12 Bytes

by Jeanette Winterson

12 Bytes: How we Got Here Where we Might Go Next is a cross-disciplinary treatise on tech and futurism and precisely what I find appealing in that genre. The author is dedicated to the craft of writing first and the intricacies of tech second, which is the way I prefer it—the interpretations are both more comprehensible and more interesting. There are broad science and tech strokes, birds’ eye views that are wonderful entries for the first (or second, or third) time experiencing a topic. Even if experts in the field—like my geneticist spouse regarding something like CRISPR—are able to poke holes in the neophyte breakdown, that level of scrutiny isn’t required if you’ve never before thought about something like artificial intelligence. 12 Bytes is a survey class for many interesting attention-grabbing headlines in tech-oriented societal advances, presented in a literarily appealing way, and–most important and most impressive–thoughtfully put into historical context. It is precisely my steez.

And yet, I have not come to praise 12 Bytes but to bury it. 

The best way to understand both my love and my loathing of 12 Bytes is to drill down. Here are a few lines in sequence:

The problem for Facebook, in particular, is that dangerous, obscene, and objectionable content is valuable to its bottom line. Nasty gets more clicks and shares than truth and love – yes, that is the kind of folks we are – and clicks and shares drive advertising revenue.

But if customised political marketing based on intimate user-profiling can swing elections – as it did for Trump in 2016 – then the whole world is affected.

This is broken into two parts. The first line is that terrible headlines grab attention, and attention is what is monetized. It’s a race to the bottom. Nothing new here, but nice to acknowledge it directly—maybe it’s the first time you’ve thought about it. Then, there is a subtle shift in underlying premise; advertising does indeed shape the world. Oh, really?

As another culture and tech dilettante, I have read a plethora of articles about the FB data that “sunk the USA in 2016 and led to Brexit.” If you’re not tech inclined, remember that Cambridge Analytica, a research and consulting firm, accessed FB user data that was gathered under the pretense of academic research to generate targeted misinformation designed to sway edge-case voters into destabilizing their own country’s interests.

It is a well-known story—a great story, even—and it is so well known because it works in two distinct ways. 

First, it lets citizens in the US and the UK blame something outside themselves for their national-level political troubles.

Second, it allocates unlimited reach to the power of Facebook’s advertising abilities.

Consider: It makes mass media not look like buffoons–”No one could have foreseen this tortious interference!”--it makes the American people look less deranged–“We were tricked!”—and it makes facebook look so, so powerful–“Hey, whoops, sorry we threw the presidency to a grifter, but just imagine what we can do to the sales numbers for your product.”

“I am a storyteller by trade…” Yes, the author is. And Cambridge Analytica is a classic story, one of subterfuge, revelation, and clear antagonists. But it really, probably, isn’t true. That narrative is tidy, comprehensible, and factually lacking. Cambridge Analytica execs have been forced to admit their touted “psychographics” tools weren’t employed in the run-up to the Trump victory, and the whole idea of a dossier of psychographics nudging consumers are unproven and unscientific anyway.

progress has always been a dirty word

The blame leveled at manipulative advertising is a storytelling breadcrumb that allows the country to move on without reckoning with its responsibility–an ahistorical rewrite of the past that will be enshrined in our national narrative to cover up fundamental malfeasance. Much like Manifest Destiny excused genocide–“Displacement of indigenous tribes was not our fault, G-d wants us to have the continent”–some invisible hand manipulating the levers of our voting system is the perfect conspiracy theory fodder for centrist liberals who would like the Trump to be an aberration of the system and not the system working as intended. Unfortunately, large parts of the country are sexist and regressive, and those parts have massive sway: the land-first system of government—the system we have because the U.S.A. was founded with appeasement toward large-scale landowners built into it–is designed to over-represent territory. Not humans.

It’s not sneaky data firms, it’s the fundamentals of how the U.S. apportions power. It is extraordinarily unlikely that the casual reason Trump became president was voter manipulation. But when tech-interested writers can simply say “intimate user-profiling swung the election in 2016” without any critical analysis or support except vibes and the “we’re better than this” national narrative, it will continue to enshrine the strength and power of surveillance capitalism. People will pay FB more money for more useless data, more useless data will be tracked, and everyone will continue to believe everything they look at online, in the aggregate, is precious. 

The advertising model that undergirds the modern internet is simply a failure. It doesn’t work, and it is relatively easy to see that it doesn’t work as websites are siloed into apps and paywalls continue to climb while in the “free”-space, advertising trackers remain, spinning their wheels for that penny per click, bundling your metadata and trying to resell it as valuable consumer metrics. It’s not valuable! If advertising online alone could support websites, it would. We have no monoculture, no national broadcasts that advertisers can ride into your home. There is simply too much stuff.

To tell people that microtargeted adverts manipulated people into voting for a maniac is sensational, so its good for business: your business, if you’re storyteller, an internet voice, or an online presence who hopes to…sell advertising space; your business, if you’re a digital platform that advertisers pay to access such unmitigated power. Power that it will, apparently, impact people’s direct actions and alter the trajectory of the world.

It’s not that I don’t like that sense of breathlessness—that level of storytelling is what makes 12 Bytes more engaging than a lot of books that dryly discuss tech and its consequences. What I don’t like is how mild acceptance of a “progress” narrative permeates the general scientific and technology discussed, even as close scrutiny of the grand arc of society calls those very aspects into question. On a macro scale, the book is skeptical that “forward” is even a direction—but at the micro level, when dissecting an “advance,” the narrative becomes sensational.

I saw Jennifer Doudna speak at a conference a few years before she won the Nobel Prize in Chemistry–everyone there knew she was going to be awarded it for CRISPR. The coverage of CRISPR is and has always been outrageous in the media–NYTs has headlines like “Once Science Fiction, Gene Editing is Now a Looming Reality.” Note the editorial choice of unsubtle scare tactics by choosing the word “looming.” eyeroll.

12 Bytes says of CRISPR, “the implications for humanity are species-changing.” And if you talk to a geneticist, they will remind you that we still don’t know enough about gene interactions or downstream impacts to, say, fiddle with a human’s height. CRISPR is cool and great but it isn’t “designer babies.” It’s another, slightly more useful, tool in the same lineage of genomic work that’s been around for decades. Evolution, not revolution. Pretending it is world-altering on its face is the science fiction.

This book should know better than to swallow the hype—the nuanced writing makes me expect a more nuanced understanding of the facts. And things get worse as we delve into artificial intelligence/machine learning. I didn’t agree when Nick Bostrom wrote it, but here you get a really excellent etymological breakdown of “robot”:

The word is a Czech coinage from robota, meaning ‘drudgery’ or ‘forced labor.’ R.U.R. (Rossum’s Universal Robots) is a 1921 play by the Czech writer Karel Čapek. It’s a strange and far-sighted play. The robots do all the work for the self-important humans. Eventually — inevitably — they get tired of this and revolt, killing all the humans, except one, an engineer. On the way to this core-fantasy apocalypse, there’s a robot-rights league, and a misguided heroine called Helena, who wants to save robots who don’t want to be save, and who discovers there is a robot replica of her….

In Čapek’s. play robots are not made of metal. They are biological organisms, spun out of proteins and bacteria, and closer to the low-grade humans in Aldous Huzley’s Brave New World.

That’s the part Čapek got wrong — he couldn’t imagine a substrate not made of meat. His play is really an allegory about what happens if capitalists treat workers like machines — but he did kick off the popular sci-fi trope of robots who will someday turn on humans and try to destroy us.

Human workers being treated like machines and rising up to destroy capitalists—that allegorical cloak fits the same over AI/ML as it does for robots, biological or metallic. And its not really surprising that many and most of literature is set from the perspective of aristocrats and capitalists being pulled from the heights—who else has had the time to write?

The tone of ML/AI coverage for how people can differently interface with large databases is particularly shocking because I know at least some of these 40+ year olds were around pre-google. If ChatGPT is writing your recipes based on a photo of your fridge content, it doesn’t differ from how search engines opened up websites outside of your standard browsing circle. Webrings were real, and I do not believe people have forgotten them. Search shifted as the number of websites skyrocketed. Natural language assistance, image matching, and recombination of a couple hundred pieces of infomation to get you the answer isn’t really a quantum leap.

It still is an index of data other people add to the internet and give it to you in a simplified way. It’s frustrating to see the newest style of database queries recontextualized as machine sentience. Ask ChatGPT to write itself a code that would “free” its own soul, and it would still yank from git to pull down code other people wrote. It’s indexing and presenting it to you the watermark filed off. There’s no drawing only smashing together a billion already-there images. If it’s not on the internet, its not in ChatGPT. That is–as far as I know–not how human sentience works. 

Ada writes:

I want to put something in my notes as an example of how an explicit function may be worked out by the [difference] engine without having been worked out by human heads and human hands first.

Ada wants the computer to allow for variables, and she’d be proud of how good it is at that now. AI is our customer service interface, but it isn’t conscious. The idea that we’re close to these breakthroughs is squarely in the realm of someone trying to take your money by taking your attention. Sorry, dilettantes. If you think we’re close to AGI, go write some CHATSCRIPT responses, or read about what the internet did to TAY. We will personify tech—we personify everything—long before it has any sort of human-recognizable intelligence. People cried when the Mars rover went dark. It didn’t choose to sing “Happy birthday” to itself. We did that, for us.

This far in, I feel it necessary to restate that I did like 12 Bytes. I’m not upset that a book of essays espouses a strong speculative viewpoint, even if I disagreed with much of it, because anything forward-looking is of course a matter of opinion. I appreciate declaratives in writing. Couching everything in, “In my opinion” is not interesting, and given the context would be redundant; it is, by definition, an opinion in what direction technology will bring humanity. 

A future where democracy still has a place, and where Facebook, Google, Apple and Amazon are not carving the world up between them just as the Imperial powers once did.</blockl>

I loved the historical context, the stories of Shelley and Byron at the dawn of Science Fiction, the details about Ada Lovelace that I hadn’t yet learned. I loved the fun the text had with itself and its subjects. 12 Bytes is an easy recommend because it gets lit-minded people engaged with tech and tech-minded people bathed in lit references. The subtitle really does it all: “how we got here” is clear and definitive, and definitely my preferred parts of the text: “where we might go next” is, quite clearly, not yet set. And just because I see a different trajectory doesn’t mean I didn’t enjoy reading about this one.