s09e06: Three Things
0.0 Context setting
I’m finishing this at night on Wednesday March 10, 2021. As far as I can tell, the children are asleep. I can hear a clock ticking. The cat is curled up in a basket but, notably, it is not any of the cat baskets, it is a toy basket which has been casually emptied of toys by the children. The children do not want to put toys in baskets, the cat does not want to sit in its basket, I am tired of writing the word basket basket basket.
Tomorrow is Thursday, which means it is nearly end of the week, and yet also what is a week, anyway when time itself feels like it’s smeared itself into some sort of miasma about a year ago.
On with the show, then:
1.0 Some things that caught my attention
A Little Printer…
On February 18, 2021, Amazon launched “Smart Sticky Note Printer, Works With Alexa, A Day 1 Editions concept” [sic].
Smart Sticky Note Printer is a little printer (wait for it) that prints little sticky notes and does some useful nice things like automatically cutting them and other “it would be nice if things worked” things, like if you tell Alexa to print a note about something, a note about that thing gets printed. In abstract, this feels pretty good! I can see how this could be useful. (I can also see how I would need to moderate usage of this function in a household with children).
But.
What was interesting about Smart Sticky Note Printer is that it very easily reminded me of BERG’s Little Printer, which went on pre-order on August 14, 2012 [The Verge], 3,110 days before Smart Sticky Note Printer.
There’s a lot that’s different between the two, from what little I can tell or infer from Amazon’s product marketing. Little Printer was also the start of BERG Cloud, which would power all the services that LP could tap into, and you need to remember that Amazon’s first Echo was still two years away. Little Printer has all of that 2012-era whimsy and personality that’s either been beaten into submission in 2021, or after about 9 years, become so pervasive as to be just noise and indistinctive. And you can tell Smart Sticky Note Printer is at the beginning of its life: it’s a small, boxy, squat black thing that looks very much like a Printer, as if Lexmark or someone made a tiny tiny laser printer. It isn’t even covered in fabric.
They’re both thermal printers, so there’s that, and I’m sure there will be lots of Alexa Skills For Smart Sticky Note Printer (really, Smart Sticky Note Printer is just dying for a nickname), but there’s something about an Amazon-ecosystem printer that just feels like it’s going to be… boring? Like I said, I can see how it’s going to be useful. Perhaps I should get one and see what fun stuff our kids end up doing with it.
And yet…
“… Ten years later”
… After noting that Smart Sticky Note Printer came out about eight and a half years after Little Printer, I wrote:
There are a bunch of people reading this for whom I suspect “look at what you thought was super amazing and were doing ~10 years ago, and then be excited about it and do it again now” my feel like a forlorn observation :(
I mean to say: there were quite a lot of things I and my peers were excited about ten years ago and we gave it a really, really good hard try and only got so far (you know who you are, and some of you may even be reading this), and for me at least there’s something very bittersweet in seeing pseudo-vindication. See, you think, see, we were right, this is absolutely a thing that people wanted and would use and look how it’s working now and I want to think we were right and yet also, we were very, very early.
I’ve written before about feeling like you can see a trend and where something is moving and trying to bring it to life and believing it so hard, that of course everyone is going to be on the internet and of course mobile data will be a thing and people will use the internet everywhere and really, social media is going to be all over the place and then before you know it, for example, you’ve won an award for creating the characters for a critically acclaimed TV show on Twitter in the UK, but it was 2009 and there were, like, ten Twitter users in the UK at the time. I mean, it was 2009: the online stuff we did (sorry, I was, uh, speaking figuratively) was sponsored by Nokia. And that was just one of the things, never mind all the other things my peers were doing, like “hey, location data is going to be really important” and so on.
In my jokey ha-ha, not serious of course, only just mildly looking at what were quite honestly some emotionally intense times, I said something along the lines of haha, what if there were Timehop, but it was prompts about your internet career, as in it’s been seven years since you launched $thing *whispers* maybe it’s time to do it again, to which the appropriate response from People Who Were There was along the lines of “I’m in this picture and I do not like it”.
Anyway, slightly more seriously, I thought about the idea of a fund that invests in products/services/companies but ten years after they didn’t quite work out, and crucially, also brings back as much of the founding team as is possible and reasonable, on the basis that you want the people who originally had that insight and, in the meantime, have seen some shit and know what they might do differently this time around. Of course, this stands to have a chance of unlocking a new chapter of emotional turmoil in the founding team, who might feel like they’re now dealing with the most delayed term sheets in the history of venture capital funding. But! I do think this isn’t a bad idea? I mean, there are likely worse ideas, and the idea of paying attention to the 10-20 year career trajectory of people who built influential products and services that nearly succeeded, and then coming back to them and seeing if they’re interested in doing it again? I think that is interesting, and something that would catch my attention.
(Obviously, there are significant structural issues around diversity, equity and inclusion: you’re falling dangerously into the trap of only backing the people who get backing, so you’d have to pay a lot of attention to make sure you’re not doing that, which I prefer to see as a novel, interesting and hopefully effective way of addressing those those issues.
Short version: if you are interested in figuring out how such an investment fund would work, or you know someone who would be and is, as they say, a qualified lead, then get in touch, because I am not entirely joking and, believe me, I have ideas.
When moat building means your call sheets bring all the crew to your lots
Via Hacker News (sorry), I learned that Netflix is setting up infrastructure for Netflix Workstations [Netflix Tech Blog]. Netflix Workstations are described as “remote workstations that allow content creators to get to work wherever they are” which in practice means supporting something called NetFX [ibc], a Netflix platform that lets “vendors, artists, and creators worldwide to collaborate on Netflix VFX content.”
In itself that is not that interesting because you can look at the above para and say: ah, virtual desktop infrastructure is totally a thing in this age of remote working, this makes sense. But what caught my eye was this phrase right at the beginning of their blog post: “To meet this need, the Studio Infrastructure team has created Netflix Workstations.”
There is a studio infrastructure team at Netflix, which means exactly what it sounds like: Netflix is creating in-house tools, software, infrastructure, services, products designed to help make sure that Netflix-produced content is the best content. That means that an otherwise boringly-titled blog post like Making our Android Studio Apps Reactive with UI Components & Redux actually means Netflix created an in-house mobile application to help production crews organize their shooting days through shooting milestones and keep everyone in a production informed about what is currently happening.
There’s another name for this group, which is Netflix Studio Technologies. You can learn a lot about what a team’s doing by looking at their jobs and hiring page, even more when you watch their recruitment pitch video.
I said moat building up top in the title, which is a reference to what happens when a startup grows up and stops being solely inspired by The New York Times Bestseller Airport Business Book: What Would Sun Tzu Do? and graduates to Castled: Business Lessons From Middle Ages Architecture. I can’t remember which of the Tech Bens it was {Thompson, Evans}, but I vaguely remember one of them excitedly writing about the concept of a moat representing your company’s unassailable position in terms of competition (all in service of delivering the best value to your customers, of course). One of the better known examples of a tech moat is Justin O’Beirnes piece from 2017, Google Maps’s Moat, comparing how far Google Maps was ahead of Apple’s maps.
The analogy here is that Netflix is building (er, digging) a moat by not just buying in and investing content, but also the means of production of that content. So when Netflix buys a show like Glow, it’s not just funding the production of the show, it’s also investing in making sure that the show gets made well. What does made well mean? It could mean something like TerraVision, from Netflix’s internal Studio Hack Day from 2019, which is supposed to help filmmakers look for and find locations:
Filmmakers can drop a photo of a look they like into an interface and find the closest visual matches from our centralized library of locations photos. [Netflix Studio Hack Day — May 2019, video]
There’s a good overview in a blog post on Data Science and the Art of Producing Entertainment at Netflix. Now, normally I’d see a title like this and assume that it’s about how Netflix optimizes cover art to get you to watch, or how Netflix is using data science to inform what sort of shows and films it’s commissioning. But, this isn’t just that.
A long time ago, I used to work somewhat closely with TV production, back when TV was terrified about the internet, in the early 2000s, which meant learning a fair amount about how TV production works. It’s a very well-oiled machine! It also has razor-thin margins on the production side. Broadly: you don’t make money producing the content, you make money selling it afterwards. The BBC or Channel 4 would order something from you and you’re pretty much allowed something like 10% as a treat, the expectation being that you make it up on the back-end with things like international sales, licensing and so on.
So, keeping production costs down is important. Here’s the blog post:
During Pre-Production, producers and executives are tasked with critical decisions such as: do we shoot in Georgia or in Gibraltar? Do we hire a thousand extras or lean on VFX? Do we keep a 10-hour workday or a 12-hour workday? Each of these choices can have massive impact on cost, timeline and creative outcome of the project. Traditionally, these decisions are rooted in human experience and intuition. Let us see how one can supplement these with data derived insights. [Netflix Tech Blog]
Again, I can kind of see where this is going: let’s use the power of business intelligence to make better decisions for our business! The blog post goes into a little about how to model and at least check gut-check, experienced intuition about whether you should shoot in Georgia or Gibaltra, taking into account hard-won experience like the kind of team you can use locally, the resources available, and so on. So far, so Let’s Use PowerBI For Making TV Shows.
But then there’s my favorite bit, which is this: some of my favorite project managers in the world are the people who have experience in tv and film production. I met a bunch of great ones working near TV in London, and again in Portland when I was at Wieden+Kennedy. Because the schedule and moving parts involved in a shoot are insane. The idea of using software to improve that, and the idea of Netflix investing in that is, well, it’s very tech.
There’s another video — it’s about 6 minutes long — that shows Netflix’s intentions here, along with another blog post introducing Studio Engineering, whose mission is “to build a unified, global and digital studio that powers the effective production of amazing content.”
The perspective that’s interesting here to me is this: Netflix can be talked about as a net-native entertainment company, and the first version of that is the kind of company that uses the net as a distribution and discovery platform and doesn’t have a whole bunch of baggage to deal with like most traditional broadcasters: infinite schedule time, but has other problems (and opportunities) to deal with, like discoverability and editorial.
But whether something is net-native or not on an axis of “is it on the internet and does it know how to use the internet, most commonly without institutional baggage of not-internet”, seems to me like it’s missing the axis of “is really obsessed with how software can be used to achieve your goal”. This is a really interesting axis. Here’s some random data points that I’m just free-associating in my head about different degrees of incorporating software into vertical integration and how important it is to your core competency:
I think I remember that SpaceX uses its own, home-grown, Windows-based Enterprise Resource Planning system rather than use something like SAP or Oracle. SpaceX is designing and building its own rockets, and yet so are Airbus and Boeing, but Airbus, as far as I know, uses off-the-shelf commercial enterprise software for this kind of thing. This isn’t to say that SpaceX doesn’t use off-the-shelf commercial software, but that there’s something it’s decided is so important it’s going to build it itself.
Pixar and Renderman, and in general the CGI/VFX’s preoccupation with tooling and production, and I am not going to say anymore about this for fear of embarrassing myself in front of certain people I know and saying something quite wrong.
I want to say Amazon here, something about the origin of AWS, but I can’t easily find the best blog post I read about how exactly AWS started. But! I’m reasonably sure that Amazon uses several SAP modules for… some things?
Vox Media, and Chorus, their CMS. Counterpoint: the newspapers like the Washington Post who’re doing the reverse and trying to turn their CMSes into products that other organizations use and buy.
Disney, and all the technology they invest in. One of my favorite YouTube channels is the Disney Research Hub where they post videos of all the… research… they’ve been doing. It’s pretty amazing? The most recent piece of research is titled A Versatile Inverse Kinematics Formulation for Retargeting Motions onto Robots with Kinematic Loops which if I squint a bit is also known as making awesome parks experiences.
Scripto, which came out of The Colbert Show, and which for me is one of the ur-examples of this genre where you get a combination of people who know they should have better tools than what they’re using and also know people who can make it, and then actually make it. Scripto, of course, is also being sold. I mean, there’s Google Docs, and then there’s a thing that is better than Google Docs for what you need to do.
Amazon does have engineering for Amazon Studios and know they have some custom production software they’ve built (hello to you people who know), but I don’t have that information to hand. It is, of course, very Amazon for them to have done this.
I felt like I was on the cusp of a nice, tingly insight here, but as I’ve been writing, I’ve realized that in some ways what I’m illustrating here is organizations doing something akin to Wardley Mapping and figuring out where they can apply software in a way that isn’t yet a commodity, and in that way can become a way to differentiate from a competitor.
Some more thoughts:
I suspect that most companies or organizations might not have a secret sauce part that would benefit stupendously from custom, tuned software. And the holy grail of custom-and-tuned is software that perfectly fits your particular niche and, I don’t know, is whatever magnitude of 10x gets you attention these days. But remember, this was the promise of configurable software in the first place: a versatile tool that you can make yours. So answer me this: why is it that I feel if I make the thing I need, rather than configuring, say, ServiceNow to be the thing I need, that making what I need is better? What sort of context and other inputs go into making that sort of determination? This has a lot of relevance in organizations like government, where a lot of thought goes into what bits should we own and take responsibility for — steward, say — and what bits make sense to be built on top of third parties?
But! Time was, organizations needed to do this because it turns out there was a time when computers did not exist and there was not widespread magic sand that had pictures etched onto it that made it look like there were giant rooms full of thinking metal. During that time, companies invented computers and programming languages and ways to bill for medical services, and it was a point of competition because if you had a computer thing, it meant you could do things better than your competitors. Nowadays, if you have a computer thing, it means… well, really, what does it mean? I mean, I guess you can operate in an interconnected late-capitalism world because people send emails and things (unless you’re a god of thunder who doesn’t have a computer, what would a god of thunder have a computer for?), or it means you can process millions of bits of paper, or substitute bits of paper for people marginally better than if you had to do them all by hand.
Anyway, there was a broadcaster who used to do stuff like this, but I’m not entirely sure how much they did on the production technologies end. It was the BBC, and as a Brit, I am required to thus link to BBC Research and Development [Kingswood Warren building, Wikipedia], one of the examples of which is the BBC Radiophonic Workshop which on the one hand was a sound effects unit, but also ended up pioneering electronic music.
But really, my mind is blown by this: there is a company out there who’s taken a look at how messy, say, production coordinating is, and said to itself: hey, you know what? We could try and make this better, and then is spending money doing that.
That Netflix one ended up being much longer than I anticipated, so I’ll think I’ll call it for this episode.
Thank you so much to everyone who has replied and said “hi”, it was very nice! How are you doing?
Best,
Dan