Share good ideas and conversation.   Login or Take a Tour!
ilex's profile
ilex


stats
following: 5
followed tags: 22
followed domains: 0
badges given: 2 of 2
hubskier for: 184 days

a little holly plant growing on the windowsill


ilex's recent comments, posts, and shares:
ilex  ·  3 days ago  ·  link  ·    ·  parent  ·  post: Uber's 5.6 Seconds of Incompetence

This is more than the classification software getting confused. This is more than a couple programming mistakes. This is more than bad software design.

Uber failed at even the most basic safety engineering. Not only did they not consider safety when building their own stuff, they even intentionally disabled existing safety measures.

The thing that really gets me is that self-driving cars have a lot of hard problems to solve, but all these factors are not open problems! Engineers in safety-conscious fields have been thinking about and preventing problems like these for fucking decades. We know how to keep drivers paying attention even when the thing they're driving is mostly automated. We know how to alert operators of unusual circumstances early so they can make informed decisions. We know, for christ's sake, how to run several unreliable control systems and take the majority vote of their results if we're worried one might be making a mistake! This shit isn't even new -- someone from 1995 could tell you how to do all these things without knowing anything about the last 30 years of technological advancement.

But goddamn it, software engineers, if they think of anything, think of security -- how to stop "bad things" from happening. Or they think in probabilistic terms -- 90% accuracy is pretty good, right? Safety engineers, though, they know that no matter what the probabilistic models say, something bad will always eventually happen -- so how do you stop that bad thing from hurting humans.

And then, on top of all that, you have some really fucking stupid programming decisions that should never have made it onto a public road in the first place. Jesus christ.

ilex  ·  3 days ago  ·  link  ·    ·  parent  ·  post: "If it's not Bitfinex, it's somebody they do business with very frequently."

"the company revealed in a court filing that tether was only 74% backed by reserves"

Given a total supply of 4,207,771,504 USDT, that's a bit over a billion dollars they've printed out of thin air.

ilex  ·  3 days ago  ·  link  ·    ·  parent  ·  post: Pubski: November 6, 2019

Conference went great! Met a lot of cool people, learned about some fun and "fun" stuff, and people seemed to like my talk.

I also got a new ruffly polka dot shirt in in time to wear it for my presentation:

Now I'm back home and my chickens are so very happy to see me.

ilex  ·  11 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 30, 2019

They are very warm, soft, and cuddly!

ilex  ·  12 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 30, 2019

Chickens

Bertha is inside until she puts a little weight on -- she is severely underweight due to her illness and as it's getting cold out I worry that she isn't strong enough to withstand the weather.

I can't really complain though because she really is an excellent snuggler:

Research

Finished pretending to be an EE for a little bit -- today I wrapped up a tiny (10 piece) production run of a little fiber-optic transceiver board we're going to use for some of the research we do here. Now I get to figure out how I'm going to explain my other research next Wednesday. I am really worried that it's going to come off either trivial and useless or pointlessly complicated (or both) but we will see what I come up with.

ilex  ·  12 days ago  ·  link  ·    ·  parent  ·  post: Facebook allows prominent right-wing website to break the rules

At the very least, it should make you suspicious of corporate pride stuff. I mean, it's cool that "people who support LGBT folks" is an advertising demographic nowadays, but it'd be nice if that meant material changes rather than companies just trying to make a buck off people being gay.

ilex  ·  14 days ago  ·  link  ·    ·  parent  ·  post: Thoughts on sympathy and empathy

Something important to keep in mind is that when shit sucks, it's nice to hear "yeah wow that sounds hard, I'm sorry it's like that" (what I'd call sympathy?), but this should be just the first part of 'helping others'. Whether you want to call it empathy or compassion or whatever, we need to understand why shit sucks and what we can do about it. And, critically, we can't predicate that understanding on having been in the exact position someone else has been in! Whatever power we might have, we need to use it to make life better for those who lack that power, and the only way I see that happening is through empathy.

ilex  ·  18 days ago  ·  link  ·    ·  parent  ·  post: Stephen Wolfram - Announcing the Rule 30 Prize

There are lots of things that are "patterns that humans created" -- that doesn't necessarily make them uninteresting. There certainly ought to be space for personal curiosity in mathematics! Sometimes it's just fun to ask questions and solve problems.

But also, cellular automata (and fractals) have a lot of relationship with dynamic "chaotic" systems which exist all over the place and are notoriously difficult to understand. This includes physical systems (which sometimes humans tend to be great at controlling, but explaining how to do it to a computer is devilishly tricky), quantum computing, and machine learning things.

ilex  ·  19 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 23, 2019

Just a poster at this one, but it went well and people seemed interested. That project needs to hurry up and get some results so there's substantive conversation to be had about it!

They're good pets; you should do it! She is, unfortunately, so she'll be back in when I'm back home.

ilex  ·  19 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 23, 2019

It's hard to break something like that off, even if it is for the best. Hopefully you'll start to feel the positives soon.

ilex  ·  19 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 23, 2019

I'm drinking shitty hotel coffee and listening to nerds talk about stuff -- which is to say, the conference is going pretty well. There's a wide variety of topics which is at least entertaining.

edit: chickens for kingmudsy

ilex  ·  25 days ago  ·  link  ·    ·  parent  ·  post: Dyson cancels electric car project

It's for the best -- it probably would have sucked if they actually made it.

ilex  ·  26 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 16, 2019

I mean, chickens will happily eat chicken noodle soup and they'd probably love it but it's too salty to be healthy for them (and they should avoid chicken...other meat is fine, though)!

She fasts today; tomorrow she gets regular chicken feed with some yogurt or other protein. Plus she always gets some of whatever I'm eating. I've seen her swallow whole penne noodles bigger than her head like it was nothing.

ilex  ·  26 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 16, 2019

Chickens

Bertha (the very hugged chicken in the above photos) has sour crop (a kind of yeast infection), so she's inside for the next few days. No food today; hopefully everything will clear up and we can get her back out to the flock in a couple days with no trouble.

Research

Conference next week! I guess I should finish the poster for it, huh?

I have at this point four projects going, which is about three more than I probably should have. But, at least it is all interesting and several things seem to be going directions. I am eternally grateful to the four-month-long checkout times from the university library because there's a lot of reading to do and sometimes it takes months before I get back to something.

ilex  ·  28 days ago  ·  link  ·    ·  parent  ·  post: Uncovering the Seeds of a Post-Lawn Future

Oh, excellent!

ilex  ·  29 days ago  ·  link  ·    ·  parent  ·  post: Uncovering the Seeds of a Post-Lawn Future

He's so good, and you'll pick up a half decent botany education if you watch enough! It'll change how nature looks -- my wife has extensive training in botany, and after a few walks through the woods now all these little details show up that I would have missed beforehand. (Also I have a newfound hatred for honeysuckle since it is an invasive species out here.)

ilex  ·  29 days ago  ·  link  ·    ·  parent  ·  post: Uncovering the Seeds of a Post-Lawn Future

Oh, and don't forget that a bunch of the students here are international and very concerned about losing their visas if they speak up about things.

    How would that even work?

You get together a group of grad students willing to lead the organization, you hold town halls 'n meetings with students and you talk to them and learn what's making life hard for them and plan how you can improve it. You talk to your union reps (here, the NEA) and eventually get signatures and form a union. Then you cross your fingers that you won't have to go to court like the union at one of your sister universities did.

    Grad students are generally in a competitive and underpaid market, aren't in that market for a significant amount of time

Exactly, which gives us so very little leverage and basically no bargaining power. If you don't like what the school decides to do, tough shit, put your head down and try to get out as fast as you can.

Just to put things in perspective, at my school we do not get health insurance or any other kind of benefits despite often working jobs or being on grants that forbid us from working elsewhere at the same time. I don't even know how grad students with kids make ends meet.

To hold a TAship and most RAships you have to be enrolled full-time. It used to be that TAships paid about $16k/yr and then you paid $8k/yr back in tuition. And, let me tell you, the cost of living here is cheap but it sure as hell isn't $8k/yr cheap. A few years ago they finally put in a program like most other schools that reimburses tuition if you hold an appointment, but lately they are trying to walk that program back because it costs money. Last year they kinda jumped the gun and people got mad, but this year they've made some smaller changes that slid by unquestioned and I fear they're just trying to carefully boil the frog.

(And, as a side note, for me that $8k/yr pays for "research credit" which translates to "I talk to my advisor sometimes".)

Being on some research grants isn't much better. I've had a couple dry up with pitiful notice (one literally the Thursday before the semester started), and the one I'm on now penalizes you for having money in the bank, then shrugs when you complain that it takes 2-4 months for them to cut you a stipend check and that travel grants require you to basically loan the university a few thousand for a few months and look it's not like I'm expecting a lavish standard of living but it'd be nice to not have to worry about bills and food when I hold a job that requires some amount of postgraduate education, y'know?

    are trying to become the employers that they'd be collectively bargaining against.

For what it's worth, my advisor is vocally in favor of grad student unions.

I'm sure some folks are in grad school to get into higher ed administration, but most are just trying to be professors or R&D industry workers. Professors' hands are often tied with respect to paying their students -- the university sets TA pay and most grants have rules about how money is spent. It's hard for one professor to say, "hey, I need more money to pay my students well"; it's a lot different to be able to work collectively towards raising expectations about grad student pay.

Basically, I just want to be treated like I'm valued a little bit, even before I hold the piece of paper that tells everyone I'm worth something.

(I definitely think that professors' attitudes towards students without degrees contributes to the whole degree inflation thing.)

    It was common to hear analysts say things like, “I just love to solve problems.” But what they were really doing was solving something closer to puzzles. It’s clear to me, for example, that the janitor at my middle school solved problems when she cleaned up trash. It’s far less clear whether analysts at Capital One are solving problems or creating them.

This is a decent description of a lot of engineers. I see so many smart people here disappear into defense contractor work or predatory-ish startups because there are interesting puzzles to solve.

Meanwhile, unionizing grad students here would solve a lot of problems, but it isn't a puzzle, employers won't care if you put effort into it, and that work doesn't help you get closer to graduation, so nobody wants to do it. Far easier to just put your head down and solve puzzles until you've done enough to write up and leave.

    Under the data-driven directives of Capitalism 2.0, you can have a bunch of friendly data scientists who don’t think too deeply about the models they’re building, while tutoring low-income kids on the side. As far as they’re concerned, they’re refining a bunch of computer algorithms.
ilex  ·  31 days ago  ·  link  ·    ·  parent  ·  post: Pubski: October 9th 2019

It's never selfish to ask for what you really need.