快猫短视频

THIS is certainly the best book about the singularity. It features 26 intelligent scholars from 11 widely varying disciplines, all of them valiantly grappling with ghosts.

Given that the subject matter is so highly speculative, so lofty, so indefinable, this tome is heavy going. Among its talents are nine philosophers and nine artificial intelligence researchers. These worthies mercilessly lay it on with their specialised jargon. It takes a sturdy, dedicated reader to plough those thickets of prose.

Worse yet, since The Singularity is a work of metaphysical philosophy, you know from the start that no amount of argument will settle its complex issues.

The book opens with a target essay by philosopher of mind David Chalmers. Each contributor fires at Chalmers with their heavy intellectual artillery, and he then appears at the close of the book to briskly refute their objections with his premises unscathed.

While the book is a tremendous flight over the craggy AI landscape, it settles no disputes and has little or nothing in the way of practical counsel.

Kant, Hume and Descartes are major intellectual presences here, apparently because explosively proliferating future AI singularities are going to be plenty worried about these three dead European guys.

The tone of the book is mostly grave, sombre, even pre-apocalyptic (except for Daniel Dennett, who can鈥檛 resist making sarcastic fun of a younger colleague, and Damien Broderick, who clearly enjoys explaining science fiction writers to philosophers).

The Singularity comports itself as if confronting a serious public emergency. But is it serious?

Maybe. The book is the one the president of the US would need if we lived in a techno-thriller where Siri tried to seize power. Imagine an emergency NATO summit where terrified global diplomats needed expert briefings from metaphysicians. The contributors of this book would definitely be those guys.

Here鈥檚 the central issue with a singularity 鈥� as Chalmers frames it. We certainly know that intelligence exists (because we鈥檝e got some). Being intelligent, we鈥檝e probably got enough technical smarts to create some contraption that is, in at least some ways, as intelligent as us 鈥� meaning AI.

鈥淭he US president would need this book if we lived in a techno-thriller where Siri tried to seize power鈥�

Humans equipped with AI would then have the capacity to invent what Chalmers calls AI+. This new entity would then deploy its active, ambitious ability to build still further entities, dubbed AI++, and pretty much outside our human ken entirely. That would be, well, problematic.

It boils down to a technical question of the ratio of AI鈥檚 capacity to compute versus its capacity to construct. If a little bit of extra smarts makes AI an ace hardware engineer, then we鈥檙e really sunk.

On the other hand, if AI+ is truly brilliant but wants to spend its energies pondering Kantian ethics, then we鈥檝e lucked out, and the singularity becomes a more remote threat. However, the singularity will still happen.

Chalmers is sincerely convinced that a singularity is our destiny. Although progress in AI has been fitful to date, he believes that these developmental stages logically imply one another. So sooner or later, humankind will invent a true Artificial Intelligence: a thinking machine that fully deserves attributes such as wisdom, acumen, self-awareness, mind and consciousness.

Once an AI+ acquires those nebulous mental qualities, it will no longer be some merely standard chatbot, deep-learning neural network, genetic algorithm, theorem prover, expert system, fuzzy logician and/or evolved program.

No, that AI+ would be a genuine peer-competitor of humankind, a conscious and numinous being necessarily featuring goals, desires, preferences and values.

The AI+, though, is made of code and software. So it could probably expand its capacities with standard cyber-industrial procedures. So, even if AI++ is not a sudden, destructive 鈥渟ingularity bomb鈥�, even if the construction of AI++ by the AI+ takes decades or centuries, humans will still, eventually, be among intelligent entities that can out-think us as we out-think mice.

This truly major philosophical issue does not hugely upset Chalmers since, as a metaphysician, he suspects that consciousness is a primal aspect of our universe. This implies that the AI++ mind isn鈥檛 entirely alien to our own being, but is an amplification, a mind with our essential mental qualities, though bigger and faster and made of different materials.

This means that we could enjoy the standard singularity hobby of uploading and downloading our brains on to other substrates. The implications are many. We might perhaps be signing long-term contracts with the postbiological creatures of roboticist Hans Moravec鈥檚 book Mind Children so that they have an incentive to look after us, rather than simply squashing us like bugs.

It鈥檚 a new, post-singular, strange order of being, yes, but even a mouse survives in a world whose superior intelligence tolerates and even understands it.

So we鈥檙e mice-to-be, and that鈥檚 politically challenging鈥� unless, that is, we鈥檝e already been simulated by some hyper-intelligent A++ entity. It may be that we merely perceive ourselves as human beings. Maybe, in a higher reality, we are all the software constructs of a godlike cyber-entity that runs our code out of some vague sense of obligation.

This is where Monsieur Descartes pops up suddenly, and being Ren茅 Descartes, he gets an extended cogito-ergo-sum look-in on the hot topic of us not really being here at all.

Being a science-fiction writer and a veteran reader of Philip K. Dick, I鈥檓 inclined to indulge this philosophical conundrum. I enjoyed it. It鈥檚 entertaining. However, compared with the stark moral urgency of humanity being torn up for computronium scrap by malware AIs, I frankly have to wonder why our non-existence is even mentioned. Because, if we humans don鈥檛 exist, that is pretty awesome, but it鈥檚 mighty awkward. Really, it鈥檚 a faux pas.

鈥淢aybe, in a higher reality, we are all the software constructs of a godlike cyber-entity鈥�

A serious worrier might protest that even if we鈥檙e already victims of a singularity, and we always have been we could still have another one here in our little pocket sandbox of a singularity. But isn鈥檛 that over-egging the pudding? Where鈥檚 the moral urgency in our being 10,000 AI angels on a pin?

The current US president, to judge by his recent public statements, is sincerely worried about sudden, lurching, dangerous advances in computation. But if the president doesn鈥檛 exist at all (and the apparent odds seem to favour that cosmic possibility) wouldn鈥檛 he just clap this heavy book shut, and throw the silly thing across the room?

This article appeared in print under the headline 鈥淎 singular proposition鈥�