Fireworks in Your Eyes

Writing about a coronal mass ejection in Popular Mechanics, Lee Billings describes the International Space Station crew’s experience of a solar storm:

Even [in the ISS’s shielded core], the astronauts received elevated doses of radiation, and occasionally saw brief flashes of brilliant white and blue—bursts of secondary radiation caused when a stray particle passed directly through the vitreous humor of the astronauts’ eyes at nearly light-speed.

What a terrifying, beautiful thought! Our eyes, themselves organic particle detectors, betray invisible apparitions when high-energy particles ionize the clear jelly between the cornea and the lens. Perhaps we are nothing more than cloud chambers made out of meat.

Prophesy

Is it gauche to quote myself? To immortalize my own words that would ordinarily dissipate, unnoticed, into the vacuous expanse of human triviality? Fuck it; this is important.

Descartes was a fraud. If being required thought, the majority of humanity would be deprived of their status as conscious entities. But just because we have some idea of how we think, doesn’t mean we would recognize consciousness in alien form.

The dawn of AI will be like that—while we mistakenly search for algorithms that simulate consciousness, dismissing emotion as a pipe dream, we will be confronted by a barely self-aware bundle of tempestuous code:

We’ll have a hungry, angry, wailing Google bot with no way to learn what new inputs satiate its animal need to crawl the web…

—@brendn, in conversation with Mills Baker

Psychologists frequently refer to our “reptilian cortex”—the raw, animal place buried beneath bicameral sociability and post-enlightenment reason—as the pre-historic origin of fear, rage and libido. It’s all too easy to confuse these things with what makes us human; we mistakenly believe that emotions are the essence of our consciousness.

We look for intelligence in all the wrong places, and as we yearn to create an entity that can understand and converse with us, we run the risk of simply failing to notice the emergent behavior of intelligence.

After all, isn’t consciousness just memory? We try so hard to simulate reason and thought and creativity that we overlook the possibility that our dogs likely exhibit emotional responses¹; we refuse to believe a chimpanzee could share that same spark that makes us human.

So here we are. The approaching equinox promises to thaw the AI winter, and we’re presented with the possibility that the next nonhuman intelligence has already been left at our threshold. My fear is that in our boundless desire to create an intelligence in our own image, we will unwittingly create one that transcends it.

The Celestial Emporium of Genetic Variation

The animal taxonomy Borges made famous in his essay, “The Analytical Language of John Wilkins” has long been an obsession of mine. He claims that instead of the thirty-five phyla familiar to any contemporary student of biology, a certain Chinese encyclopedia classifies Animalia into

1. those that belong to the Emperor,
2. embalmed ones,
3. those that are trained,
4. suckling pigs,
5. mermaids,
6. fabulous ones,
7. stray dogs,
8. those included in the present classification,
9. those that tremble as if they were mad,
10. innumerable ones,
11. those drawn with a very fine camelhair brush,
12. others,
13. those that have just broken a flower vase,
14. those that from a long way off look like flies.

I remain unconvinced that such an absurd classification system could be little more than the quasifictional imagining of a professional yarn spinner—the truth, so they say, is stranger than fiction. I have my doubts too that the modern mind could even conceive of such a tangled menagerie. Is it instead the lure that draws you into the chaotic mind of an unhinged genius?

Those questions remain unanswered, mainly because my research led me elsewhere. The numerous citations of the passage, whether found in dusty library stacks or obscure Internet posts, all exhibited slight variations. Eco and other prominent scholars of Borges usually credit a translator, but many reproductions of the list do not. And curiously, it appeared no two instances were perfectly identical.

The following table illustrates the differences between the original Spanish and three selected English translations. The substance of each entry remains mainly the same between renditions, but there is remarkable diversity in the inessential bits—the pluralizations, possessives, capitalizations, and extraneous modifiers—which mark the individual whims of the translator.

Read more →

<TIME> Lapse

The outrage has blown over, the thunderhead safely recedes towards the horizon, and the sparrows are chirping and tweeting away in the rain-drenched foliage. But for the brief span of the storm’s darkest moments, we hunkered down knowing full well what was in danger of being taken away.

The HTML 5 <time> element, that precious jewel of markup, that essential semantic mot was poised at the precipice, prepared for its fate. It meant so much to so many, and yet we were too quick to forgive its shortcomings:

1. Formal protocols are nice, except when they aren’t.

   If you’ve ever written software to deal with dates and times, you know how much of a pain it is. The rigid logic of ISO 8601 is a refreshing comfort in a world of locale-dependent date formats, but it’s not without its flaws. The most obvious of which is the impossibility of representing dates prior to the beginning of the Gregorian calendar. Dates of important events in Hellenistic Greece are therefore woefully unable to assume a machine readable form. In my opinion, the <time> element is under-spec’d in this regard, and QuirksMode generally agrees.

2. Ad-hoc protocols are better than no protocols.

   I love the idea of and philosophy behind microformats. Before we had semantic markup, we were able to construct logical, sane, machine readable structures in HTML, and they could be adopted or not by a site’s developers. Sharing these ad-hoc formats freely online propelled their adoption without the need for endless committee meetings in a standards body.

   To this end, we could have explored alternatives to the <time> element. Some suggested we use <data> instead, only to be admonished by fearful developers. Ultimately, we did not explore our remaining options for annotating documents with date and time information.

3. Trust no one, but share your secrets openly.

   This is a variation on Postel’s Law. Ultimately, we have no way of ensuring everyone would microformats, <time>, or <data> elements correctly, if at all. We can only enforce sanity on systems under our own control. Smoothly exchanging structured data requires you to choose a format that works for the particular case you’re concerned with and document the hell out of it. And we all know that developers always adhere to the spec. (I’m looking at you, Facebook.)

[Update: as of 7 November, the <time> element is still missing from the HTML 5 Editor’s Draft, but the W3C HTML 5 Working Group mailing list has a revert request for r6783, which implies imminent restoration of the abducted tag.]

The Future Will Be Stranger Than Anything We Can Imagine

Maybe it was walking through Chinatown in a misty drizzle, neon aglow, empty restaurants next to storefront print shops inexplicably full of creaking machinery. Or maybe it was the dot com launch party I had just left, buzzwords and business cards flowing like cheap booze, another humming machine in a storefront. But I became acutely aware of a sensation that can only be described as melting into a William Gibson novel. It was unreal—the physical world resembling fiction—but everything about that walk home from the bar seemed like it had already been written on a dusty page. And because of that feeling, anything seemed possible in this world. We are limited only by the imagination of a writer behind a keyboard, except instead of prose, we write code.