applicants pretty much as hard as they want but aren't. the talent deficit posted:O1 is for like justin bieber and sidney crosby, not a dude researching hidden states of matter or whatever At my last job I worked with someone whose pre-tech resume was 'very good but not great postdoc in EE' and he got an O1 so I don't think it's quite as extreme as that. You need a really good publication record and to have other peer-reviewed work talking about you, from what I could tell. My guess is that with the right immigration lawyers -- which I'm pretty sure is not the case for most applicants -- a lot of tenured profs could get O1s.
|
|
#
¿
Mar 2, 2017 10:05
|
|
|
TheScott2K posted:Oh hey this is neat and sucks: https://arstechnica.com/tech-policy/2017/03/public-university-lays-off-79-it-workers-after-they-train-h-1b-replacements/ Note that this type of labor market arbitrage is more or less the worst abuse of the h1b system as is and also something that would be directly addressed via auctions or high price floors.
|
|
#
¿
Mar 2, 2017 17:43
|
|
|
Discendo Vox posted:Auctions and high price floors, as previously mentioned, create a different form of abuse. There are elements in the current h1b system that are intended to prevent this form of outsourcing, and a variety of other more effective reforms that don't open the door to greater abuse. I'm thoroughly unconvinced that these abuses exist or would be likely to exist. Auction off visas, carve out a separate quota or new visa for postdocs and tenured professors (which should probably also be allocated via auctions -- having more postdocs making 35k/year isn't helping anyone) and you're basically there.
|
|
#
¿
Mar 2, 2017 21:29
|
|
|
Discendo Vox posted:You don't understand why it's a bad idea to give technical visas solely to the companies with the most money to spend on them? Yes. I think that you overestimate the willingness of companies to overpay solely for the purposes of getting visas as well and probably also do not correctly estimate the composition of the absolute highest paid positions. They primarily aren't, and won't be, software engineers or tech workers. They will be executives, doctors, people working in finance, etc (feel free to verify this using http://h1bdata.info/ or any similar site). Visas are primarily allocated to generate economic value and attract top-notch talent and compensation is an efficient proxy for that. Even if you disagree with that (it's certainly not a perfect proxy) it's certainly a massive improvement over the existing system, where an engineer making 500k/year at Google is put on equal footing with an engineer making 60k at Infosys who, if selected, will probably end up displacing some American worker making 80k. blah_blah fucked around with this message at 00:25 on Mar 3, 2017 |
|
#
¿
Mar 3, 2017 00:05
|
|
|
|
| # ¿ May 10, 2024 22:26 |
|
|
Baby Babbeh posted:but we also don't want to create a system where Facebook and Google monopolize the spots because they're the only ones who can pay to bring in $200K unicorn engineers in bulk. There were 85000 H1Bs granted last year and Microsoft, Apple, Google, Facebook, Amazon had about 13000 applications in total. To contextualize that, the market cap of the S&P 500 is around 21 trillion dollars and the market caps of Microsoft, Apple, Google, Facebook, and Amazon combined are approximately 2.5 trillion. Even if every single one of their H1B applications was approved it would be only slightly disproportionate to their share of the economy. This doesn't even address the separate (wrong) assumption that people seem to be making, where these companies outbid the entire market across all 13000 applications, including higher paying fields like law and medicine, as well as their entry-level employees versus senior and executive levels elsewhere.
|
|
#
¿
Mar 11, 2017 04:57
|
|