s06e09: What’s the target?
0.0 Sitrep
In an act of radical transparency, I will share that I accidentally found a whole bunch of what appear to be official instrumental tracks on Youtube at about 3am in the morning last night, and this is the official reason why I have been listening to, er, the studio instrumental track of Christina Perri’s A Thousand Years Pt. 2 on single track repeat (look, I just like what the strings do).
I’m on an Airbus A320-200 at 10.3k meters with a ground speed of 776km/h. Even though the inside of this plane looks new, the in-flight entertainment system say it’s ship number N333NW, which Planespotters.net says is over 27 years old (it used to be owned by Northwest Airlines). When I search against that ship number I also find about 2 years ago, this plane accidentally landed at Ellsworth Airforce Base instead of Rapid City Regional Airport because the NTSB helpfully includes ship numbers in its reports. Anyway, the reason why the inside of the plane feels new is because Delta refurbished all of its A320s starting in 2015 and because I’m the kind of person I am, I’m interested by the fact that the interior was designed for Delta by Zodiac Aerospace and that you can learn about their cabin interior solutions.
It is kind of stupendous that I can find out this information *while I am on the plane* and for relatively little money.
1.0 Things That Have Caught, Etc.
1.1 Digital Classical Musicians
You’d assume the New York Times piece When Classical Musicians Go Digital would be about classical musicians using iPads because a) electronic, high-resolution copies of scores are, broadly, more convenient than carrying around all that paper; b) “pencils” are pretty good these days and c) digital files are easy to share. Which is not new news. There were a couple of parts that really stood out to me, though.
First: Mr. Sheppard-Skaerved, a violinist and scholar at the Royal Academy of Music who organized (“curated”?) a centenary tribute (exhibition?) to Yerhudi Menuhin says that “the advent of the mass-produced graphite pencil in the second half of the 19th century coincided with profound changes in the way a performer engaged with a musical text. The generation of musicians who benefited from the new tool — capable of making durable, but erasable, markings that didn’t harm paper — were, he wrote, “the first where practice was aimed at perfection of execution, and not developing the skills for real-time extemporization on the material in front of them, or improvisation ‘off book.’””
... which feels like a great reminder that the simple pencil is a tool, and that the technology and tools we grow up with (obligatory Douglas Adams quote), are treated as part of our environment and “natural” and less things that were strictly “invented”. There is a thought here about how tools in general might be additive? The sphere of influence of a pencil might, I imagine, be relatively easy to understand. When you get to the technology of a “book”, never mind the “mass market paperback” and the attendant moral panic of people sitting around all day *reading*, these things still feel [citation needed] relatively comprehendible. The possibility space is not too big. Although, books, as tools: instruments of knowledge that effect the spread of knowledge. Very terrifying indeed. Anyway, I get ahead of myself: the simple thought was this, a pencil as a tool might be easy to understand. It has relatively few degrees of freedom or influence. The networked, mobile, sensing processor? Significantly less so, never mind when technologies are multiplicative, combinatorial and not simply additive. And, just based on that thought about how tool combinations feel a bit exponential, I’m wondering: where in nature might we experience exponential growth? I mean, it doesn’t make *sense* for exponential growth to exist because... it would die out. You’d see a quick bloom and a die-off. Perhaps this ability to only understand exponentials *just a little bit* is both a fundamental part of our creativity and imagination, and also a part of the problem. We know enough to step past simple addition, but when it comes to fully understanding exponents, I can handwave about the story dating back to 1256 (according to Wikipedia, at least) about placing a single grain of {rice, wheat} on the first square of a chessboard and doubling it until you get to the last, 64th square. That this is such an old, re-created story, practically a fable, feels like [citation needed, etc] a basic reminder of something we don’t instictively understand. There is a very dry piece of Wikipedia writing in the opening section of that article, stating “the total number of grains [on the last square] equals 18,446,744,073,709,551,615, *much higher than most expect*” (admiring asterisk emphasis my own).
Anyway, we were talking about the pencil and its effect on classical music, and the idea that the pencil was a tool that enabled (guided, maybe?) repeated revisions at getting to a sort of perfect execution of a score and I promised a second bit.
The second bit is this, and kind of has nothing really to do with iPads. It’s that archivists have been diligently digitising their collections and making them available online, which means that musicians can read music directly from (in some cases, I’m guessing, reproductions of original) manuscripts. Deciding how to perform a piece can be as much a historical, opinionated exercise, and the musicians interviewed in the NYT’s article talk about how in the olden days (ha), they’d would “wait until I could go to the library to seek out” early editions and manuscripts for clues to to a composer’s intentions. Another violinist talks about how he can retrieve late Beethoven quartets directly off manuscript and display and mark-up on screen - which can be quite different from the definitive printed editions that are treated as a sort of ur-text. I mean, this is kind of amazing and wonderful and brilliant, right? Classical musicians can get access to reproductions of original manuscripts for a relatively trivial amount of money (the Beethoven-Haus archive charges 6 euros for a 600/400dpi hi-res work, plus a 10 euro processing fee).
I mean, this is basically the device/content deal, right? You’ve got two sides of an equation:
1) A portable device that has a) natural input (high DPI and refresh rate pen-type input), b) “good-enough” networking, e.g. WiFi; and c) a high resolution display that’s comfortable enough to read from
2) Content that is a) exists, b) is relevant / meets the needs of specific users; c) is of high enough quality (high dpi); and d) relatively affordable
But! The portable device probably wouldn’t have been affordable enough if it were solely targeted to the niche audience of classical musicians, whom, I am reliably informed, are not necessarily a class of consumer blessed with tremendous amounts of disposable income. In other words: people have tried to create networked LCD screens for e-scores in the past, and part of the problem is you’re just not going to get the scale. What feels like a difficult conceptual sticking point is that the *opposite* happened, which was *instead* of identifying a specific, niche audience for portable networked devices with good screens, *somehow* they instead went mass-market enough to be... affordable to classical musicians?
And this is where, I think, the internet argument comes in. By producing a (somewhat) fair playing field of low-cost digital distribution (significantly lower cost than the previous practice of having to actually print out copies of scores and then have them hang around on inventory to sell them), then you’ve got another piece of the puzzle/value chain filled in. Sorry for using the phrase value chain. But it is a chain of things and there are definitely things that people value in it.
This feels a bit like the whole 3D television fad and a point that I keep coming back to (which probably says more about me): 3D televisions felt like they were going to be such an obvious slam-dunk consumer failure *compared to the other thing that was happening at the same time*, which was TVs that you could hook up (whether via HDMI or built-in) to the internet. The relevant content for a connected TV is *any video ever produced*, whereas the relevant content for a 3D TV is... the tiny subset of 3D content. Why bother, especially when the premium for 3D support is so high?
Anyway. One other point that I think is important here is that the article implies musicians can get access to digital manuscripts at relatively sane prices (e.g. in the 10s as opposed to the hundreds) so, say, within an order of magnitude of existing print prices — which, I should point out, are for print scores *that don’t exist* or that can only be seen but not even touched.
I would *like* to think that the archival institutions in possession of such manuscripts see more value in ensuring the spread and availability of those manuscripts rather than, say, deciding to go down a path of MAXIMUM VALUE EXTRACTION due to some sort of combination of structural and supply scarcity (there’s only one score, we have it, you don’t have it, etc.). I think there is a Lesson Here, etc.
Anyway. Of course the pencil is interesting. A piece of chalk is interesting; anything can be interesting if you’re curious enough.
1.2 What If Agents, But With More A.I. This Time
I saw that Mark Gurman over at Bloomberg is reporting Essential Products (Andy Rubin’s thing after running Android at Google) are developing a new phone that “will try to mimic the user and automatically respond to messages on their behalf, according to people familiar with the plans” and I admire Gurman’s strategic (and, I would like to read, somewhat shade-throwing) insertion of qualifier “try to” between “will” and “mimic the user”.
The idea is that Essential will {handwave} make some artificial intelligence software {end-handwave} that will let people “book appointments or respond to emails and text messages on its own”, but don’t worry, you’ll also “be able to make phone calls from the planned device”, to which last point I am seriously, like, why bother?
In a quote that is very possibly out of relevant context and thus a *stunning* example of hubris and lack of being in-touch with current concerns and sentiments, Rubin said in an interview with Bloomberg last year that “If I can get to the point where your phone is a virtual version of you, you can be off enjoying your life, having that dinner, without touching your phone, and you can trust your phone to do things on your behalf,” and that part of the reason for doing this is to “solve part of the addictive behavior”.
On face value, constructing a somewhat Turing-test passing intelligent agent to transact on your behalf *as a way to deal with intentionally designed addictive behavior in consumer mobile technology* feels a bit like, I don’t know, pre-emptively amputating the feet of anyone who’s pre-diabetic, or pre-emptively fundamentally altering urban planning and human behavior to reduce and simplify the problem-set that autonomous cars would need to solve in order to not kill people.
I mean, really! I am probably going to get some criticism for extrapolating what Andy Rubin may have mentioned offhand in an interview about long-term future intelligent agents proxying (autocorrect originally turned proxying in to *parodying* which... good call, autocorrect) you and representing you to a sort of Uber-intent floating over Silicon Valley. But - it’s not like people have started to notice that GMail’s Smart Reply feature is... a bit twee? Anecdotal evidence: emails worrying about Dr. Blase with suggested replies being something like “That’s awesome!” (Aside: Smart Replies now apparently account for 10% of all emails sent on Gmail)
On the one hand: 10% is a lot of emails! On the other hand: 90% is even more!
And of course, everything old is new again. Because what Rubin is pitching (at a high level, at least) is what General Magic was pitching in the 1990s and the kind of thing that, unsurprisingly, Douglas Adams got really excited about. It should not surprise you that Rubin was at General Magic and worked on Magic Cap, the agent-based operating system that had to invent an internet that didn’t exist yet, or that Rubin was also the co-founder of Danger which was the one thing that made texting cool to American teens.
(Here are the Telescript white papers describing the high level architecture of Magic Cap, the agent-based system design and the scripting language. They are one of those things that feel like the description of superstring theory: they fell out of the future into our hands and we didn’t know what to do with them because we just weren’t ready yet). And if you want to know more about General Magic then you should probably try to get to a screening of the documentary; there’ll be one at SFMOMA on November 3rd that I’d dearly love to go to, and there are others scheduled around the U.S.
Anyway, there’s a bunch of history here. Rubin is one of those who has been directly involved in trying to do *exactly* what he’s pitching now, only *twenty-six* years ago. Back then, Magic Cap and Telescript would allow explicitly programmed somewhat autonomous programs to contract and act on a user’s behalf inside of specified parameters, explicit intent.
The dream is still the same: why can’t I have an intelligent assistant that does stuff on my behalf? Twenty six years ago, I’d be more intentional about it, the hype right now (and I am totally in the camp of using what people want to deliver what they need) is to use deep learning and A.I. which, fine, sure, use that too as a marketing and getting-interest hook.
But I skirted over a really important point in that previous paragraph. My understanding of how Magic would’ve worked was that it didn’t rely on mining a crap-ton of personal data, it relied on you *telling it what you wanted to do*. The potential difference with what’s being pitched right now *and* the difference in implementation is that mimicking the user requires the sort of large-scale data collection that we’re rightly suspicious of these days.
Look: if Facebook, a company, not a person, can commit the egregious “error” of using a communication method provided with the specific intent of *securing their platform* for *advertising* in a way that to people might feel representative enough of an entire industry, then even if we didn’t have episodes of Black Mirror to slyly reference, we’d have reasonable grounds for suspicion and distrust.
Everything is connected, of course, so even though I’d been studiously ignoring reading Wired UK’s piece about Stripe, it all came back via a Metafilter post calling out this crucial point - blockquote time:
“When Berners-Lee and his team were building the world wide web and designing HTTP and HTMP standards, they included error codes such as “500: internal server error”, or “404: page not found”. In the early 90s, they were trying to realise Licklider’s vision and setting out the rules for how we were all going to interact over this information network. One long-standing error code is “402: payment required”. The original intention – the reason 402 is reserved for future use – was that this code would be used to transact digital cash or micropayments. It has never been implemented – and the Collisons argue this is the reason tech is turning from an equal access opportunity to an oligopoly controlled by five companies now worth more than $3 trillion.”
Magic Cap and this intelligent agent architecture of autonomous programs wandering around the “internet” doing business for you, like all of the stuff in the AT&T You Will ads: paying to have your car fixed, getting concert tickets, buying flowers and stuff are of the era that didn’t anticipate an attention-based economy. As far as they were concerned, the internet would be, in a general sense, business as usual. Sure, it’d be democratising and the network would flatten access and the whole leaving the meatsack body behind and mind floating in an egalitarian network of pure thought. But people would still pay for things. There would be a monetary exchange of value.
That original vision of Magic Cap *doesn’t work in today’s attention economy*. Where’s the advertising? Where’s the opportunity to get in front of my eyeballs? If these are indeed independent programs with partitioned, segmented user data, then… where’s all the aggregatino? Where are all of the native content opportunities?
I mean, if you suddenly built and deployed a way of doing business that didn’t need all of those things then sure. But… that’s not what we have right now? What we have right now is a result of this weird faustian bargain where the high-level strategy went a little bit like this:
Objective: get all the users.
How do we get all the users?
Well, ideally users don’t have to pay for our software. Users having to pay for software is an adoption barrier.
So, I guess we should make our software free?
Yes, we should totally do that.
Wait, that’s how we used to make money.
Well… now we have all this information about them?
Okay, can we use that to sell advertising?
Sure?
Great, now we’ve got lots of users and… now we have infrastructure costs. I guess we’re going to have to sell more advertising?
The above is essentially a re-telling of Maciej Ceglowski’s concept of investor storytime which I have to admit at this point feels a bit like an early version of the “hey, what happens if you put a grain of rice on the first square of a chessboard and then you double it for the next square” only it’s “children, gather around the nuclear pile that is the source of heat in our post-apocalyptic landscape and see if you can solve this riddle I now pose: what happens when someone gives you money on the promise that you will make even more money” and this answer is something like “what if the hedonic treadmill applied to economic growth”, in which case is the answer: “we’re fucked”?
Anyway. A vision of an intelligent agent that works on your behalf to do what you ask it to do and spends your money in ways that you approve (oh how in any possible way could anyone have a problem with this) is much more difficult to work with (maybe?) when you don’t pay with money, you pay with Attention. Does your intelligent agent then become some sort of intermediary? Is the deal: hey, I’ll go do this stuff for you so you don’t have to do it in exchange for viewing these short messages from our sponsors? When will there be enough sponsors? What are enough messages?
And all of this is orthoganal to what you’d have to do, to even get to something like the difficulty of producing a good-enough Alastair Reynolds gamma-level simulation (2000), or the simulation of you that would read your email for you that Greg Egan would write about in Permutation City (1994) or the agent proxies Karl Schroeder would write about in Lady of Mazes (2005).
1.3 Mr. Ocean, the purpose of this meeting is to determi-
Here’s a throwaway thought:
A legacy technology modernization and replacement project portrayed as an Ocean’s 11-style heist movie.
(I think this is because I was talking about government technology and legacy modernization projects at yesterday’s Code for Canada fellowship onboarding workshop (sigh. Lots of long words there) *and* because I’d seen something to do with the Met Gala *and* I’d recently watched Ocean’s 8.)
This is funny to me (ha ha, my sense of humour) because:
* Incredibly complicated goal dismissed as “impossible”
* Incredibly complicated goal dismissed as “impossible”
* Requires expert people skills as much as technical proficiency
* Complexity requires meticulous planning, but...
* Plans A through Z invariably go out the window because the job always end up more complicated than even the well-prepared anticipated, so also requires comfort, familiarity and expertise at improvisation
* Numerous interests invested in preventing success
* Professional team management and co-operation and clear communication
* Easy temptation to target side goals that must be resisted
* Some people actually think doing this is "fun"
* If all goes well, nobody notices
I mean…
17. INT. CANTER’S DELI - NIGHT
They’re sitting in a booth over coffee.
RUSTY: How’s Tess?
Danny stares at him: next subject, please.
RUSTY: Alright. Tell me.
DANNY: It’s tricky. No one’s ever done it before. Needs planning, a large crew.
RUSTY: Exploits?
DANNY: Not zero days. It has to be very precise. There’s a lot of security. But the take…
RUSTY: What’s the target?
DANNY: Eight figures. Each.
RUSTY: What’s. The. Target.
DANNY: (deep breath, then): When’s the last time you were in Northern Virginia?
RUSTY: What? You wanna knock over a VA core data center?
Danny puts down his coffee. And shakes his head. And lifts three fingers. Then a pause. Five fingers. Thirty five data centers.
Rusty must put down his coffee, too.
What’s not to like?!
Also, this is me belatedly realising that heist movies are like competence porn! The drama isn’t in wondering *whether* they’ll get away with it (they should get away with it), the drama is in seeing *how* they get away with it, which in a way is like all the behind-the-scenes Aaron Sorkin Sports Night, West Wing stuff: here’s this big thing we’ve got to pull off and all these people are involved and it’s all got to go just right. How fun!
1.4 More Things That Caught, Etc.
* Charlie Stross asks what the blindspots in contemporary science fiction might be, with four suggestions: 1) near-term, late 21st century stories (e.g. from the 1950s to 2001); 2) frustration at the majority of visions of the future being “what if capitalism, where people have “jobs” and work for “employers”, but with more computers, and in space” instead of something different (for which I highly recommend Cory Doctorow’s walkaway); 3) what if the patriarchy goes away (for which I highly recommend Naomi Alderman’s The Power); and finally 4) a super easy one, “what does a future social contract and legal system look like?”
There are some interesting points in the comments, one phrase I liked was “we are the asteroid” in reference to the slow-moving threat that comes out of nowhere and happens to be an extinction event, in exactly what we’re doing with our ignorance of externalities, climate change and so on, just because I find it funny and it’s a tiny step to the “the asteroid was already on the planet!” / “the caller was inside the house!” etc. I mean sure, the idea of the twist that you’re the architect of your own foretold disaster isn’t a new one, but I just liked the juxtaposition of the “we’re the asteroid” because, well, we know what asteroids look like and they’re relatively big and flamey or “the size of Texas” and worry Billy Bob Thornton.
* via Scott Alexander, there’s evidence that peaks in ultraviolet radiation from the 11-year solar cycle have an effect on human lifespan - people born during the 3-year long peak exposure period appear to live about 1.5 years less than people who weren’t born during that period. Wikipedia says we’ve probabl yjust come out of solar cycle 24, with peaks in 2011 and 2014. But hey, it’s only 1.5 years if you happen to have had children born around then and it’s not like we’ve got other things to worry about on this planet. Also this is interesting because a long time ago I did a tiny tiny bit of work on Solar Stormwatch, a citizen science project to comb through coronal mass ejection data from the STEREO spacecraft.
* when tech companies publish principles I always get interested, so when Google launched its new Google Home thing and said they have “three very specific ways” in which they provide a “helpful, personal Google experience”, I pay attention. Those three specific ways are, paraphrased, i) providing experiences that are “unique to you”; ii) security, i.e. simple, powerful ways to safeguard your devices; and finally iii) being in control of your digital wellbeing.
It is interesting to see how quickly the industry has clustered around the phrase “digital wellbeing”, starting with the time-well-spent movement. I mean, it’s not really a surprise, wellbeing is a good thing, and who wan’ts to talk about addiction and distraction? Always good to focus on the benefits, right?
* Paul Ford, who can explain what code is, delivered a bunch of literally stunning tweets when I read them this morning. For full disclosure I should probably point out that they turned up in my Twitter notifications as tweets that the people I followed had favorited, which… is probably what I wanted? Anyway, I’ll just excerpt Ford’s tweets:
“It is too bad, given how urgently action is needed, that the science-loving leaders of massive social platforms refuse to band together and use their vast reach to communicate with their billions of users about the real risks of climate change. Or to put it in clearer terms: The first person who goes all in is going to win a Nobel Peace Prize within 10 years. Could be Mark, Alexis, Jack—even Reid. Can't be all of them. No Nobel Prize for fiddling.
“2040: “I accept this Nobel Prize on behalf of the millions of people who headed the call to respond to climate change and came together on [platform] to create a safe and more just future for all.”
“OR
““At [platform] we take your privacy s—” (doors burst, room floods, all flee).
“I mean even five years from now. “For his work with the Reddit Action Network on Climate Change...” soooo many global rewards and soooo much truly earned respect for the very wealthy man who is NOT afraid.” You all who work for these dudes know what motivates them. History books can say, “while the means to reach billions at nearly no cost was in the hands of these few, they did not act.” Or they can say, “among the brave few who clearly saw the threat and acted to save the earth—”
The part that I loved was the observation of the “means to reach billions at nearly no cast was in the hands of the few”. There are tiny baby steps: sure, Facebook will put in a tiny call to action asking people to donate to {most recent example of climate change induced humanitarian disaster}. Sure, it’ll be in iTunes and the App Store. Google might even put that on the front page. But, as the saying goes, that’s not really 10x thinking now, is it? That’s not really moonshot thinking, is it? I mean, we heard lots about moonshot but I think one of the things Ford is pointing out here is that as much of the climate change problem is about *people* and communication as it is about engineering solutions. Moonshots and [x] are a needed part but people have got to want change. Sure, you might get called in front of congress for using Facebook to tell people about what Trump and the GOP think is junk science where the jury’s out. But maybe that’s worth it? Or are you too afraid?
I am optimistic too, though. Or, rather, I’m optimistic about the technology. Maybe that’s naivety and idealism but what you can’t refute is that it’s easier to get one, single, simple idea out to *billions* of people than ever before. This is going to sound hypocritical of me, but at the same time I’ll just say that everything is context-dependent: I am in favour of moving fast and breaking complacency to save the world.
* ha, someone came up with a low-coding data-oriented “programming” language called Alan and it’s funny because it isn’t Turing complete.
* on the other hand while I am optimistic *in general* I am pessimistic in the micro. via jwz, this Kara Swisher/Anand Giridharadas interview about “engine problems” versus “crime scene problems”, which feels like one of those super useful abstract binary distinctions, a bit like askers/guessers. Anyway, Giridharadas brings up this axis of looking at problems where on the one hand you have engineering problems, which might as well encompass systems thinking problems: the analogy is that some sort of process is broken and you can recover by adjusting the variables and behavior of the particular process or piece of abstract machinery. There is less, if any, impetus in an engine problem to figure out *why* something happened, just that: ok, here’s where we are, let’s solve this and figure it out. There’s totally competence porn here.
Crime scene problems though, they’re the other kind. Crime scene problems are, Giridharadas says, ones where you absolutely do not approach them like engine problems. Giridharadas points out that crime scene problems are backward looking and interested in cause (which, I acknowledge, can be a difficult thing in today’s complicated environment!). But the broader point that I understand Giridharadas is making is that “you don’t show up at a crime scene and say ‘let’s just solve this’” where “solving this” is “okay, I guess we have a dead person, let’s move on and find a replacement for this dead person”. Solving “crimes” in this way is figuring out why it happened and who did it. And the particular examples in the interview are that frustrating genre of non-apology situation avoidance, where Swisher tries to talk to Zuckerberg about the shit that went down in Myanmar and Facebook’s involvement, and Zuckerberg replies with very well coached variations on a theme of “What we really want to do is fix the problem. We want to get soutions. Getting to solutions is important”.
It is almost as if, taken to an extreme, focussing on an engineering culture can lead to avoidance of acceptance of responsibility. I mean, this is complicated. At the same time there are people whom I really look up to, like John Allspaw and the concept of blamesless postmortems, where you focus on the fallibility of huamns instead of seeing mistakes and accidents as fault. But there’s a really obvious key here, which is that blameless postmortems are about accidents. They are when what happened *was not what was intended*. Now, you can say that obviously Facebook did not *intend* to bring about discrimination and violence [citation needed] but there’s a difference between accident and, say, wilful ignorance or negligence. (I should, for example, know the difference between them because I’m pretty sure I was extensively lectured and supervised on this topic while at university, but it turns out I’ve forgotten everything about it).
I mean, it would be really awkward if, at one level, the appropriate answer to “why did this happen” was “because providing a reasonable duty of care was not scalable and thus profitable”.
(side-note: I did not know that Facebook subcontracts moderation to Accenture.)
Just *typing* that makes me think of making a connection between stories about corporations doing the numbers and figuring that it’s cheaper to have a broken product and settle claims than it is to actually do something about it. And yet the first few comments on (sorry) the hacker news conversation about Heathrow Airport being fined GBP 120k for a data protection breach point out the calculus in assessing the fine (up to GBP 500k) versus the cost of actually training the workforce (the ICO found that only 2% of the workforce was trained in data protection) or, even, the cost of FTEs responsible for data protection. I mean, why should this actor bother, when the regulatory cost of non-compliance is so low and there’s a monopoly?
* my internet friend Kelly Link got a goddamn MacArthur genius grant and is celebrating with a sale at small beer press, the publisher she runs with her partner, and the only reason why I know about Link is because of Sean Stewart.
* about this time last year I went to Juvet in Norway and had my brain smashed to pieces. I don’t think it was just the literally awe-inspiring landscape or the humbling aurora borealis I saw for the first time, but also the combination of people I was there with, and what we were working on and talking about. This year, a different group of people went, and internet friend Damien Williams went and wrote up his experience at the Second Future of A.I. Retreat. You should read it, it’s good.
—
Apparently travelling across the country and spending time on airplanes to deliver a workshop is conducive to writing. I probably could’ve stopped earlier, but ¯\_(ツ)_/¯ and anyway, there were words that were waiting to come out.
I’ve done my workshop. I think it went okay? I pointed out to the attendees that every time I give it, I figure out more things so at the very least I get something out of it, which feels like it comes out wrong. One of the points that I tried to underscore this time was the idea of not negotiating with yourself and the particular context of being deployed as a team that’s going to Do Innovative Technology Stuff Using Innovative Methods and is being invited into a Big Old Institution.
Anyway.
How are you?
Best,
Dan