Erik Haugen @ech@qoto.org
Computer programmer
"From what we can tell, Haugen works at Google. So much for "Do no evil."" – Kent Anderson
Joined Mar 2023
@TruthSandwich @freemo I'm having a failure of imagination. How does his candidacy help Trump?
@rmathematicus @josephwelch @Sheril Yeah, I get filmmakers have to change things to fit the format, but do they really need to change this much? Especially when it's tarnishing Turing's and Denniston's reputations.
Erik Haugen
boosted
Academics: stop being coy about #SciHub and start treating it like basic research infrastructure. If you dont include it in your syllabus already as a normal way to access research, you should start. No more winks and nods, just link directly to it and accept no criticism for doing so from the researchers that necessitate its continued existence by their publishing practices
@ratel @Strandjunker They'd quickly be blocked from some big instances if they were outed.
@onideus@med-mastodon.com @Strandjunker https://theintercept.com/2021/09/11/september-11-saudi-arabia/ conspiracy mongering; or, at best, calling the Saudi Crown Prince "the people behind 9/11" is pretty silly, at least based on anything we know publicly.
I mean, there are other pretty serious concerns we maybe should have with the Saudi government, don't get me wrong.
@danluu Are there data for this sort of thing about other respiratory illnesses as well? I know flu/etc can cause long-term problems as well, but I suspect not at these rates.
@rudy How would an LLM get this correct? I would assume if it got it correct that it saw the puzzle in its training data and is essentially parroting the answer. Isn't that how it answers questions like how to code quicksort or whatever?
@antares @georgetakei For president, pretty much yes. check in on electoral-vote.com to make sure your state is still in the "swing" category. (I'd maybe include North Carolina and make it 7 at this point?)
But if you live in California or Wyoming vote for whomever makes you happy. For president.
@danluu Or maybe I'm still misunderstanding you: maybe a more charitable reading of Redacted's "more participants, more revenue etc" is: "hobbies/sports that haven't languished, but instead have 'grown' are generally going to be such that today's 16yos are roughly at world-class 1970–80s level". IOW: you two are saying basically the same thing: training/etc has matured for a lot of sports since that time. (except for hobbies where it hasn't.)
I think the problem around "that could've been done by a SWE-2" has to do with the relative difficulty of telling how good someone is in software engineering vs. other fields.
For SWEs, it's easy (relatively) to recognize a complex system. It's quite a bit harder to recognize that a simple solution was simple because of the brilliance of the designer, and not because it's a simple problem. (That's my assertion here, anyway!)
It's kind of reminiscent of the "streetlight effect": we look at the easy thing to look at because we're better at looking at it.
It's critically important to combat this when designing an evaluation process for SWEs. If you're a SWE evaluating other SWEs, you need to be self-aware around this. If you are a SWE at a firm, leave if they get this wrong.
I think this is the right way to look at it: You want to promote people who (1) work on hard problems and (2) do the work well. (There's also leadership and various other factors, but let's just consider 1 and 2 here.)
(1) is important – another way to put this is that if SWE-2s can do all your problems, then you can just hire SWE-2s. (And they should be wary of working at your firm for long – there's little growth potential; they aren't going to learn much.)
(2) can be accomplished by finding a really simple solution that nobody else could think of. e.g. sr. folks built something really complex to do the job because they thought it was necessary, or "everyone" thought we needed a complex system, and then this candidate redid the system in a really simple way that worked at least as well. You have to go out of your way to be open to recognizing this.
Worth noting: "that could've been done by a SWE-2" might be legitimate! Maybe it just wasn't a hard problem. IOW: would a SWE-2 have recognized it as something they could do? If so, no promo.
@danluu Did you know that the best HS swimmers today can beat Spitz' times from 50 years ago?
Does this cause you to reconsider whether this comment is obviously absurd?
@danluu I'm not sure I understand what you're saying about Redacted's null hypothesis. It *sounds* like you're saying it's stupid because it should be obvious that a 16yo today should be nowhere near as good as an elite athlete 50 years ago, even in a sport that grew in significance?
In more than a few sports, high-schoolers today could roll with elite athletes from 40–50 years ago – even sports that were pretty big back then – so that assumption seems like a "wrong to anyone who has contact with the real world" kind of thing.
@burgerdrome @SubElement I was thinking that, too, but the barrel is not plastic. 😂
@NatureMC @stefano open an incognito window or a different browser and you can easily read that person's posts. With Mastodon it's even easier: just use a different instance that you aren't logged in to – no incognito window required.
"Logged out of the Fediverse, you can see more than my public posts?" – ah, I only had public posts in mind with what I wrote. I have never used non-public posts with Mastodon, so I have no idea; if that's what you were referring to beg pardon.
@NatureMC @stefano "If you block haters, depending on the instances, often they can still see your profile and read your posts." <-- This feels impossible to fix. maybe it's a failure of imagination on my part or a different set of assumptions, but if it's possible to read your posts while logged out, then it's fundamentally going to be easy for me to read your posts while logged out after you block my account.
@jmcrookston @dangillmor Well, they do send data off the car for "safety-critical events"
They say about Tesla:
* good that they don't sell data to 3rd parties (but you can opt-in, and maybe that is confusing)
* reminder of the scandal regarding employees sharing pictures from the cameras
* privacy policy is somewhat vague in some areas (sharing with law enforcement, and so on)
* if you opt-out of all data sharing, you don't get software updates/etc because they cut all connectivity, so that is stupid
I think those are the highlights.
my note: the picture sharing scandal implies super-poor internal processes for handling customer data, which is unfortunately far too common for low-end tech firms. (I've worked in several, and it is quite shocking.)
@jmcrookston @dangillmor although, the whole "safety-critical event" thing makes me wonder – suppose you're speeding in an accident; will that be subpoenaed?
But I guess they could probably even subpoena your cell records to get that, so I suppose that ship has sailed.
@jmcrookston @dangillmor Is it really the worst, though? They mostly don't send data off the car unless you opt-in or there's a crash/security event or something, and they let you delete everything associated with your account. They seem pretty upfront about what they do. (Unless they are outright lying about it, which I suppose is a possibility.) This article is mostly mocking the agreements, which I appreciate, but as far as what they actually do, it seems about what you'd expect? am I wrong?
I suspect most car makers are actually pretty good, as well. Same problems with any bush-league online merchant apply, of course – financial details stored using less-than-best-practice.
@goc @dangillmor @mozilla The article is mostly making fun of privacy agreements & policies and not very usable about practical questions like that. For example, while Teslas are more than capable of collecting all kinds of information on you, as a practical matter, unless they're just outright lying, which I suppose is possible, they don't send that data off the car in most cases.
I suspect the items about sexual activity are there just because any cameras on the car might pick that up, so they're enumerating it – it is something to keep in mind, I guess?
I suspect most data selling is the same sort of data selling every other merchant does.
This is hilarious.
@EU_Commission Dear EU, The following page appear blank which using the privacy-focused #TorBrowser: https://ec.europa.eu/commission/presscorner/de...
Computer programmer
"From what we can tell, Haugen works at Google. So much for "Do no evil."" – Kent Anderson
Joined Mar 2023
