|
signed my offer for newjob wednesday and submitted my resignation yesterday.  bout a 50% cut in TC because I’m leaving a shitload of RSUs on the table, but a decent base pay bump and getting back into startup land which is what I wanted. working at a big public company was killing me, literally (I developed hypertension this year from all the bullshit going on there) so getting out of it is worth it.
|
#
?
Jul 28, 2023 13:11
|
|
|
|
| # ? Apr 27, 2024 06:52 |
|
|
rotor posted:yeah i mean i get it, just sayin its very weird to me that academia doesnt self-publish. Get with the program. eggheads. definitely, especially since no one in the whole chain of publishing papers actually gets paid except for the journal
|
|
#
?
Jul 28, 2023 19:29
|
|
|
You may not get paid on a per-paper basis but if you stop publishing you will discover the limits of tenure pretty quickly.
|
|
#
?
Jul 28, 2023 19:32
|
|
|
kitten emergency posted:signed my offer for newjob wednesday and submitted my resignation yesterday. One of my former coworkers was recently telling me that I should never, ever let my work impact my health and I think it has to do with her recent diabetes diagnosis  Congrats on the step forward in any case!
|
|
#
?
Jul 28, 2023 20:02
|
|
|
in academia it's not JUST publishing. _just_ publishing is a stand-in for people actually paying attention to what you write. the assumption is that if you are publishing in good journals, people will actually read them. self-publishing, pay-to-publish, etc type stuff basically doesn't count because those journals are full of trash and aren't really worth reading. it's a very very broken system but "have no editorial oversight" ain't the solution. at least in the humanities. i kind of expect that the grant-funding-review process could stand in for this in a lot of cases in the sciences since people would still read and respond to the output of the prestigious 2023 Poodle Grant or whatever. but the lack of external funding sources in the humanities (among other things) means that the self-published world is an absolute cesspit
|
|
#
?
Jul 28, 2023 20:10
|
|
|
Achmed Jones posted:it's a very very broken system but "have no editorial oversight" ain't the solution. at least in the humanities. i kind of expect that the grant-funding-review process could stand in for this in a lot of cases in the sciences since people would still read and respond to the output of the prestigious 2023 Poodle Grant or whatever. but the lack of external funding sources in the humanities (among other things) means that the self-published world is an absolute cesspit my field of chemometrics (well used to be my field but that's neither here nor there) gets published in all sorts of Elsevier journals and it's a cointoss if articles that are just through and through terrible gets through. Basically imagine someone waving their hands and going "AI" could just publish whatever the gently caress they wanted as long as the word spectroscopy was tacked on at the end.
|
|
#
?
Jul 28, 2023 20:57
|
|
|
wash bucket posted:Gonna have to disagree with this one. I mean, I'm sure there are some decent places to work but for a lot of them "efficiency" means having the bare minimum of staff and resources. yeah, on second thought I’m walking manufacturing back
|
|
#
?
Jul 29, 2023 19:25
|
|
|
champagne posting posted:everyone should just post to free journals like arxiv instead of the terrible money machines that are Nature et al. so "anything goes" places like arxiv end up in the lower end of this thought process
|
|
#
?
Jul 29, 2023 19:37
|
|
|
personally i think that at least in theoretical cs good international conferences are a better vehicle than journals. bakes in a willingness to communicate with people, and as the program committee is to a great extent trying to make an interesting program (many of them often being attendants as well) it filters better for the common incredibly irrelevant esoterica. the theory is that the "pure" journal stuff is further refined, but i really don't think that is how it works out in practice.
|
|
#
?
Jul 30, 2023 18:36
|
|
|
Maybe it works on the theory side, but the conference system has been an absolute disaster for anything with a strong empirical component. You could maybe patch that a bit with stronger training in how to do empirical work, but really the whole system is a relic of the pre-internet days and we could safely bury the whole thing.
|
|
#
?
Jul 30, 2023 18:50
|
|
|
ultrafilter posted:Maybe it works on the theory side, but the conference system has been an absolute disaster for anything with a strong empirical component.  e: maybe I'm misunderstanding what you're saying, but it isn't clear how a journal submission is fundamentally better than an conference submission except that there are more words in the former Dijkstracula fucked around with this message at 18:54 on Jul 30, 2023 |
|
#
?
Jul 30, 2023 18:52
|
|
|
Going into depth on this is probably a bit much for the interviewing thread, but here's the very short version: About ten years ago for various reasons a group of psychologists became interested in replicating some of the experiments that had been published in their journals. The aftermath of that is described very nicely in Stuart Ritchie's Science Fictions: How Fraud, Bias, Negligence, and Hype Undermine the Search for Truth but the tl;dr is that in the last decade a lot of people who thought their fields were on solid evidential ground have found out that it isn't so. Some of it's because people aren't always great about publishing with someone who really knows how to do rigorous empirical work, but a lot of it comes from the incentive systems created by the tenure/promotion and grant systems. In computer science, those problems are exacerbated by two factors. First, it seems to be very unusual for researchers to receive formal training in empirical methods to the degree that you'd find even in a psychology program. Second, the bad incentives that exist elsewhere are mixed in with a bunch of deadlines meaning that there's less time to get things right. Now that doesn't prevent good work from being done, but it does make it harder. I'm most familiar with machine learning and I can tell you the floor is set pretty low there. Of course none of this constitutes proof that these things are being done poorly in other subfields, but given the issues I've described above I'm going to claim that the burden of proof is on researchers in a field to prove that they're doing things well. If that has been done in any other parts of CS I'd definitely be interested in seeing it.
|
|
#
?
Jul 31, 2023 03:02
|
|
|
ML reproducibility gets pretty bad for like, a formal science but its not quite psychology-of-2005 bad. lots of CS is still true formal science, just not ML "how am i gonna repro this proof that IP=PSPACE" "read it and if you find counterexamples you failed to repro, namaste" so it does tend to work in formal-land bob dobbs is dead fucked around with this message at 03:40 on Jul 31, 2023 |
|
#
?
Jul 31, 2023 03:36
|
|
|
Time for my yearly interview run. LinkedIn Jobs is still the one to hit?
|
|
#
?
Aug 1, 2023 00:32
|
|
|
Rudest Buddhist posted:Time for my yearly interview run. LinkedIn Jobs is still the one to hit? Yes
|
|
#
?
Aug 1, 2023 00:50
|
|
|
|
|
#
?
Aug 1, 2023 01:03
|
|
|
ultrafilter posted:About ten years ago for various reasons a group of psychologists became interested in replicating some of the experiments that had been published in their journals. The aftermath of that is described very nicely in Stuart Ritchie's Science Fictions: How Fraud, Bias, Negligence, and Hype Undermine the Search for Truth (Also, apologies, I think I confused the previous poster bringing up journals and your post, and I conflated the two, thinking you were advocating specfiically for journal publications but not conference ones. Sorry!)
|
|
#
?
Aug 1, 2023 03:11
|
|
|
Dijkstracula posted:this one looks interesting, I see the library has a copy so I'll put it in my queue! In my perfect world computer science would switch to a journal system for the reasons I laid out, but I recognize that there are other improvements that are more likely to happen in the short term.
|
|
#
?
Aug 1, 2023 04:04
|
|
|
got my last rejection I gave a poo poo about from the start up. i'm officially on sabbatical and I'll get a full 2 months in Japan. Was worried I'd have to cut it short to go back to work. i've been cramming basic japanese since last week and now can read hiragana and 12 kanji
|
|
#
?
Aug 2, 2023 02:49
|
|
|
cheese eats mouse posted:got my last rejection I gave a poo poo about from the start up. i'm officially on sabbatical and I'll get a full 2 months in Japan. Was worried I'd have to cut it short to go back to work. be sure to get a pinku bento box
|
|
#
?
Aug 2, 2023 02:50
|
|
|
Captain Foo posted:be sure to get a pinku bento box i had to google this cause i assumed forums lore and lol
|
|
#
?
Aug 2, 2023 03:01
|
|
|
upon review i would like to qualify my previous statement about liking academia with the very important caveat that i had no aspirations of professorship or tenure so i did not give a single poo poo about publishing. if i helped someone with something enough to get my name on their paper, neat, but i didn't really care either way.champagne posting posted:everyone should just post to free journals like arxiv instead of the terrible money machines that are Nature et al. Quality of papers clearly don't matter to them (at least not for my field which is extremely hit or miss) so why not just post papers wherever in my (former) field arxiv was specifically where you dumped garbage that none of your peers thought was any good. i know that not all fields are like that though.
|
|
#
?
Aug 2, 2023 03:15
|
|
|
ultrafilter posted:In my perfect world computer science would switch to a journal system for the reasons I laid out, but I recognize that there are other improvements that are more likely to happen in the short term. yeah I guess I'm back to not seeing how journals help if the problem is "too many people are trying to submit low-value junk to neurips" because those people will just staple together three low-value conference papers and call it a journal paper, but I'm happy to concede the point if the ML community promises to never put me on the program committee
|
|
#
?
Aug 2, 2023 04:42
|
|
|
pure malthusian population pressure. jonathan swift knew what to do
|
|
#
?
Aug 2, 2023 05:11
|
|
|
cheese eats mouse posted:got my last rejection I gave a poo poo about from the start up. i'm officially on sabbatical and I'll get a full 2 months in Japan. Was worried I'd have to cut it short to go back to work. grats im considering doing this myself
|
|
#
?
Aug 2, 2023 05:18
|
|
|
cheese eats mouse posted:got my last rejection I gave a poo poo about from the start up. i'm officially on sabbatical and I'll get a full 2 months in Japan. Was worried I'd have to cut it short to go back to work. がんばって！
|
|
#
?
Aug 2, 2023 06:49
|
|
|
AnoHito posted:がんばって！ はじめまして！
|
|
#
?
Aug 2, 2023 07:26
|
|
|
cheese eats mouse posted:got my last rejection I gave a poo poo about from the start up. i'm officially on sabbatical and I'll get a full 2 months in Japan. Was worried I'd have to cut it short to go back to work. i only ever learned one kanji
|
|
#
?
Aug 2, 2023 09:41
|
|
|
more like kan'tji
|
|
#
?
Aug 2, 2023 11:21
|
|
|
Cat Face Joe posted:cjs: just had an interview with a company. 20% raise, probably a lot more work, and the reporting manager seems a little unhinged. it's stuff ive got a lot of familiarity with and it sounds like there could be some actual success instead of here where ive got zero upward mobility. only hiccup is he asked if i worked with [previous manager who fired me] so we'll see how that shakes out
|
|
#
?
Aug 2, 2023 22:19
|
|
|
congratudolensces
|
|
#
?
Aug 3, 2023 06:37
|
|
|
Jabor posted:more like kan'tji Immanuel kantji
|
|
#
?
Aug 4, 2023 14:02
|
|
|
i got sent a hackerrank for an fpga engineering internship. very curious what thats going to look like; i thought these positions would be less competitive than software positions. and it's definitely way harder to practice "problem solving" with an hdl since to my knowledge there's no leetcode equivalent
|
|
#
?
Aug 4, 2023 17:41
|
|
|
They probably just make every _____ Engineer position do the hacker rank despite whatever the job listing/qualifications actually say. Touching a computer? Hacker rank!
|
|
#
?
Aug 4, 2023 17:54
|
|
|
lord fifth posted:i got sent a hackerrank for an fpga engineering internship. very curious what thats going to look like; i thought these positions would be less competitive than software positions. and it's definitely way harder to practice "problem solving" with an hdl since to my knowledge there's no leetcode equivalent i think they still let you register and use some commercial simulators for free, and have some tutorials, i did a take home interview for cirrus logic on this site. https://www.edaplayground.com
|
|
#
?
Aug 4, 2023 17:57
|
|
|
EIDE Van Hagar posted:i think they still let you register and use some commercial simulators for free, and have some tutorials, i did a take home interview for cirrus logic on this site. oh this is neat. usually i just rawdog vhdl using iverilog and vscode
|
|
#
?
Aug 4, 2023 18:06
|
|
|
lord fifth posted:oh this is neat. usually i just rawdog verilog using iverilog and vscode
|
|
#
?
Aug 4, 2023 18:07
|
|
|
i have performed shameful posting errors
|
|
#
?
Aug 4, 2023 18:10
|
|
|
shame quad
|
|
#
?
Aug 4, 2023 18:10
|
|
|
|
| # ? Apr 27, 2024 06:52 |
|
|
this eastern european man just got very angry when i asked if they had any plans to update their process to use a modern architecture within their ios app because they use mvc.
|
|
#
?
Aug 4, 2023 20:09
|
|
