In this essay I argue the following:
Brain emulation requires enormous computing power; enormous computing power requires further progression of Moore’s law; further Moore’s law relies on large-scale production of cheap processors in ever more-advanced chip fabs; cutting-edge chip fabs are both expensive and vulnerable to state actors (but not non-state actors such as terrorists). Therefore: the advent of brain emulation can be delayed by global regulation of chip fabs.
Full essay: http://www.gwern.net/Slowing%20Moore%27s%20Law
Predictions that improvements in manufacturing will lead to lower prices are made ceteris paribus; rising prices caused by a temporary disruption cannot be used to conclude manufacturing costs have gone up until the original conditions have been restored or been shown to be unable to be restored. Since R&D has largely gone on unmodified, there is no particular reason yet to expect that hard drive prices per unit capacity will be any higher in 2013, after most manufacturing facilities are restored and the market has had time to readjust, than an extrapolation made 1990-2010 would have predicted.
And the relevant question as to whether a facility is too expensive to rebuild is not one of the size of firms in that business currently, but of the expected rate of return on the capital. Sunk costs in the form of destroyed fabs will not prevent new capital from coming in to build new fabs (though it might bankrupt specific firms). For sabotage to actually have a long-term effect, it would have to happen regularly enough and effectively enough to significantly drive down the expected rate of return on capital invested in building fabs.
Related (about memory chips, but probably still relevant): A 0.07-Second Power Problem at Toshiba Chip-Plant May Affect Digital Device Availability/Prices.
http://www.evolvingexcellence.com/blog/2010/12/the-value-of-007-seconds.html :... (read more)
Ok, but what kind of actor are you envisioning as wanting to blow up the fabs? Nukes and bunker-busters seems to me to indicate nation-states, which - if genuinely convinced of the dangers of Moore's Law - have all kinds of other options available to them, like regulation. (Just look at what they've done to nuclear power plants...) If you decided to go terrorist and had access to a nuke, would a chip fab really be your highest-priority target?
A variety of links are broken- these include the link about suppression of firearms, the Cheyenne mountain link, the 2011 Thailand flood link, and the experience curve effect. It appears that something has messed up all the links that were to Wikipedia.
This piece seems to be proposing a solution to something that isn't obviously the thing to worry about. There are a variety of other threats to long-term human survival that require technological improvement to combat. Three obvious issues are asteroid impacts, large scale disease (due to modern infrastructure allowing the fast spread of otherwise localized diseases), and resource limitations (such as running out of easily accessible fossil fuels). Some of these are not directly connected to chip improvements- the Apollo program happened with mid 1960s level technology, and it is likely that the technological barriers to dealing with an annoying asteroid or comet are not strongly connected to computer tech level. However, others are not so limited- better computers mean better treatment of disease from better drug design and detection. Similarly, more efficient chips mean less use of oil (since less energy cost for the same comput... (read more)
TBH i'd rather share my planet with the mind uploaded folks than the folks that bomb factories. Both of those folks are potentially non-friendly non-me intelligences, except the latter type is for certain non-friendly while the former type might end up friendly enough in the sense in which corporations and governments - as meta-organisms - are friendly enough.
Looking at the essay more, I would say: the chip fabrication labs are amazingly cheap and even highly coordinated efforts by all terrorist groups can not make even a small dent in the progress (excluding scenarios such as big nuclear war)
The several billion dollars apiece for a fab is cheap . The margins may be thin when everyone else, too, is making chips. But I can pay 10x of what I am paying for my electronic equipment, if needs be, and all chances are you can too if you are posting here. The income distribution being what it is (power law, see Pareto distribution), the costs can be raised massively while retaining the revenue as long as the cheap alternatives are not available. More than 50% of your customers can and would pay >2x (if the alternatives weren't available). Think about it. 20% own 80% of everything (or even more skewed), and its like this all the way to the top. With this kind of income distribution the luxury market alone can support entire industry along with progress. Look at Apple, and how successful it is even though the products are not all that popular worldwide - they make huge margin.
Anyhow, with regards to the de-novo AGI, really, all we have is extre... (read more)
Are you actually suggesting that people attack chip fab plants in attempt to prevent WBE from occurring before de novo AGI?
I think if you were successful, you'd be more likely to prevent either from occurring than to prevent WBE from occurring first. It takes a whole lot of unfounded technological optimism to estimate that friendly de novo AGI is simple enough that an action like this would make it occur first, when we don't even know what the obstacles really are.
This is haunting the site. I see that your perspective is: "Does this imply that regulation could be accomplished by any modestly capable group, such as a Unabomber imitator or a souped-up ITS? No (reasons)" and that your position is that Terrorism is not effective. However, I have found several mentions of people being creeped out by this article around the site. Here is the last mention of someone being creeped out I noticed.. I think there is a serious presentation problem with this piece that goes like this:
Person clicks article titl
I wasn't sure that this was worth acting on, but I see that another person seems to be taking it the wrong way, so I guess you are right. I've done the following:
Isn't the first rule of Fight Club that you don't talk about Fight Club?
It seems to me that the key point is Moore's second law, leading to centralization. Centralized facilities are easy to sabotage. But if this law keeps going, it might end Moore's law all on its own.
If the capital expense of a fab keeps growing exponentially, pretty soon there will be a monopoly on state of the art silicon. What happens to the economics then? It seems to remove much of the incentive to build better fabs. Even if pricing keeps on as normal, the exponentially increasing cost of the fabs seems hard to finance.
The obvious solution is not to ma... (read more)
Great analysis. I am skeptical, though, that a campaign of targeted disruption could permanently derail Moore's law as long as the economy as a whole remains strong. Markets are really good at responding to dislocations like the destruction of a chip fabrication plant: other facilities would increase production in response to higher prices, unnecessary consumption would be curtailed (think of all those college kids using their fabulously advanced computers to surf Facebook), and future facility development would incorporate the threat of attack into their designs. We might even see companies start to trade special forms of insurance against such attacks.
I have some ideas for slowing Moore's law as well and, I'm wondering what you guys think of them (Gwern, too). I'm thinking of making these into post/s of their own and am curious about whether they'd be well-received or what, if anything, should to be done first.
Upvoted; very interesting.
I think it might be a hard sell to convince governments to intentionally retard their own technological progress. Any country who willingly does this will put themselves at a competitive disadvantage economically and defense-wise.
Nukes are probably an easier sell because they are specific to war - there's no other good use for them.
I think this might be more like Eliezer's "let it out of the box" experiments: The prospect of using the technology is too appealing to... (read more)
Sabotage? Isn't that going to be illegal?
Depends on whether the ones doing it are also the ones making the laws.
Maybe there's an intermediate possibility between WBE and de novo AI-- upload an animal brain (how about a crow? a mouse?) and set it to self-improvement. You'd still have to figure out Friendliness for it, but Friendliness might be a hard problem even for an uploaded human brain. How would you identify sufficient Friendliness when you're choosing humans to unpload? I'm assuming that human ems would self-improve, even if it's a much harder problem than improving a de novo AI.
Moore's Second Law reminds me of a good enough for sf notion I've got. Chip fabs k... (read more)
The real question is how to accelerate it, not ti stop it. And more likely to happen, too.
In some homes, the electricity used to power computers is already a significant fraction of total household power used. If a carbon tax were applied to the natural gas bonanza from fracking, prices would discourage buying just straight additional CPUs, in favour of making better use of the CPU speed we have already - it would simply be uneconomic for a company to run 100 times the CPU power it currently does.
EDITED TO ADD:
I guess I should expand on my reasoning a little.
Moore's law continues, in part, because there is the demand for additional computing... (read more)
Brain emulations are a joke. Intelligence augmentation seems much more significant - though it is not really much of an alternative to machine intelligence.