Episode Eighty One: Board-Level Algorithms; You Speak GIF, Right?; Intent; Odds
0.0 Station Ident
I'm back in Portland, where for some bizarre reason it's hotter than Dearborn, Missouri. I've left my wife, son and mother behind with my father-in-law. Now I have a week or so of bachelorhood before wife and son come back. And there's only one episode left of Agents of SHIELD left to watch.
1.0 Board-Level Algorithms
Aging Analytics Agency[1], whose website look like the kind of thing a production office would cook up for a movie or an ARG, issued a press release on Tuesday stating that "A Venture Capital Firm Just Named An Algorithm To Its Board Of Directors"[2], which press release is about as full of hyperbole as you could expect.
As an aside, the press release cites as a source a Business Insider article[3], which cites a Beta Beat article[4], which cites the press release, in a somewhat suspicious circle-jerk of attribution.
In any event, anyone who knows anything about VCs will know that there's a significant difference in terms of making investment decisions and *being named to the board of directors*, the latter of which opens up all sorts of legal issues around what would qualify as a Director that would make Ray Kurzweil super excited.
Only Deep Knowledge Ventures[5], the Hong Kong-based VC company, knows in detail how Aging Analytics' algorithm will influence their decisions in practice. Their website says that they partner with "the top analytical companies utilizing machine learning and [a] large panel of experts" to "construct complex decision trees showing risk and return at every stage of company evolution."
All of which is, you know, blah blah big data blah blah deep learning blah blah investment decisions which is all fine and well until you find out Microsoft have incorporated a Deep Learning Wizard into the next version of Excel and all of your modeling spreadsheets are out of date.
Absent substantive information from the press releases or public interviews, what we have here is a sort of expert system that the VC firm and its partner supplying the algorithm have seen fit to *let the rest of the world know* that it's helping them make investment decisions. Deep Knowledge Ventures is already an interesting company because of its investments in In Silico Medicine[6], and if you know what *that* means, you're wondering why a Gibson-esque science-fiction novel has again busted out into the real world perhaps a couple of years early.
In Silico Medicine - apart from being the name of something you'd expect from a Michael Crichton novel and the phrase cropping up as the title of a good book from Jim Munroe[7] - is the practice of Moore's law encroaching further into the medical process. Faster processors and more sophisticated software means higher fidelity simulations which means faster scientific iteration. At least, so the theory goes, if you're a singulatarian. But, of course, whilst those systems are more complicated to simulate than generally thought, it's never wise to bet against Moore's Law.
So no. An algorithm has not just become a legally defined person capable of holding corporate office. It is merely "voting" in an advisory capacity to assist a board of directors in making investment decisions. The real story here is yet another mark in the book showing how intertwined we are becoming with the technology we create (which has always been the case), and the gradual prising of open of the Overton window for *real* algorithmic representation in corporate law.
[3] http://www.businessinsider.com/vital-named-to-board-2014-5#ixzz31dVwrSEo
[4] http://betabeat.com/2014/05/v-c-firm-names-robot-to-board-of-directors/
[5] http://deepknowledgeventures.com
[6] http://www.insilicomedicine.com/#!/
[7] Everyone In Silico:
Amazon: http://amzn.to/RFQMJ1
Powell's: http://www.powells.com/biblio/9781568582405
2.0 You Speak GIF, Right?
Balaji S. Srinivasan posted a provocative multi-tweet-essay in the style of his a16z colleague Marc Andreessen yesterday on the evolution of our species communication technology[1], from the written word through to animated GIFs. Srinivasan's got an interesting point - written laws were a breakthrough, he contends, because they allowed "lossless digital replay over distance" - and we'll just leave aside the inaccurate usage of the word 'digital'. They won out of oral laws because oral communication was limited not only in travel radius and memory, but also in a temporal sense, because human memory necessarily decays over time. So yes, written language allowed scaling a culture to much larger areas than ever before.
The tweak that Srinivasan adds is looking at communication through the lens of the current trend of photo and image sharing. Language, he says, began with hieroglyphs and pictograms, then became ideograms and then became abstracted into alphabets. So, you can expect language-forming behaviour around photo/video and image sharing, thus leading inexorably to the memes that we have on the internet. And then we get to a sort of not-quite-Dawkinesque memeplex - that is, conglomerations of memes that convey complex concepts.
What we have with photo-sharing and the culture of meme/photo/video sharing is, Srinivasan says, the "ability to communicate what we see/hear digitally, losslessly over great distance" without using just characters. So it's not that we're entering a post-literate society because memes are incorporating written language - you still need to understand English to grok the full nature of the meme. But memes again are built upon a shared understanding of culture - they're built on the stump of literacy, or cultural literacy, as a base.
I think the point that I take issue is this idea of lossless communication. Human communication is necessarily lossy - because there's that part in our heads where we interpret what's going on and integrate it into our own understanding. True digital lossless communication would be along the SFnal lines of transferring mind states, and whilst the literality of what Srinivasan says is true: ie communicating what we *see* and what we *hear* losslessly, what is lossily transmitted is the *intent* or the meaning behind those images.
Weirdly, this is a part that animated gifs are, in part, fixing. In terms of animated gifs ripped from popular culture, what's interesting about them is that they appropriate the entire context of the surrounding material. So a shakes-my-head animated gif from Doctor Who is communicating a very different shakes-my-head animated gif from The Good Wife, for example - and both only work if both the sender *and* the recipient share a similar cultural reference frame: or, we've both watched that episode, or understand those characters and their motivations and their personalities and their settings.
So whilst image and photo sharing is undeniably a leap in terms of literally being able to *see what I see*, it's not the same as *understanding what I see*. There's still lossy communication there.
[1] https://twitter.com/balajis/status/466762126934016001
3.0 Intent
Via Gruber, John Moran wrote about what makes Apple unique[1] (because in our world, we will never be short of opinions about what makes or made Apple unique - and yours truly is as guilty of this as anyone else), and takes a look at what it is that makes Apple's particular application of that woolly word "design" special as opposed to any other company's. Morgan's point is that design is about intent - and that it's not just "how it works as well as how it looks" but that the intent covers "why it does what it does" or "how should this be". With this, Morgan pulls out quotes like Steve Jobs potentially being at his most frustrating in his home life, with the story of him taking forever to choose some furniture (or, if you've read the Isaacson biography, a washing machine) because he first had to work out what the *point* of a sofa was in the first place before he could choose the right one.
Morgan contends that no other company makes intentional design decisions like Apple, and it should be clear that Apple only occasionally holds itself up to that particular high standard, otherwise we wouldn't get things like iCloud or iPod socks. Instead, Apple's competitors and other companies make what Morgan categorises as Design Evasions that fall under the category of preservation (see: Microsoft), copying (see: Android) and delegating (see: Samsung and Project Ara).
I wrote in a previous episode[2] that Astro Teller's designs for Google X - a sort of sufficiently advanced technology indistinguishable from a do-what-I-mean magic - was a bit ambitious and perhaps misrepresented when he spoke at TechCrunch Disrupt. But in case it was ever not obvious, that lack of editing seems to be a weakness for Google. Sure, Apple appear - from the outside - to lack the technical chops these days, but at least you can rely upon them to be ruthless (ish) on the what and the why of the utility that they're providing. What I do find interesting is how this philosophy of what-a-thing-does and curation/editing means in terms of smart objects that are comfortable being connected with a network. Sure, they'll be good at what they do, but how do you curate/edit how they talk to the rest of the network?
[1] http://rampantinnovation.com/2014/05/13/design-is-about-intent/
4.0 Odds
- It's Siggraph technical paper preview time. Here's 2014's, full of more jaw-dropping algorithmic awesomeness: https://www.youtube.com/watch?v=u3Z1hDwGEmM
- Stross on Pikettty: http://www.antipope.org/charlie/blog-static/2014/04/the-prospects-of-the-space-and.html
- Amazing Military Infographics: https://medium.com/message/1ba60bdc32e7 - this one I think I might write a bit more about tomorrow.
--
Seriously, you lot are underperforming with notes. I can only assume that all the new people are far too embarrassed to introduce themselves, to which I say: it's just an email. Just say hi. Also: only nineteen episodes to go til I hit one hundred!
Best,
Dan