0.0 Context setting
Wednesday April 21, 2021, and I started writing this on the deck outside in the morning, again, because it’s still beautiful here, again.
I’m still in an in-between projects space, still practicing getting into a rhythm of writing in the mornings and having time in the afternoon. It’s kind of working?
Only about 3 weeks now until our Second Shots.
1.0 Some things that caught my attention
Three things in today’s episode:
- Cursed Takes
- Multiplayer collaborative productivity applications smushed with latency
- Digital isn’t always better than paper
I like to persuade myself that writing these or blurting them out is “funny” or even more seriously and simultaneously somewhat self-defensively, that they’re a sort of bad-idea vaccine, a minimally viable Bad Take that might inspire a healthy immune response. But I also know that’s probably not really true, and that what’s really going to change things is better, implemented alternatives.
For example, I think given the intersection between Diet Libertarian (as well as the smaller Libertarian Classic contingent), what I think is a bit of (misguided? Missing the point?) anarchism, there’s a non-zero chance that there are a bunch of Crypto boosters out there absolutely convinced that the only viable way to defund the police is to use Crypto (i.e. blockchain), of all things. And so, I can imagine someone fictional saying something like this:
“Blockchain is actually the only viable way of defunding the police.”
Or something like this:
“Storing bodycam footage on the blockchain with citizen-minted EnforcementChainTokens is the only way to ensure trust in the police system.”
“Violent state systems like policing and incarceration are ideologically equivalent to fiat currencies. Therefore police reform requires a system of decentralized trust based on Ethereum.”
And then perhaps a horrific, doubling-down end goal of:
“We are looking for a partner city for our Decentralized Policing startup and EnforcementChainToken ICO. Purchasing an ECT in our ICO will fund and invest in participation in a radically transparent, full-trust, decentralized system for public safety from all crime.”
In which case I have perhaps “discovered” one of the worst cursed takes given the last 24 hours.
Caught my attention because:
- Blockchain blockchain blockchain, will people please shut up about blockchain.
- Blockchain tends to be used as a snake-oil techno-solutionist approach to a problem that doesn’t exist, or to one that is completely attacking the wrong problem. Yes, people don’t trust the police. But that’s not because bodycam footage isn’t reliable, or that bodycam footage isn’t publicly available even though in general it should be. It’s because police are killing people, and are racist, and are operating in a system that doesn’t see that as a problem. Blockchain doesn’t fix that, you know?
- How many times do we need to yell that technology doesn’t save us, not on its own?
- I saw this paper, Trust in blockchain-based systems, via the Orange Place. On an irresponsible skim-reading, I liked that it pointed out (opened, even) that “trust” means different things to different people, and in the context of blockchain, two of those meanings are “trust in people and institutions”, and the second of which is “trust in the context of technical systems, like being free of errors and being secure”. The fact that the same word has two different meanings in different contexts is apparently super confusing and has led to lots of misunderstandings.
Make Work Play Again And Ruin Both Things? Also, Where The Internet Comes From
Two things caught my attention today and are a great illustration (ha) of how my brain works and the associations it throws up.
- Figma, a design tool, had its annual conference (because that’s how products and services announce new features now), and announced [tweet] real-time collaboration for up to 500 users on a single design file.
- I quipped that this is like massively multiplayer PowerPoint, which I expect a bunch of people have already experienced in terms of a bunch of people messing around in a Google Doc/Spreadsheet etc. It’s an interesting experience.
- The massively multiplayer connection isn’t just because of an existing association in my head, though, it’s because Figma’s announcement explicitly referenced “multiplayer”:
Just announced at #Config2021: Next-level multiplayer. Now you can work and play with up to 500 of your closest design friends in Figma. [Tweet]
- Earlier today, Mike Davidson complained that “all of the time I used to spend commuting, I now spend signing back into SaaS products.” [Tweet] which, in my head, commuting is a function of time and distance (and frustration). While Davidson is saying the frustration is the signon process (which I imagine involves 2FA and security policies that require authentication every 2 hours), I will obstinately make the somewhat silly connection between time/distance and the distance to a data center hosting the SaaS, so my helpful advice was “Perhaps try living closer to the cloud?”
About 35 minutes later, after stewing, this fictional news story emerges:
“Prineville, Oregon is experiencing massive population growth — thanks to remote workers moving closer to the data centers hosting the SaaS services they use for work.”[Tweet]
along with a fictional quote like this:
“I’ve reduced my ping by at least 10ms, which is helping me get an edge over my peers. Plus it’s cheaper to live here.”
and then smushing it all together:
“More tech workers are planning line-of-sight microwave links to the data centers hosting design applications like Figma.”
Here’s some more of what’s catching/triggering my attention in this melange:
- Direct line-of-sight microwave links are what’s used by financial traders, and are part of what enabled/is required to make high frequency trading work. Ingrid Burrington wrote about these back in 2015 for The Atlantic [A Network of Fragments] and anyway, you should also read her book, Networks of New York. There’s more about high frequency trading in Michael Lewis’ book Flash Boys.
- Look, here’s how important time is: Kim Pallister has anecdata about a trader asking in the early days of high-frequencyt trading, “Where is the internet backbone, and can I buy a house on top of it?”
- It is, of course, dumb that you’d want to move closer to a datacenter just to be a bit faster at using a tool that your peers are also using (you’re optimizing the wrong thing, duh), and yet humans are very good at doing dumb things. Occasionally doing dumb things leads to orthogonal discoveries. Fuck around and find out, as we say. It is doubly dumb because internet infrastructure companies are well aware that centralized data centers are by definition geographically distant from, well, other places, which is why content distribution networks exist, and why Intel wants to sell you more processors to go not just in your existing big datacenters, but for you to build more datacenters in more places, i.e. “the Edge” which also seems weird because… do clouds have edges? Is there a strict place where a cloud has an edge? (Asking this question is also dumb because of course there is a strict boundary in an architecture, I think?)
- Anyway, back to gaming. Back in the 90s, if you played twitch first person shooter games — the “twitch” comes from the games requiring fast reaction times and good eye/hand coordination — then you’d be super annoyed about LPBs, low ping bastards [Urban dictionary]. In fact, I think LPBs were a thing even before playing over the internet was a thing, and you’d be dialing into banks of modems to play Quake over a 28.8k/56k connection?
- Day traders hate low-ping bastards. Time is money, right?
- I don’t know, if you’re a super-competitive person, then once you’ve managed to choose to live somewhere that has a wide internet connection (i.e. download speed, you don’t want anything less than fiber and a gigabit/s right?), the next thing to get one up over your rivals is to get one that has fast internet connection to the locations that matter to you. And that has everything to do with physical distance.
- And why did I pick Prineville, Oregon? Datacenter locations are as much real estate deals as they are tax incentives but they are also about the cost of electricity. What’s interesting about Prineville? It’s got cheap electricity because back in the 1930s, the Federal government got the U.S. Army Corps of Engineers and the U.S. Bureau of Reclamation to build the Federal Columbia River Power System. Here’s a handy PDF from the Bonville Power Administration explaining How the Federal Columbia River Power System works for you.
- It is ironic, in a way, that companies like Facebook are benefiting from 1930s federally-funded infrastructure for cheap hydroelectric power, huh?
- Anyway, that’s why a bunch of datacenters and hundreds of millions of dollars of “investment” are in Prineville, Oregon: cheap, green power plus your usual tax breaks.
- The dams aren’t without cost, of course, both to the environment [Columbia River System Operations Final Environmental Impact Statement, BPA, July 2020] and to native tribes [Smithsonian Institute, and this 2015 paper, Tribal Water Rights: Exploring Dam Construction in Indian Country]. Another example of how the internet is built on stolen land.
Meanwhile, the Figma->tool->gaming association brings up interesting (to me, anyway) directions like:
- Can you have campers [Urban Dictionary] in tools like Figma? People who don’t play fair by just loitering around strategic locations to “win”? What would camping look like? Is that a good thing? Could camping be like figuring out where “good” work or ideas are coming from, and then taking credit for work that’s not yours?
- Multiplayer games have lobby services: you can play by having a team (and those teams evolved into the teams that now play in esports leagues, for which see the next paragraphs), or you can just pick up and play on your own. If you want a game to be fun, though, you want to play with people roughly your own ability otherwise you’re just going to lose a lot. So lobby services are much about matching people for, hopefully, a fun game. Is there a need for a lobby service in SaaS work/productivity tools? Is there a collision here with something like intranet tools for talent/“human resource” discovery? All those wikis and tags for “finding the right person for x” in a giant multinational or, to be honest, any sufficiently large collection of human beings were you can’t remember enough about what everyone does. Is there anything to learn from how the problem of player matching is approached in games? This is super interesting!
- You know, let’s just go for it. If you’re really good at Figma, then why not just have a Figma eSports league? There’s already Battledecks, which is hilarious for certain values of humor. But, you know. Competitive Figma. You know it’s already a matter of time before LinkedIn is integrated with Xbox Live.
- Actually, team-based hackathons are esports for… whatever it is hackathons are for. They’re certainly not… great.
- Oh, here’s another one: Battle Royale Figma, where the canvas just gets smaller and smaller and smaller and smaller until the One True Designer’s cursor is left standing.
Ben Hammersley [tweet] and Anil Dash [tweet] both chimed in later to remind me about Layer Tennis, which used to be both a website (perhaps initially community run, and then unsurprisingly and smartly sponsored by Adobe) and then a live event (Ben’s got a photo from 2009). Which is just a good reminder that any sufficiently specialized profession can totally have a sport. Just like how Republican men (and Democrats, I’m sure — being a misogynistic bastard is not dependent upon political party alignment) in Congress have a sport of rating women.
Not better than paper
In my feed via Emily Tavoulareas, a piece in Slateby Hana Schank and Tara Dawson McGuiness, What Happened When the U.S. Government Tried to Make the Immigration System Digital, of which this quote caught my attention:
A common mistake people make when trying to improve or modernize something is believing that digital will always be better. But digitizing a broken paper process doesn’t make it better. Sometimes it even makes it worse…
… [ELIS 2] faced an even bigger problem in the agency’s assumption that digital would unquestionably be better than paper.
(ELIS 2 was/is the replacement effort for the previous technology modernization project, ELIS 1 (which after 11 years managed to digitize two out of 94 forms, at the cost of $1 billion).
Caught my attention because: I’ve seen this happen before, and I bet you have, too.
One of my most recent examples is a ServiceNow workflow implementation of a previously paper/ad-hoc digital based process for approvals. The current implementation doesn’t allow for fallbacks, so if someone is out/sick/on training or not available to sign for whatever reason, and a substitute hasn’t been configured, or the substitute(s) aren’t available… the workflow cannot proceed. No “oh, I’ll just run this over to the next person and get them to sign it”. You’re just stuck. It may even have been the position that you couldn’t go higher up the chain to get an override.
Another frequent thing I’ve heard in interviews is that people qualify current systems in a “well, it’s not like we can go back to paper”. Digital systems usually allow for increased throughput (ish?) or at least allow for people to handle volume that they couldn’t handle if it were all paper. But just like Schank and McGuinness’s article says:
In the case of ELIS, the team members believed their job just was to take what was on a form, digitize it, and call it done—there was no place adding sticky notes or placing the folder in a special filing spot based on where it needed to go next. They did not factor in the colossal amount of filing, categorizing, and handwritten note making that the people processing forms did on a daily basis.
I couldn’t agree harder. Digital isn’t better than paper. Assuming that it is is the kind of thing that leads to procurements that essentially boil down to “we want some magic that will make everything better and all the hard things go away”. The world doesn’t work like that, and anyone who says that they’ll sell you that is lying.
I last wrote about this in 2017 [s4e10: Stages of Transformation] (sigh), quoting the wonderful “If you apply digital to a thing that’s broken, you’ll have a broken digital thing”, but that equally applies to broken digital implementations that don’t take into account how people actually work, and just examine what abstracted business processes say should happen instead of what does happen. Because what happens is how things happen. Again: the business process diagram is not the territory.
That’s it for today. ~2,500 words, which is comfortably under Buttondown’s limit.
The plan is for another episode to come out this Friday.
How are you doing?