1,911 karma · joined August 28, 2018
1) Excluding Googlers, which isn't very nice.
2) Applying pressure to de-evil Google, which is very nice and benefits everyone.
I'd like to have a non-evil Google again. I really miss the old non-evil, brilliant Google. Organizations change.
But from my perspective, I'm fed up enough with Google's anti-AGPL FUD that I don't mind not getting contributions from Googlers for a while. It's gotten worse than Microsoft in the nineties.
And with critical mass, Google will hopefully make a similar turn. Microsoft is now the world's largest open source contributor by some metrics. I don't know if I buy those metrics, but it's big. Same thing will happen to Google, hopefully.
If I'm buying into an open ecosystem, it's because I want an open ecosystem. I don't want a use-open-to-gain-market-dominance-and-than-milk-customers ecosystem.
AGPL is ideal for things I don't need to link proprietary code to.
What I most want is a plausible, paid, AGPL competitor to GSuite.
Others can also provide hosted services, They can't get a competitive advantage with proprietary "value-adds." Usually, the authors have significant reputation and branding advantage, so unless they f- up, competitors are bottom-barrel low-cost providers.
AGPL without CLA is whom I tend to gravitate to.
On paper, there is also path by which you can marry into Chinese citizenship, but it's exceptionally rare in practice.
"Renounce other citizenships" is complex. Many nations don't provide a citizens with a means to stop being citizen. Which is to say, if you take a Chinese (or any other) citizenship, and renounce your old citizenship, you may still end up with both citizenships.
I don't know specifics about Honk Kong, but I do know the standards are laxer than mainland China.
On the whole, I find these 20% performance memory upgrade (leading to a fraction of that in real-world performance) more obnoxious than anything, but I'd love ECC!
This thing is a nice first prototype, but it looks like a death trap. There is absolutely no reason for the giant lever arm coming up from behind you and all the way around.
Look at the altwork design. It's a lot lighter and more robust. It doesn't place a ton of weight overhead, and there's a lot less leverage on the joints. Or look at ErgoQuest. The monitors are stationary, and only a lightweight desk/keyboard tray moves.
I appreciate your confidence, but I'm afraid it's misplaced. I've been doing woodworking and engineering for a lot more than 15 years. A couple of things to consider:
"When you want to build this kind of structure, I'm not sure you can plan everything beforehand, there are always things that will happen that you didn't expect." You absolutely can. It just takes practice and experience. When I build something, I plan out what I do, and 9/10 it comes together on the first go. For my mentors, it's 10/10. The first few years I was doing this, it was 0/10.
"After a few failing choices, I ended up with a main pole of 7cm x 7cm." You generally want about a 10x safety margin for dynamic loading and similar issues. If you had a few failing designs, you're way too close to failing in your current design.
Back to the overengineering point: Simply using massive beams of wood and bolts doesn't necessarily make things robust. A 7cm x 7cm piece of wood weighs a ton, and puts that much more strain on everything. That's true for all the pieces. I could build a more robust design with normal-sized boards. That has the distinct advantage that if there is a failure, what falls on me my hurt me, but won't kill or cripple me.
I'm glad you're trying, but by overestimating your expertise, you're putting yourself at risk. I'd encourage you to keep trying, but to also take this sort of constructive criticism seriously.
Think through how the desk ought to move so you don't need the giant lever arm, whether you want the monitors to move at all, and where to mount these so you can get in and out easily. I can think of a half-dozen designs where there are no massive lever arms, where nothing moves or just the keyboard tray moves, where if things break, I don't die, and generally simpler, lighter, and more robust.
https://www.homedepot.com/p/Everbilt-18-in-Full-Extension-Si...
They're rated for loads up to 100lbs, and they're plenty tough enough to hold up a full-sized MIDI controller. They're stiff enough too.
If you're worried, though, you can double them up (two on each side). Now, you're at $30. Typically, when you double up, the load-baring capacity more than doubles.
I have a planer, but I didn't use it at all in this project.
Might be regional. US has a ton of land and resources, so maybe wood is cheaper here. You're pricing in Euros, and Europe is a bit more densely populated.
I look at the first paper as generating a hypothesis. That's the first step in any scientific process. A few replications later, we've got a theory, and then a few more later, a fact. The first papers are as important as the rest of that pipeline.
What causes outright harm is:
1) Papers with poor methodology;
2) Authors overselling their results; or just as often
3) Readers / reporters giving too much confidence in preliminary results
To be clear, I was calling em500 on #3, not the authors on #1 or #2.
On a skim, I don't see obvious problems (but, then, I wouldn't expect to find them; most require a deep read). It's published in an open access venue, which is always a good sign. It also helps that you're from BC, which for the most part, (still) has a pretty good reputation for scientific integrity (there's been a rapid decline in scientific integrity over the past two decades, and that's hit the institutional culture across the river from you particularly hard).
So you're okay by me!
Compared to commercial, what he's got is:
1. Lighter (2x4s are light)
2. More robust (2x4s are tough)
3. Looks better (An off-the-shelf product would look better in a showroom, but this one fits his room exactly, and fits his equipment exactly; that makes a much bigger difference).
I'll mention: That's true for most of the crude furniture I make. I have fine woodworking, and I have not-so-fine woodworking. The fine woodworking pieces tend to be generic (e.g. a bookcase). The not-so-fine-ones tend to be special-purpose. Right now, I have custom stuff built up for remote work, for example, with cameras/cables/etc. properly routed.
* Bosch drill/driver kit: $100
* Circular saw: $100
* Tape measure: $20
* Square: $10 x2 (one big, one small)
Total: $240. That's 1/10th of what you quoted.
"Step it up a notch" kit as next step:
* Random orbital sander: $70
* Router: $170
* Planer: $120
* Guides: $20
* Clamps: $100
And we're up another $500. Still a good ways short of $2500. And that's at current Lowes prices. If you wait a few months, you can find complete 5-tool kits on sale sometimes.
I don't think $250 is unreasonable to spend on a hobby starting out. If OP is at FAANG salaries, then dropping $750 isn't unreasonable either. And if OP is a student, use a university woodworking shop, or buy used. These things go for a song on Craigslist sometimes.
Planer (or belt sander) saves money in the long run. You can recycle (free) old wood, use cheap S1S2E wood, and similar. For the most part, you just want to build a lot of stuff. Given the random orbital recommendation, I thought of the two, a planer was more versatile. It has unanticipated beginner uses too. E.g. If you mess up and make something 1/16th of an inch too small, you can trim a bit, and it will fit. Or if you can't do layout/cuts precisely and well, you can leave a bit of margin, and keep trimming a piece of wood until it fits.
You want a good pair of parallel clamps or pipe clamps too, for when gluing stuff, which can hold something precisely for a long time with a lot of pressure, but most of the time, you want something fast.
I also recommend a few clutch-style bar clamps, since a lot of time, you're reaching over something.
2. That's okay! If it's not quite straight or square, it's really fine for most pieces. Unless you've really s-ed up, you won't be able to tell.
3. I have plenty of pieces of furniture made from construction wood, especially in the kids' room, and they do just fine. Kids grow fast, and an elementary school table won't be what we want in middle school. Heck, we might want something different if and when hobbies change.
The point is to make a bunch of pieces quickly, and to develop the technique to be able to:
* Make a (roughly) perpendicular screw hole without having the wood splinter on the opposite side, and perhaps with a proper countersink hole
* (roughly) cut a piece of wood to length, with a (roughly) square edge
* Get intuition for the order to measure/cut/assemble pieces so everything fits
* ... and so on.
There's a bunch of mechanical skills (like making a square cut) and spatial reasoning skills to develop. The best way to do this is to do it over, and over, and over many times. Once past that stage, poplar is ideal. Starting out, make as much stuff as you can as quickly and cheaply as you can.
Even now, I'll sometimes cut a piece of wood to the wrong length, and I've been doing this a long time. The first few pieces? Forgetaboutit.
I am dismissing it as:
One paper = interesting hypothesis
Multiple replicated papers = theory
Multiple replicated papers, with substantially different methodologies = scientific fact
> Yes I have, on both sides. Maybe it differs by field, but I (and my fellow PhD students) took it serious enough that we rarely read and finished writing a peer review within a single day. (This was mostly on economic theory / methodology papers, where even understanding the paper can take multiple days.)
Interesting. I've worked in several fields, and if this is the case, economics is pretty special. Good on you!
It wasn't always this way. As an economist, you study incentives. Think about how to prevent your field from sliding the way others did. Right now, academic incentive structures dictate cursory skim, and cultures+behaviors tend to (often after many years) align to incentives. As someone in a field (apparently) further along that path than economics, all I can suggest is to act sooner rather than later. It's easier to preserve cultures of integrity than to fix cultures of the opposite.
A good place to get started is construction lumber -- just 2x4s. They're not bad once you clean them up with a sander and potentially planer and finish them, and they have a certain industrial aesthetic raw. They're cheap enough that if you mess up or change your mind, you can throw them away and try again. And if you want to redo a piece of furniture, you can do that too. If you want to make a custom mount for something, just drill a hole.
Homes look surprisingly good when all the furniture is custom -- floor-to-ceiling wall-to-wall bookcases, desks designed for what's on them, etc. -- even if it's just cheap construction wood.
From there, you can move upscale. You can go quite far with Home Depot plywood sheets. Done well, they look great.
I have a mix of fine woodworking pieces (which take a month or two to do), to things which take a couple hours to throw together. And it feels awesome once most of the furniture in your home is handmade. And it's a much better use of space.
Get a few 2x4s, a pair of cordless drills, some nice screws, a circular saw, a carpenter's square, a tape measure, and safety glasses and you're good to go to get started.
If you want to step it up a notch, get a random orbital sander, proper guides, (GOOD) clamps, a router, and a jigsaw. If you can at all afford it, buy NICE tools. Better beats more. I'd rather have a good Bosch cordless circular saw and drill than a full kit of Walmart toolkit. And they'll cost less in the long run, with less damaged work pieces, and fewer broken tools.
Also, Youtube videos are great. So are online woodworking forums. Classes are nice if you can afford them, but hardly necessary.
> This is true, but it's true for almost every observational study, and IMO too easy a dismission.
It's also a correct dismissal. The vast majority of results don't replicate. It's super-easy to come up with results like this paper just by playing around with what you control for (with a 5% of a false positive each time) until you get the result you want. It's almost impossible to tell if this happened. There are many similar issues.
> You can use this reasoning to dismiss, or more likely, to cherry-pick probably 95% of all scientific results from economic/medical/social/nutricion/climate science.
And you should! You should believe results only once they've been replicated multiple times, by multiple communities, with multiple methodologies.
> I'd say the onus on sceptics in this case
The onus is on the people presenting the result, in all cases. That's the scientific process. The authors need to defend the null hypothesis as vigorously as you can.
> This paper is likely headed for a serious economic journal, where it typically goes through at least days of labor to scrutinise the execution/details/methodology in peer review. I'm not invested enough to volunteer for that.
Have you ever gone through a peer review process? Seriously?
I understand you're not invested enough to do that, but neither are the reviewers. It's a volunteer position with no credit or upside. Most give a cursory skim, and either accept or reject. It's no better than chance.
Seriously. NIPS did a study, and found it is literally no better than chance.
Of my own papers, I've only had one where peer reviewers did an in-depth read. For the rest, acceptance was completely random. Reviewer comments often have little relation to the actual paper (many reviewers just skim the abstract and the intro, and don't even fully read those).
I'd like to know sensor size. Not "pixel width." Sensor size+aperture, to a first order, is what you want to know about a camera.
Most don't want to do the work (not implying laziness -- it's about interest / fit / etc.).
From my time in grad school, I don't mind working at 3 am (and most other grad students I knew, at least without families, didn't mind that either). The problem was the abusive nastiness. You're presuming student and faculty interests are otherwise aligned. They're not. The most successful paths to academic success I've seen include:
* Spending time raising money, hiring a ton of students, and getting a few good publications by law of large numbers
* Attaching students to many projects, and when one shows promise, sprinting to do most of the work to claim first authorship on a breakthrough result (at the expense of collaborators), and neglecting the rest of the projects / students (at the expense of the rest of the lab).
* ... and so on.
There is also a question of who becomes associated with the breakthrough: the student or the advisor? Yes, the first/last author split in publication helps, but at the end of the day, someone goes out and gives talks, media interviews, etc. I've seen plenty of student s-ed because the faculty member (who often did minimal work) takes credit for their work.
And there are megaprojects, where there's one nice result at the end, and a half-dozen graduate students who work on it doing menial engineering/lab work. Labs like that end up with zero students going for faculty (or other highly-competitive) jobs.
In other words, aside from baking data, most tactics to academic success involve carefully optimizing the odds, and burning a number of students and/or collaborators to have tenure-granting results associated with the faculty member.
As a footnote, back in the day, I was blessed with a terrific advisor, but the MAJORITY of faculty at MIT are abusive douche bags, and the system is set up entirely to protect them. And most students I know who left MIT into faculty slots did that by cheating at least a little bit, whether through credit theft, baking research results, or otherwise. Most MIT grad students don't cheat, mind you, but most highly successful ones do (and yes, I am in a position where I've seen many of these case play out at MIT, and a few at other institutions). Those go into tenure-track positions, and abuse their students the same way they were abused.
Former students also have a better viewpoint. Many people don't realize they were abused until years later, when they can look back on an experience objectively. Among my cohort, the number of people who would say they were abused today are much greater than the number of people who would have said it at the time.
(2) Quant hiring is super-competitive. most physics students wouldn't get jobs. A lot of quants are physicists, but the converse isn't true.
(3) And students? Not so much. Quants like to hire Ph.Ds or possibly ABDs.
Ones who are ALSO good programmers can go the CS route, but most aren't. You need to do basic programming to do physics, but not industry-grade software engineering.
Talk to them, and look up what they're doing. If a lot of former students ended up where you see yourself in 10 years, that's a great sign. If a lot of former students are in dead-end careers, that's a bad sign.
Of course, that discriminates against tenure-track faculty, but avoiding tenure-track faculty is good advice too. Tenured professors have the luxury of being able to care about you. Tenure-track need to do what it takes to succeed.
If they can, the next part is fit. You need to get over both bars to be hired.
All of this is incredibly noisy, so part of what happens is you have to set the bar high, since 37% of the people you interview will perform one std. div. above where they actually are, and 5% will perform 2 std. div. above where they actually are. Setting a high bar means you lose a lot of good people due to random error. But the costs of one bad hire are EXTREME compared to missing a few good hires. So it makes sense to do that.
People take rejections personally, but in most interview processes, it's pretty random. You have a bar. You hire people who interview 1-2 std. div. above that bar. You might be perfectly qualified, and I won't hire you if you had a bad day (or even an average day). The risk is too high. If you're 1 std. div. above, I'm moderately confident you're qualified. If you're 2 std. div. above, I'm very confident. And having hired people, that's how it plays out: I mostly hire people who interviewed well, and their performance on a typical day is typically lower than at the interview (not always, but usually).
Hiring is about finding and bringing in good people as efficiently as possible. I'd rather interview 20 people, hire two I'm confident are good (miss six qualified people among those 20 who had bad days), than I would bring in 10 people and hire two of them.
And bad people still slip in. You just can't do that much in an 8 hour interview.
For startups, I go for personal recs and github repos. Does that mean I miss people who aren't in my network, and who don't do open source? Of course. But that's okay. I still find enough really good people that way. (And yes, my network IS diverse, thank you very much).
But big companies need standardized processes, or things get rife with discrimination, nepotism, and all sorts of other nastiness. That easy to avoid small-scale, but with 1000 employees, it takes one bad apple to land on front page NY Times.
Even if you assume 150% developers at 300% the cost, it's worth it. The number of communications channels grows quadratically with the number of people in the org. You'll come out ahead at larger organization sizes. Fewer, higher-productivity people is a huge win in software with compounding effects.
And the next problem is most markets in software are winner-takes-all or winner-takes-most. Being the runner-up to Netflix is not a fun spot to be.
The place where 1x developers do well is routine work, especially on smaller projects. For example, in-house IT work at a bank, manufacturer, etc. tends to not have those same economies.
I was responding to: "I am puzzled why there aren't more split keyboards on the market, they are such a good idea."
I would argue, for all the reasons I gave, most are horrible for most people. That's the context. You have to read in context. That doesn't mean they won't work for you. Most people want $5 keyboards. That's followed by a long tail of oddball keyboards, of which ergo are one breed.
Most of the keyboards I use would be horrible for most people too. But they work for me.
If I give up a few things I like about my existing keyboard to get what you like about your keyboard, that's a bad deal. I think that's true for most people.
The type of mechanical keys on my keyboards really reduce the amount of strain on my wrist relative to membrane keyboards. It's not just a matter of taste. There's much less of a bump on the bottom if the key can continue moving after the keystroke is triggered. Membrane keyboards can do a little bit of that, but not nearly as well.
Trackpoint, implemented well, means I can do casual mousing (e.g. switching focus) without moving my hands to the mouse. That reduces shoulder strain.
You're welcome to call things I like "a bad idea," and I could do the reverse just as well, but at the end of the day, I ain't buying. I know what works for me, and I haven't found an "ergo" keyboard which does.
One of the reasons is overlap. Your hands do occasionally hit keys on the other side. You want a pair of 2/3 keyboards, rather than 1/2 keyboards. You /especially/ want a lot of the "special" keys on both sides.
Another is key action. You want proper mechanical keys.
A third is mounting. That's a problem which hasn't really been figured out. The middle hinge or fixed setup is definitely sub-ideal. Two pieces are a bit clunky. Armchair mounts are $$$.
For me, I also like having a trackpoint. I'm not saying that's common, but a lot of people want /some/ oddball feature like that. For one person, it might be lighted keys. For another, swapped capslock. Etc.
And then there's the mass of people who don't care about keyboard and buy the cheapest possible model. Split keyboards are $$$.
On the whole, the concept is good in abstract, but all models I've used have had one fatal flaw or another.