s5e09: At Any Cost
0.0 Station Ident
1.0 ... Caught My Attention
(An assortment of things that have...)
It feels like everyone in my bubble is required to have an opinion about Facebook. It does not follow that you, reader, need to care about it. Feel free to skip this section if you just can't even deal with this shit right now.
Casey Newton has been writing The Interface, an evening newsletter about Facebook, social networks and democracy and it's the best daily roundup on... everything Facebook right now. There's a lot that I'm just not writing about Facebook because, well, there's a lot *to* write about Facebook and I think I'm pretty OK with not devoting that much brain time to Yet Another Take On Facebook And Cambridge Analytica And What It All Means.
But! Some of the secondary effects (like the almost (blackly) comical pile-on happening to what I think is now approaching Apple-levels of "embattled, beleagured" corporate descriptions) are throwing up some pretty interesting stuff. Like an internal memo (of which the *word* is funny, just because... memos are yes, a business thing, but... not how, er, digital native businesses necessarily work?) posted by Andrew Bosworth leaked to Buzzfeed News[0]. The memo stakes out the position that Facebook values growth above, well, anything else. Look, I'll excerpt it here:
“We connect people. Period. That’s why all the work we do in growth is justified. All the questionable contact importing practices. All the subtle language that helps people stay searchable by friends. All of the work we do to bring more communication in. The work we will likely have to do in China some day. All of it,” VP Andrew “Boz” Bosworth wrote.
(After publishing, Boz took to Twitter to point out he [didn't] agree with the post today and I didn't agree with it even when I wrote it."[1].
But, you know what? I read Boz's post (sorry, memo), and what it reads like is your regular corporate rhetoric to justify a position, and in this case, that position is: "Hey, everyone. Growth is, like, really super-duper important. It's, like, a first-order priority, right?"
My (long, belabored, oft-repeated) point here is that *everything else* about what Facebook does (e.g. the connection map that fills in the world[2]) and the way that it's behaved and its rhetoric and communication over time has *always* emphasizes "connecting people", and that the key is connecting people is a means, not an ends.
When I've written about this before, I've used the somewhat facetious device of poking fun at the agile user story ("as a user, I want x so that I can y") because it provides context about a means through the addition of a goal. One example: "As a user, I want government protected free speech, so that I and those who come after me can pursue life, liberty and happiness."
In this formulation, my intuition is that Facebook never *really* got good at saying why connecting people was good. They just double down, by saying things like "We connect people. Period."
I've written about this before but haven't referred to it here and I'm kind of saving it up for a longer newsletter post because I suspect there's some people in this particular venn set that aren't in that other venn set: but, in No one's coming - it's up to us[3], I wrote about how some axioms of "doing things on the internet" like fervent, unquestioned belief in Metcalfe's Law[4] about the value in comms networks leading to, well, not so great outcomes.
Boz's post says, hey, it's OK (even going so far, you might say, that it's unfortunate) if in trying to connect people someone dies due to bullying or a terrorist attack coordinated on their platform. Why is that OK? Because in a utilitarian calculation, the value in connecting outweighs the bad. And, taken to an extreme, what's one death when compared to the potential good ("Maybe someone finds love. Maybe it even save the life of someone on the brink of suicide")?
Some other observations:
* Facebook has formerly been a stupendously tight ship. Now it's leaking, and I'm of the opinion and intuition that there'll be more leaks if things don't change. Quite how they need to change (ie: what, and how quickly) is a bit murky though, right?
* Boz says that he wrote the post to encourage debate around hard topics ("we have to be able to consider even bad ideas, if only to eliminate them. To see that post in isolation is rough because it makes it appear as a stance that I hold or that the company holds when neither is the case.") to which I immediately start thinking: OK, if you "care deeply about how our product affects people and [I] take very personally the responsibility I have to make that impact positive" then... where's the evidence of that? Where's the *inverse* of this straw-man memo that says: we seek to maximize connection because *these* things are good and *those* things are bad - ie, the bullying etc, and it's as important to minimize the bad things as it is to increase the potential and probability of the good things. Because sitting here (and also having sat on the inside, to some degree!) I *always* saw [citation needed, confirmation bias alert] things framed in terms of growth and positive *without* much, if any, consideration of the potential for the negative.
* If Boz wants to reject the narrative that people are projecting on to Facebook (because Facebook is practically a vacuum and has in the past had a *very* well managed comms strategy in terms of execution), then one way of doing that would be to... provide *evidence* that rejects the narrative! And one hard piece of truth might be: you know what? There is not an opposite equivalent to the straw-man memo. We thought connecting people would just be a de-facto good. And maybe it's not? Maybe it requires more work than that? So show us on the outside that leadership does encourage that view. Where's the post (sorry, memo) for that?
[0] Top Facebook Executive Defended Data Collection In 2016 Memo — And Warned That Facebook Could Get People Killed
[1] Boz on Twitter: "My statement on the recent Buzzfeed story containing a post I wrote in 2016… "
[2] You know the one, only, well, it only fills in *bits* of the world, right?
[3] No one’s coming. It’s up to us. – Dan Hon – Medium - this is the thing I am slowly, methodically, turning into a book (or rather, I had the idea for a book and this talk and then post was an attempt at, uh, the minimum viable product)
[4] Metcalfe's law - Wikipedia
2.0 Grab Bag
* Game engine sizzle reels for GDC, the Game Developers Conference are a bit like a tiny tiny slice of the SIGGRAPH papers video. Anyway, here's the Unreal Engine sizzle reel, which I felt was a bit better (showed more interesting stuff) than the Unity one.
* And inside that Unreal Engine sizzle reel was this gem, which if I describe it in this way for British readers of a certain age, I bet you'll totally click right now: Knightmare, The Adventure Game and The Crystal Maze, but rendered in realtime through greenscreen using Unreal for a 42 minute tv show that's already finished production - Lost in Time, by Future Universe and Fremantle Media.
* I'm researching dark UX patterns for my book and a related project. If you know what that phrase means, who beyond the usual suspects do you know of who's been doing long-term work in this area?
* Graffiti that fools machines is now a thing (we've imagined stuff like this for a while, remember poor Bobby Tables)
* Last time I built a PC must've been in the late 90s, early 2000s and over the last 5 years my interest in having a PC (and, thus building one) again (something to do with feeling like I need to know more of what's happening in games and VR) has slowly crept up. Which makes PC Building Simulator really interesting because building a PC does still feel totally stressful (not just the building part, but the picking-the-right-parts part, which is like some sort of solve-this-combinatorial-problem with multiple dimensions of constraints, a kind of hellish paradox of choice combined with a desire to optimize - I mean, this is just for playing *games*! for god's sake). But what I found *even more interesting* was the conceit of PC Building Simulator, which in the review is of "uncle Tim suddenly leaving me in charge of a small computer repair shop" and my reaction is omg that is a fantastic writing prompt along the lines of "hey, let's check out what's in the basement/attic/abandoned house etc" and ties in with the "oh, I just found a phone". At least for me, I felt like the stories you could tell from a beginning like that were deliciously tantalizing.
* In book research, a reminder that some tech organizations have manifestos: here's Mozilla's which no doubt was ridiculed at the time ("the internet is an integral part of modern life") and now looks nicely prescient.
* Here's a way AI can help create AI-Physicist/scientist centaurs: teaching machines to spot essential information in physical systems which as a naive, interested layman I'm interpreting as "machines can help us spot the interesting things to pay attention to that may help us understand/explain underlying system behavior"
* The big story is that we found one (*one*) galaxy without a dark matter halo when every other [I think?] galaxy does. Bad Astronomer Phil Plait has a good writeup as to why this is interesting and the original paper is in Nature. *That* is interesting enough in itself, but what's *also* interesting is that the discovery was made using something called the Dragonfly Telescope Array (paper here at arxiv) which is a sort of amazing MacGyvered telescope made of 24 (originally 8) Canon telephoto lenses with a nanoscale filter removing scattered light on a commercial robotoic mount controlled over USB by three Mac minis, some custom software including a bash script (that you can get on github!), and some commercial software from the robotic mount to make sure you point the array at the right thing. Plait points out that this new telescope array *sees things that other telescopes don't see* and before it found the *one* [citation needed?] galaxy that doesn't have a dark matter halo, it had already discovered a kind of galaxy we never even knew existed. This is amazing science. And people say the pace of discovery has slowed down! (Of course now I want to go and make one of these things so I can pretend to be Ellie Arroway. I am OK TO GO.)
* I cannot remember who sent me this via Twitter, but here is a terrible government software project in France. Spoilers: 6m loc, 20 developers, 35 managers and an example of one bug that flagged "already solved"/WONTFIX: the load data from CD-ROM function wasn't broken, it was working as designed when it took 7 days to import 700 megabytes. I am happy to report that none of the projects I've worked on have been this bad.
* There's a remote code execution vulnerability in Drupal and one of my first thoughts is how many unpatched government websites there will be.
* I am slowly making my way through Andy Clark's Surfing Uncertainty outlining the theory of predictive processing as a sort of theory-of-everything for cognition, intelligence, consciousness and so on. One connection I've made is that if our brains work through a combination of top-down prediction combined with bottom-up sensory data *at every level* and that what gets propagated upwards is the *difference* (ie you pass up/along the chain only the bottom-up data that clashes with the top-down prediction at that level, which concept I have set next to "surprisal"), then... does this partly explain the aggressive attention race? Things that grab our attention are when what happens *doesn't* fit our model, so then does that explain the ratcheting-up of headlines and claims? If I keep going with this thought, I don't entirely like the conclusion. More on Andy Clark in the recent New Yorker, too (oh god this newsletter feels so pretentious).
* From a regular trawl of IEEE Spectrum, I did not know that these days lasers are made using fiber optics.
* There will be a super interesting interactive fiction convention in Boston in Summer 2019 that I wish I could go to.
* I saw a render of the Nexus, a concept SSTO orbital launch vehicle by Convair and thought it was fictional from an imagined 1960s (in a way, it was) but it turns out that no, it was totally a *real* concept and it would've been ridiculous. Speaking of space art, Don Davis who did all those O'Neill etc paintings you love is on Twitter now.
* I asked on Twitter if there was an information/data/personal data equivalent to fiduciary responsibility and oh wow here's a whole set of articles in the UC Davis Law Review about it: Information Fiduciaries and the First Amendment from Jack Balkin. Sorry, a bit of Cambridge Analytica crept in there.
--
OK, the end. Thank you! Have a good weekend.
Cheers,
Dan