Sunday, May 21, 2017

Insurgence, by Ken MacLeod

Ken MacLeod's The Corporation Wars: Insurgence, is the second book of a trilogy. It (along with the first book in the series, Dissidence, is a finalist for the Prometheus award this year.

Insurgence continues the story of awakened robots struggling for freedom, and uploaded human ex-combatants fighting to retake the planetary system the robots had been mining and exploring.

This installment focuses less on the robots' claim to be agents worthy of separate respect, and more on the uploaded warriors struggle to figure out the nature of the reality they inhabit while mostly following orders to fight the battles their supervisors are pursuing. Their ultimate worry is that they don't have enough information to tell which side they're fighting on or who they are battling to subdue. When you live in a simulation (particularly when you can tell that someone else has access to the control panel) it's a little difficult to be sure that your choices aren't effectively controlled by someone else.

Next, cracks appear in the simulation, and "real" revived people see the shortcomings, but non-player-characters (MacLeod calls them philosophical zombies) think everything is normal, so the real people can tell who's just a simulated person. The idea of zombies in philosophy (sometimes "p-zombies") is an exploration of the idea of consciousness. What if there were beings that acted just like people, but had no consciousness? Would it make a difference to them? Should we accord them lesser rights?

I consider the idea of p-zombies to be incoherent, but many smart people treat the question as exploring an important distinction. MacLeod here undercuts the point of the argument since there are actual behavioral differences. It isn't an exploration of whether consciousness matters, it's just that some characters in the story are imperfect simulations without an inner life, and the actual thinking beings can tell who they are. At the same time, MacLeod makes sure we notice that the robots and AIs who are active in the battles and the scheming do have an inner dialogue, and are making plans and collaborating with others to get things done.

The starting position for the agencies that represent the current Earth government and act under its protection is that only humans are allowed to be sentient. Even AIs' powers are circumscribed. Whenever self awareness arises otherwise, it must be stamped out. It's not clear why this would be a plausible stance, since it's clearly the case that the AIs can become self-aware for short periods, and autonomously operating robots have the capacity for spontaneous self awareness given the right trigger. So they must be constantly battling to defeat uprisings, and track down newly minted sophonts who either try to escape from control, or hide in occupied systems. It would make more sense to forbid use of tools with the capacity for self awareness, than to constantly try to stomp them out. I'd also have a hard time going along with a regime that wanted to outlaw and destroy a class of beings because they were self aware. Self aware and hostile is a separate thing, but that's not the distinction they've settled on.

Before one of the final battles, one of the leaders of the simulated humans challenges the combatants to each eat a slice of p-zombie flesh to prove that they believe they're in a simulation, and that there can't be any moral issues with simulated eating of simulated meat from simulated people that were never actually alive or aware. Except for a few who object to the initiation-ceremony aspect of the act, they all partake.

So there's a lot of exploration here of of philosophical questions of identity, and what it means to be human. The questions of liberty are mostly focussed on what kinds of agents deserve respect as actual people, though I think MacLeod fumbled some of the issues. The action is interesting and the conflict exciting. Besides there are also weaponized communications packets, interrogations of potentially hostile agents by sending them into a dungeon simulation, double and triple agents, and terraforming. It's a pretty good read, and the lead-in to part three, of course leaves a few things to be resolved.

Saturday, April 15, 2017

Dissidence, by Ken MacLeod

Ken MacLeod's The Corporation Wars: Dissidence, is the first book of trilogy. It (along with the second book in the series, Insurgence) is a finalist for the Prometheus award this year.

The story starts with a scene in which a pair of mining robots exploring an asteroid (in a distant solar system) and representing different corporate interests have an encounter, which leads them to realize they have opposing interests, which leads them each to recognize that they have interests, which leads them to self-awareness. The corporations are in a tenuous situation, trying to assert their ownership of the robots, trying to be civil about their contractual cooperation, but objecting strenuously to breaches by the opposing robots. The corporations end up fighting one another, and the robots band together and spread the concept of self-awareness to other nearby robots with sufficient computing capacity. Since the corporations don't seem likely to grant them independence, the robots form an independent faction in the upcoming battle. The corporations are loath to destroy their valuable property just yet.

When they do decide that military actions are called for, they end up dredging up opposing troops of uploaded warriors from past wars. All the AIs and non-self-aware robots, and other actors are under a deep compulsion that only humans and their uploads can actually be armed for combat, even against rogue self-aware robots. So the "humans" spent parts of their time embodied as people in a planetary environment, training and relaxing between missions. In the missions, they're downloaded into articulated space battle suits. Every time they die in battle, they return to the training site to start again. Over time, they find reason to doubt the reality of their home, and eventually detect serious cracks.

The uploads gradually learn enough about their realities to doubt that they're still fighting for the side they were loyal to in their first lives. Apparently part of the distinction between uploads and awakened AIs is that the operators can't tinker with opinions and loyalties directly, but they can easily lie and mislead about who they're representing, and what their opponents are fighting for. Of course, it wouldn't be an interesting story if the operator's control couldn't be subverted.

Ken MacLeod tells a good story, and gets us to think about what kinds of entities should have rights. The authorial point of view allows him to show the action in the eyes alternately of the awakened robots and the revived soldiers, so we feel their fundamental humanness. The characters, ex-human and non-human alike, think about who they should allow into their coalition, whether other actors are actually aware or just act like it, and have varying motives.

My biggest complaint about the story and the characters' attitudes is a simple acceptance among all the characters that some other characters are not real, based simply on statements from people in authority roles. In war, it doesn't make much sense to worry about whether the people shooting at you are actually thinking beings, but deciding that some category of bystanders don't have inner lives should be a cause for more intensive investigation. It's an easy allegation to make, and not far from standard attitudes about our enemies that we've mostly moved past.

Saturday, April 08, 2017

Johanna Sinisalo's The Core of the Sun is a finalist for the Prometheus award this year.

It has enough SF elements to qualify as standard near future fiction, and provides biting social commentary. In feel, it reminds me a lot of Margaret Atwood's The Handmaid's Tale, but I liked this better in several ways.

The story takes place in a future Finland that has managed to selectively breed its women so that they're either docile sex dolls and mothers ("eloi"), or sterile, powerless but competent workers ("morlocks"). They've also outlawed psycho-active drugs from alcohol to heroin, and somehow included capsacin (hot peppers) on that list. The protagonist (Vanna) is a morlock who was raised as an eloi, which allows her to pass in polite company. She's also hooked on hot peppers, and has started dealing in whole, dried, and preserved peppers in order to afford her next fix.

Compared to Handmaid's Tale, the viewpoint character is a more active agent, with more freedom to act for her own interests and to undermine the system; her allies against the state are more fully bought into the fight; the state she fights has taken more reprehensible steps, though it seems to have less thorough control of its subjugated females.

The story is told with a mix of present-tense action, and recollections by Vanna of how she got to her present situation, mostly written as letters to her long-lost eloi sister, Manna. The two were raised away from the city by their eccentric aunt, which gave Vanna the opportunity to act naturally most of the time, and mimic her sister when visitors were around. This gave her the tools to pass as eloi when she grew up.

After the aunt dies, Manna finds a husband who Vanna suspects to be after the family farm, since neither Manna nor Vanna (passing as an eloi) can legally hold title to it. Vanna finds a man to partner with who values her for her unusual intellect and her ability to act independently (a useful tool for his black market activities).

Vanna pursues the secrets behind her sister's disappearance until events force her to escape with her partner. I found the prose (and occasional poetry) to be delightful and very evocative. The characters were convincing, and Vanna's struggle to be her own person in the face of societal expectations was heart breaking.

Saturday, April 01, 2017

The Vital Question, by Nick Lane

Nick Lane's The Vital Question seeks to explain why all eukaryotes share a large number of traits that are completely absent from all bacteria and other simple organisms. As Lane says in his opening page

All complex life shares an astonishing catalogue of elaborate traits, […]. Why, if all of these traits arose by natural selection, in which each step offers some small advantage, did equivalent traits not arise on other occasions in various bacterial groups?

Life arose around half a billion years after the earth's formation, perhaps 4 billion years ago, but then got stuck at the bacterial level of complexity for more than a billion years, half the age of our planet. […] In stark contrast, all morphologically complex organisms […] descend from that singular ancestor about 1.5—2 billion years ago. This ancestor was recognizably a 'modern' cell, with an exquisite internal structure and unprecedented nanomachines encoded by thousands of new genes that are largely unknown in bacteria. There are no surviving evolutionary intermediates, no 'missing links' to give any indication of how or why these complex traits arose, just an unexplained void between the morphological simplicity of bacteria and the awesome complexity of everything else.

Lane claims that this gap in our understanding should be glaringly obvious, and the scientific community should be struggling mightily to fill it in, but (he says) few are working on it, and fewer are talking about it as an important item on the agenda.

Lane's argument is that the combination of bacteria and archaeon that allowed the formation of eukaryotes happened once, and must have quickly evolved to have mitochondria, cilia, and to rely on sex for reproduction, and that all complex life descended from that single event. One of the surprising things is that eukaryotes didn't replace their ancestors; even though they have enough advantages that all complex life descends from that single event, there are still plenty of opportunities for the ancestral forms. The explanation Lane presents is that there's a delicate balance in the energy economy in bacteria and archaea, which doesn't allow the cells to grow much larger, and puts serious constraints on what kinds of mechanisms can be powered inside the cell. When that single archaeon engulfed a bacteria and turned it into the primal mitochondrion, the energy balance changed, and it became possible to store energy and distribute it around the cell, which made it possible to power more kinds of mechanisms, which led to the explosion in the variety of life and ways of living.

The usual story is that the environment changed (the Great Oxygenation) which enabled more styles of living. But what you'd expect if that was the cause would be a separate explosion from every kind of living creature, while what we really see in the evolutionary record is that when there are events like this (the cambrian explosion, e.g.) they radiate from a single progenitor, which tells us there was a significant discovery in that line that enabled the new directions of evolution.

Lynn Margulis' research shows that the form of modern eukaryotes derives from a series of mergers of adjacent bacteria and archaia. (One of the .sig lines I use refers to this) Lane says that while her results hold up, the mergers all occurred in a single line of descent, and all existant eukaryotes radiated out from the same end point of the serial events. Apparently none of the intermediate forms were good platforms from which to generate new life forms.

There are some simpler organisms (giardia among them) that are like eukaryotes in many ways, but lack mitochondria. They have long been viewed as an intermediate evolutionary point between archaia and eukaryotes, but modern phylogenetics (tracing the descent via gene similarity) shows that they're actually descended from eukaryotes, and merely discarded some of the internal structure because it wasn't needed in the environmental niches they occupied. This buttresses Lane's contention that all plants, animals, algae, fungi and protists share a common ancestor.

The common ancestor stored its DNA in a nucleus with a double membrane. The cell itself has a membrane with pores that were inherited by all the branches of its descendants. All the DNA has telomeres as well as introns which are spliced out using common machinery before proteins are built. The golgi apparatus, the form of the cytoskeleton, mitochondria, lysosomes, peroxisomes, the endoplasmic reticulum and the intra-cellular signaling mechanism are also common.

If you're interested, Lane goes into a lot of detail on his hypothesis on the energetic mechanisms that could have led to the evolution of the mitochondrial pathway starting from deep sea hydrothermal vents, where hydrogen and oxygen are bound in a way that can produce positive energy when the bonds are broken. I mostly understood it as I read it, but I'm going to have trouble doing it justice. Here's a precis of the argument; ATP is the end product, and is both stable and easy to extract energy from. A simple mechanism that can produce ATP has the effect of making many energy consuming processes possible.

Hydrothermal vents at the ocean bottom ("black smokers") are places where constantly renewed magma is in contact with sea water, which results in hot acidic water. Lane picks out nearby "alkaline vents" (also on the ocean bottom, but not where magma is exposed) as the plausible site for metabolism to arise. The alkaline version is rich in dissolved hydrogen, accompanied by "other reduced gases including methane, ammonia and sulphide". The rock is riddled with micropores from micrometers to millimeters in size. The flow of warmed sea water is relatively slow, so there's plenty of time for percolation and reaction. There are eddies in the flow, which allows reactive products to accumulate and concentrate locally. Before the Great Oxygenation, the most common gas in both the atmosphere and the ocean was CO2. In this environment, CO2 will react exergonically (releasing energy) with H2 to form CH4 (methane), but it needs a catalyst.

Lane considers it a crucial clue that all living cells drive their energy metabolism via proton gradients across thin membranes. To expand that, the claim is that the production of ATP always happens in the presence of cell membranes that separate proton-rich from proton-poor regions of a cell, and require a constant supply of H2 on the low-density side, and produces ATP on the high-density side. On the low density (alkaline) side, the H2 donates an electron, which is gobbled up in the production of ATP. Both of these reactions happen spontaneously.

Lane then describes a path via which permeable membranes (which don't benefit from better pumping) could evolve to be more selectively permeable, which would allow better pumping to be a benefit. This change makes it possible for the cell to escape from the natural proton gradient, since it can sustain its own internal gradient. Lane hyphothesizes that once selective permeability arises, archaea and bacteria evolved different membrane pumps (evidenced by the fact that they use steroisomers of glycerol) and split into evolutionarily distinct lines.

I'm not sure I explained that very well, but this felt like the first time I've read an explanation of basic cell metabolism that presented a mechanistic picture of the benefit of ATP (stores energy in an easily-extracted form), how the production of ATP is paid for energetically (proton gradients maintained by membranes and selective pumping), and why these designs are fundamental to the difference between bacteria, archaea, and eukaryotes, and eventually lead to the development of chloroplasts as an alternate energy source. Lane gives an explanation at a similar mechanistic level of what happens during apoptosis (programmed cell death; also conserved across the eukaryotes!)

Lane also argues that anti-oxidant supplements interfere with the apoptosis pathway, and thereby reduce health. He presents this as the currently accepted scientific viewpoint, though it's news to me. I need to do more research here.

I learned a lot of biology from this book, and thoroughly enjoyed it.