Some, for example, could point to the possible linguistic similarity between Kerberos, the guardian dog of the Greek Hades, and the epithet sabala ‘spotted, varicolored’ (*kerbero?), the standard epithet for one of the dogs of Yama, the Indic god of the dead. And even after more force than the comparative method in linguistics will normally allow, all one gains by postulating such a correspondence is the somewhat incongruous image of a Proto-Indo-European canine guard of the realm of the dead who answered to the name of 'Spot!’
bitzdoesthings-deactivated20190 asked: AI revolution or peace with humans?
Peace with the raccoon before too long.
The San Francisco Examiner, California, March 1, 1933
reblogging for swain
(via birdblogwhichisforbirds)
My little loft library via https://ift.tt/2KV2ATB
This is how institutionality reproduces. Even the call to think critically about power becomes a clever smoke screen. There is a whole dissertation to be written on intellectuals using the word neoliberal to mean “rules I shouldn’t have to follow.” “If we focus on this one case, these details, this accuser and accused, we will miss the opportunity to think about the structural issues,” wrote Duggan. This was code. It meant, “You can talk about structural issues all you want, so long as you don’t use examples of people we know.”
— Andrea Long Chu, “I Worked With Avital Ronnell. I Believe Her Accuser.” (via baroquespiral)
“And that’s why evil will never prevail over good,” said Kevin sardonically to the dying sea monster, Haereals.
A sovereign wealth fund might make some sense if:
- The government just thinks it can invest better than the private sector. (This was the whole idea behind socialist efficiency, after all.)
- The government thinks the private sector is investing too little, so it taxes consumption in order to invest more.
- The government expects tax revenue to decline in the future. (E.g. the oil running out.) They could of course just lower tax rates now and let the private sector grow faster, but maybe they are worried about the money going overseas. And they don’t want to spend the tax money lavishly now only to have government spending crash in the future.
- The government thinks that somehow it would be an easier sell to grab a bunch of money now to start a sovereign wealth fund than to take the money out of future taxes.
I think your last point here – in which public reception and political viability matters – breaks some of the assumptions of your other points and your earlier post.
There, the idea was that the government can always tax whatever it wants in principle, which means it can treat any asset in the economy like its investment, and “liquidate” it via a tax when it wants to do so. Thus there’s no point to buying specific assets in exchange for the right to “liquidate” them later, which it already has. Or as @collapsedsquid writes
As for the second stock-flow point, the economy is going to grow in taxable wealth regardless so “acquiring“ the asset beforehand through the SWF doesn’t make sense, the money’s going to be there all the same, it’s just going to be taxed rather than gained as investment returns.
While this is technically true in an abstract “monopoly on use of force” way, in practice there can be political downsides to new taxes. Of course, setting up a SWF is not politically costless either. But there are reasons why an SWF may be more palatable to various actors than an equivalent use of taxation, and these reasons have to do with my original question “why shouldn’t the government invest?”
If you raise revenue X with taxes, buy assets with it, then liquidate them later to get X + dX, your profit dX was obtained in a politically costless way, in that it was obtained through normal participation in the economy. If the government buys some shares, everyone knows they can sell those shares for revenue, just like any other actor. There might be objections to this behavior if, say, enough is done at one to destabilize the market, but one can’t make the objection that the government is just arbitrarily or unexpectedly demanding money, as one might about the equivalent taxation plan where it just levies dX worth of extra taxes at some point.
Or, to put it another way, investment is good financial planning for the government like it is for many other actors, except with the twist that relative to all other actors, the government has an unusual ability to just take money from people. But taking money from people isn’t costless if you depend somewhat on the approval of the people you’re taking it from. And from the perspective of those people, it seems more palatable to say “I take care to cultivate ‘normal’ sources of revenue like anyone else, but on top of that I really need this money to do such-and-such good things,” as opposed to “I’ve never cared about cultivating ‘normal’ sources of revenue because I can just demand it from people arbitrarily.”
(This has veered pretty far away from central SWF topics in that I’m now talking about the state’s relation to owners of existing capital, not to the everyman with no capital. But the two are related: allowing the everyman to capture a fraction of “normal” capital gains in the same controlled way as any other owner is different from whimsically exerting the right to take arbitrary capital gains at arbitrary times and then giving them to the everyman. It decouples social spending programs from complaints about the usual ways of raising revenues for them, letting the everyman make money “like any other capitalist” rather than by expropriative whim. This feels like a very non-leftist justification, but seems important just as a real difference between the two.)
The lightning bolts of thy mind’s eye to furthest ends, like la-
ser beams,
Leap out at once, with all the depth and weight and height
and strength of dreams!
A Response to Mike Konczal on the Social Wealth Fund →
Matt Bruenig has published a rebuttal to the criticism of sovereign wealth fund but I think he’s missing the very real things that Konzal’s complaint was based on, such as fund managers for union pension funds working to privatize infrastructure and bust unions as part of their investment strategy.
His rebuttal of the “Decrease Inequality and the Power of Owners“ section is pretty weak too because unless the social wealth fund is really big most of the proceeds will still be going to the rich. It’s reminiscent of the way that 401ks can function as basically a political guarantee of the stock market to the benefit of the rich by ensuring that a large number of politically influential people have a lot riding on the stock market.
His rebuttal of “Regulate the Economy“ seems to be pretty weird too. If I work at one company but gain money from all companies, then the purely selfish solution is for everyone at all those other companies to work as hard as hell even if I would prefer not to work. One of the things about the financial system is that in the course of it’s natural operation it basically makes that decision through maximization of profit, which immediately, and measurably benefits people rather than hard-to-measure things like “worker conditions” or “environmental damage”.
Given the fact that this actual managing of the funds all will be done by fund managers what this can end up doing is just re-contextualizing ownership, everything will be “owned by the public“ but they will be under the control of fund managers who have maximizing profit as their goal which can mean that you’ve just re-shuffled some words and justifications around.
Bruenig characterizes Konzal here as “giving up“ and seeing the left as basically a perpetual opposition who should not be interested in actually seizing capital but I see it as the need to reconsider the entire system rather than just reassigning, reshuffling, and rejustifying titles.
discoursedrome: isn’t the whole appeal of a sovereign wealth fund basically that it’s the smallest possible change that might have a substantial effect? That’s usually how I see it pitched, anyway, like “ok everything is fucked right now but here’s the minimum possible thing that would reduce fuckedness to within tolerances”
which I guess is not a very compelling argument for serious leftists but I do like practicable policy
Part of the thing here is that Matt Bruenig has been sort of selling it as “socialism“ in itself which is I think some of the reason for this. But part of it is I think that if you want to take money from or control business, then just use a tax or regulation. Why play this “ownership“ game that seems harder to get for less benefit?
I’m kind of confused by this discussion. As I understand it (which is not very well, probably), the issues over SWFs break down into:
(1) A purely financial question about how much the state should invest its income, vs. immediately spending it. In principle, this could be assessed from a ruthless state-income-maximizing perspective, without touching on how the income is actually to be spent or invoking socially just outcomes specifically.
(2) Issues about the idea of framing state investments as “publicly owned,” e.g. by typing them to dividend payments for the public. (How much does this matter? If it’s a good thing, is that justification for a higher rate of investment than the mere ruthless accounting of #1 would advise? Can we ensure the management of the fund is actually responsive to the investing preferences of the public?)
The issues under (2) are real and difficult, but some of the anti-SWF arguments seem to be arguing against the very idea of state investment as a financial choice. It’s hard for me to see how these don’t just reduce to “investment is always a bad choice” (clearly false!) and “investment is often a good choice for other entities, but never for the state” (in which case the pertinent difference should be spelled out).
For example, Konczal writes:
The United States needs to spend more money. It needs to spend more on a variety of goals, from basic income to infrastructure to removing key spheres of life like health and education from the market. Let’s abstract from what form that spending should take. Bruenig’s American Solidarity Fund would pay out in a basic income, but we can also imagine the profits of a SWF going to free college, the expansion of Medicare or better infrastructure.
A SWF is a terrible way to try and increase spending in the United States. It introduces a new failure point in the otherwise straightforward mechanism of spending and requires extensive and unnecessary prefunding. Normally we increase taxes on income X to secure spending Y. Here we increase taxes on income X to buy assets Z whose profits then secure spending Y. What is the point of this extra step? There is no economic argument presented that argues you can raise revenues more effectively or efficiently this more complicated way.
This sounds like a generic argument against investing: “why bother putting income into investments so you can pay later expenditures with the returns, when you can just pay expenditures with returns directly, which is simpler?” Not investing is indeed always simpler, but this does not mean no one should invest. The whole point is getting returns, which can make it easier to pay for all future Ys.
Konczal continues:
It also requires saving in order to spend, postponing resources necessary to meet urgent, contemporary needs. Let’s take free public college, which costs about $80 billion a year. We can do that by passing a financial transaction tax. But if we did this with a SWF, assuming a 5% returns on the fund, we’d need to first spend $1.6 trillion dollars buying up financial assets. We’d have to prefund 20 years of free college in order to give the very first person free college. You need to do this in order to secure the assets necessary to make the income necessary to get this spending. This problem presents itself no matter how to go about funding that $1.6 trillion dollars, be it printing money, taxing incomes or putting requirements on public firms. Why not spend that trillion and a half dollars in a useful way right now?
This seems like a stock-flow conflation. $1.6 trillion is a stock; if we spend it on free collage, we’d (under the same simplifying assumptions as the post) get to fund it for 20 years, and then we’d have to find more money. The hypothetical $80 billion returns are a flow: they could fund free college forever, all else being equal.
There is an argument to be had over how much we are willing to forego important state expenditures now in order to have more in the future (a kind of “discount rate”). But once you are arguing about what the discount rate ought to be, you’re implicitly arguing about how much the state should invest, and so you’ve left behind these strange qualitative arguments against investment per se. Well, unless you take the position that the future doesn’t matter at all, only the present, but that’s an extreme position and ought to be defended if taken.
Konczal is a finance guy and I can’t imagine he doesn’t realize all this. Either I don’t understand something, or … I dunno. ???
The issues under (2) are real and difficult, but some of the anti-SWF arguments seem to be arguing against the very idea of state investment as a financial choice. It’s hard for me to see how these don’t just reduce to “investment is always a bad choice” (clearly false!) and “investment is often a good choice for other entities, but never for the state” (in which case the pertinent difference should be spelled out).
[…]
This sounds like a generic argument against investing: “why bother putting income into investments so you can pay later expenditures with the returns, when you can just pay expenditures with returns directly, which is simpler?” Not investing is indeed always simpler, but this does not mean no one should invest. The whole point is getting returns, which can make it easier to pay for all future Ys.
The difference between the government and every other organization is that the government has the power just to take as much wealth as it wants, whenever it wants, in the form of taxes to pay for its expenditures.
Other people/organizations have to invest because if they spent their whole incomes on present consumption, they would have no financial security or ability to expand operations.
The government can let other people invest the money, and then take it whenever it wants for its consumption.
It’s really little different from, say, an investment account. If I have an investment account, I’m entitled to all the money in it. But if the companies I’m invested in are better at finding profitable business opportunities than I am, I can let them keep the money, let them keep compounding it for me, and then take it out as and when needed to fund future consumption. I benefit more by letting them handle it for me than by directly trying to put those funds in my own business.
In the same way, the government is entitled to (or at least has the power to take) all the wealth in the economy, so the whole economy is like the government’s investment account. Anything the government doesn’t take, it’s leaving for the private sector to compound for its potential future use. They can come knocking at any time to take it. So, unless the government is just better at investing for some reason, there is no reason for it to invest directly.
This is not a general argument against investment because the government’s power to take wealth from everyone else makes it fundamentally different.
My current thoughts on MIRI's "highly reliable agent design" work - Effective Altruism Forum →
This is from 2017, but I only just read it, and it did a lot to clarify for me why MIRI thinks work like Logical Induction is relevant to AI. Also does a good job crystalizing the reasons I (and the author) find this unpersuasive.
This in particular is really good:
I understand HRAD work as aiming to describe basic aspects of reasoning and decision-making in a complete, principled, and theoretically satisfying way, and ideally to have arguments that no other description is more satisfying. I’ll refer to this as a “complete axiomatic approach,” meaning that an end result of HRAD-style research on some aspect of reasoning would be a set of axioms that completely describe that aspect and that are chosen for their intrinsic desirability or for the desirability of the properties they entail. This property of HRAD work is the source of several of my reservations:
- I haven’t found any instances of complete axiomatic descriptions of AI systems being used to mitigate problems in those systems (e.g. to predict, postdict, explain, or fix them) or to design those systems in a way that avoids problems they’d otherwise face. AIXI and Solomonoff induction are particularly strong examples of work that is very close to HRAD, but don’t seem to have been applicable to real AI systems. While I think the most likely explanation for this lack of precedent is that complete axiomatic description is not a very promising approach, it could be that not enough effort has been spent in this direction for contingent reasons; I think that attempts at this would be very informative about HRAD’s expected usefulness, and seem like the most likely way that I’ll increase my credence in HRAD’s future applicability. (Two very accomplished machine learning researchers have told me that AIXI is a useful source of inspiration for their work; I think it’s plausible that e.g. logical uncertainty could serve a similar role, but this is a much weaker case for HRAD than the one I understand MIRI as making.) If HRAD work were likely to be applicable to advanced AI systems, it seems likely to me that some complete axiomatic descriptions (or early HRAD results) should be applicable to current AI systems, especially if advanced AI systems are similar to today’s.
- From conversations with researchers and from my own familiarity with the literature, my understanding is that it would be extremely difficult to relate today’s cutting-edge AI systems to complete axiomatic descriptions. It seems to me that very few researchers think this approach is promising relative to other kinds of theory work, and that when researchers have tried to describe modern machine learning methods in this way, their work has generally not been very successful (compared to other theoretical and experimental work) in increasing researchers’ understanding of the AI systems they are developing.
- It seems plausible that the kinds of axiomatic descriptions that HRAD work could produce would be too taxing to be usefully applied to any practical AI system. HRAD results would have to be applied to actual AI systems via theoretically satisfying approximation methods, and it seems plausible that this will not be possible (or that the approximation methods will not preserve most of the desirable properties entailed by the axiomatic descriptions). I haven’t gathered evidence about this question.
- It seems plausible that the conceptual framework and axioms chosen during HRAD work will be very different from the conceptual framework that would best describe how early advanced AI systems work. In theory, it may be possible to describe a recurrent neural network learning to predict future inputs as a particular approximation of Solomonoff induction, but in practice the differences in conceptual framework may be significant enough that this description would not actually be useful for understanding how neural networks work or how they might fail.