Ian Castles has been a frequent and welcome poster here. It must be very gratifying for him to see the following announcement on Oct. 12, 2005:
The [U.K.] Chancellor announced on 19 July 2005 that he had asked Sir Nicholas Stern to lead a major review of the economics of climate change, to understand more comprehensively the nature of the economic challenges and how they can be met, in the UK and globally.
The Terms of Reference for the review have now been announced and are annexed to this Press Notice. A call for evidence has also been issued.
The review will be taken forward jointly by the Cabinet Office and HM Treasury, and will report to the Prime Minister and Chancellor by Autumn 2006. It takes place within the context of existing national and international climate change policy.
Ian’s associate, David Henderson, has written the following background note on the Stern Review, which I’ve reproduced here.
David Henderson writes:
Over the past three years I and an Australian colleague, Ian Castles, have presented and developed a critique of the treatment of economic issues by the Intergovernmental Panel on Climate Change (IPCC). The Panel, through its Chairman, has dismissed us as purveyors of disinformation and described us as “Åso-called “two independent commentators”‘.
A feature that we have consistently stressed in our critique is the strange and culpable failure of treasuries and finance ministries to involve themselves in questions of climate change and policies relating to it. Despite the large amounts that are at stake, these ministries, including Her Majesty’s Treasury, have been content to leave the economic aspects to be dealt with exclusively, as they have seen fit, by environmental departments and agencies.
Whitehall showed no interest in what we had written. But in April 2004 Lord Taverne asked in the House of Lords “ÅWhether [the Government] are satisfied with the economic and statistical work of the Intergovernmental Panel on Climate Change’. On behalf of the government, Baroness Farrington replied:
“ÅMy Lords, we are satisfied that the economic work of the IPCC is the most comprehensive assessment available. We note that it represents consensus between governments based on careful analysis’.
This long-established official position remained unchallenged until this summer, when a new element appeared on the scene.
The new element is an impressive report, entitled “ÅEconomic Aspects of Climate Change’, from the House of Lords Select Committee on Economic Affairs, a high-powered body. The report was unanimous. In the opening paragraph of the Abstract, the Committee “Åcalls on the Government to give HM Treasury a more extensive role, both in examining the costs and benefits of climate change and presenting then to the United Kingdom public, and in the work of the [IPCC].’ The decision to launch the Stern Review was almost certainly prompted by this and other criticisms of the official government line, as also of the IPCC, voiced by the Select Committee.
Does the decision imply that government polices on climate change have been, or are about to be, radically recast? In my view, this is not the case. All the same, the government has shifted its position in one important respect, though without saying so.
The complacent words that her officials put into the mouth of Baroness Farrington in April last year, as quoted above, have been tacitly eaten. If the IPCC’s handling of economic issues was truly beyond challenge, the Stern Review would have no point. By commissioning the review, and by suggesting that it could contribute usefully to the IPCC’s Fourth Assessment Report which was launched almost two years ago, a British government has conceded for the first time that the IPCC process is less than wholly authoritative and could be improved from outside. This is a welcome though belated development.
David Henderson
Westminster Business School
David Henderson CMG was formerly Head of the Economics and Statistics Department of the Organisation for Economic Cooperation and Development (OECD). In 1985 he gave the BBC Reith Lectures. [Currently he is a Visiting Professor at the Westminster Business School in London.]
Climate Audit 
78 Comments
Unfortunately, although the economic arguments made by the Lords Committee were fairly sound, the scientific parts on climate change contained so many howlers that I cannot recommend it.
“… the scientific parts on climate change contained so many howlers that I cannot recommend it.”
For example?
re #2
Page 10, "The Uncertain Science of Climate Change", para 6
No, it ain’t. The Sun is very definitely a weakly variable star, with both short, mid and long term variability in its output.
Same para, next few sentences:
No it doesn’t. I think McKitrick and Essex might have something to say about the concept of a "global temperature", as well.
Further down the same paragraph, on energy balance:
Why is water vapor separated from other greenhouse gases? Also the whole paragraph suggests a static equilibrium of energy input to output, which is just completely at odds with practically every investigation done on past climatic variation. It does not mention for example, that water vapor completely dominates the greenhouse effect.
Box 2, Page 11, "The Greenhouse Gases"
No mention of the far-and-away predominent greenhouse gas, water vapor.
The whole of para 7:
This is highly disputable, tendentious or just plain wrong. Again the authors assume the natural climate has a "balance" which holds it steady, and that the tiny amount of extra carbon dioxide produced by man relative to the natural background is somehow "tipping the balance". That paragraph paraphrases Peter Hearnden’s output in another thread comment, again without evidence for where this balance exists, or that the climate is stable without this "tipping of the balance"
And so on.
The science part is a compendium of oversimplified, tendentious or just scientifically wrong statements. I’ve no doubt that others with a better science background than I, could spot even more. Steve McIntyre is given a PhD according to the text, something I never recall Steve ever claiming.
Is that enough, Ian?
sigh… Bookkeeping John A, it’s the NET Co2 emission that counts, not the bulk flux.
For what does it count? My point was that the paragraph assumed an unstable equilibrium for its radiative forcings.
There ia a very interesting post on Jennifer Marohasy’s blog :http://www.jennifermarohasy.com/blog/archives/000938.html, which has the testimony of Lord Nigel Lawson to the Senate Committee on Environment and Public Works. His conclusion is that the IPCC has outlived it’s usefulness and should be closed down.
A primary source for his testimony is http://epw.senate.gov/hearing_statements.cfm?id=246944
John A can’t recommend the House of Lords Committee report on “The Economics of Climate Change” because “the scientific parts on climate change contained so many howlers”. The US Senate Committee on Environment and Public Works thought otherwise and circulated copies of the report to its members before Lord Nigel Lawson, former Chancellor of the Exchequer, appeared before it last month.
Lord Lawson explained to the Committee that “the Economic Affairs Committee is one of the four permanent investigative committees of the House of Lords, and fulfils one of the major roles of our second chamber as a forum of independent expertise and review of all UK government activity. It is composed of members of all three main political parties. Its climate change report, which was agreed unanimously, was published on 6 July 2005, just ahead of the G8 summit at Gleneagles in Scotland.”
In less than six months after deciding to undertake the inquiry, the Committee had produced an 84-page report and a 310-page volume of evidence. They had examined, among others, Dr. Pachauri (Chair, IPCC), Sir David King, Sir John Houghton, British departmental officials and Professors Colin Robertson, David Henderson, Richard Lindzen, Bjorn Lomborg, Richard Tol, Nebojsa Nakicenovic, Dennis Anderson and Michael Grubb. The thoroughness with which their Lordships approached their task is evident in many of the 400-odd questions they asked of the eminent scientists and economists that appeared before them.
The Committee also received voluminous written statements, both from those who appeared before them and from others. Many of the latter are published in the Committee’s Evidence volume. I would especially recommend reading those submitted by Dr. (sic) Ian Castles (I’m given a Ph. D according to the heading – something I never recall ever claiming) and Professors Angus Maddison, Ross McKitrick, Julian Morris, Paul Reiter and Fred Singer.
In his written submission to the Committee, Professor David Henderson said:
“It may be – though this goes outside my area of competence – that the IPCC peer review process is likewise under challenge in the debate that is now in progress about the validity of the famous ‘hockey stick’ diagram which Lord Lawson referred to at the Committee’s first meeting … The Committee may wish to keep an eye on the debate, if only because of the weight that has been placed on this particular piece of evidence.”
The Committee duly devoted 2 pages of its report to a review of the hockey stick debate, in the course of which they perpetrated two “howlers” (referring TWICE to Dr. McIntyre).
By way of placing these errors in perspective, my submission to the Committee drew attention “to eight further errors in the [IPCC Special Report on Emissions Scenarios] and/or in the responses of the SRES Teams to our critique, in addition to those identified in our previous papers [published over the preceding two-and-a-half years]. The nature of these errors is evident from the sub-headings in my submission: “Claim that energy intensities in developing countries are comparable with past levels in the now-industrialised countries”, “Omission of United Kingdom from ‘Definition of SRES World Regions'”, “Differences in Estimates of Regional and Global GDP in the Base Year of the Projections”, “Lack of Correspondence between global total and estimates for regions”, “Inconsistent assumptions about scenarios parameters”, “Confusion between GDP(PPP) and GDP(MER) growth rates”, “Erroneous calculzation of Growth rates for the 1990-2000 decade” and “Claim that Castles and Henderson endorsed MESSAGE PPP scenarios.”
Most of these errors, and the more serious ones which David Henderson and I had previously brought to notice, went unrecognised or disregarded by the 53 authors, 89 expert reviewers and numerous though uncounted government officials who were involved in the preparation of the SRES.
Apart from the note on the ‘hockey stick’, the Committee’s short section on the science of climate change does not purport to be other than an explanation of the “strong majority view”. They cite the IPCC 2001 main scientific report, the Royal Society’s submission, Sir David King’s oral evidence, Sir John Houghton’s “Global Warming: the Complete Briefing”, and M. Maslin’s “Global Warming: A very Short Introduction”.
But they also cite Professor Richard Lindzen’s dissenting view of the scale of the enhanced greenhouse effect, and they register their concern “that the IPCC has not always sought to ensure that dissenting views are given a full hearing” (para. 14). They document these concerns in some detail, not only in relation to the economics (citing Castles, Henderson, McKitrick and Tol) but
also the science (see the discussion of Professor Paul Reiter’s experience in paras. 115-116).
I’m unmoved by John A’s comment that “McKitrick and Essex might have something to say about the concept of a ‘global temperature’, as well.” Some of the Committee members may well have known the views of M&E on this point, or have been told of it by the Committee’s adviser, the deeply respected environmental economist David Pearce (whose sudden death since the publication of the Committee’s report is an enormous loss to the discipline). A unanimous report of this calibre could never have been produced by a committee whose members argued about everything about which someone had had “something to say.”
The report is lucidly written and makes eminent good sense. Many of the members of the Committee would make no claims to economic expertise (although several of them certainly can). One can only wish that the economic content of the IPCC’s reports was up to the standard of the scientific content of this Committee’s report. It will stand as lasting evidence of the great merits of the system that produced it. And, without it, the British Government would certainly not have taken such a dramatic U-turn about the need for Treasury involvement in climate change policy.
John A’ reference to Essex and McKitrick on global temperature estimates suggests to me that he is thinking of the impossibility of doing that by simply averaging, whether on cell grids or other arbitrary boundaries, of station temperatures.
The problem is that temperature is an intensive variable which is used to factor an extensive variable like mass, to numerically quantify thermal energy associated with a particular mass. Just because an ice cube at 0 deg Celsius and an million tonne iceberg at 0 deg celsius does not mean that there is any sensible equivalence with the two readings of 0 deg Celsius except that both physical objects are at the temp of 0 deg celsius.
My own experience is that most scientists have forgotten about the crucial difference between intensive and extensive variables, and those who graduated in the humanities and social sciences would not understand it at all.
There is a reason why the meteorologists cannot get their prediction for weather right, whether short or long term.
In any case I suspect there is an additional source of energy that the earth receives, which dwarfs solar energy- electrical that the plasma physicists and cosmologists work with, and that they can test in laboratory settings. Except it is, at the moment, scientific heresy.
Louis,
You might want to clarify your statement:
“In any case I suspect there is an additional source of energy that the earth receives, which dwarfs solar energy- electrical that the plasma physicists and cosmologists work with, and that they can test in laboratory settings.”
particularly the “dwarfs solar energy” part. As worded, your statement seems to me to suggest that Earth receives more energy from GCRs than from solar electromagnetic (i.e. light) energy, plus “solar wind”, and, or including, CME, energy.
Re #3 and #*: I’m with Ian on this. John A., I think that the statements that you’re objecting to are all reasonable as first-order statements and that you’re losing sight of what the committee is doing in making first-order summaries.
1) solar radiation is "fairly" constant over time. It varies a little, but for the purposes of a first-order exposition, this is surely true. There’s no harm in saying this to a lay audience.
2) there’s a little inventory change in energy on an annual basis, but, again, on a first-order basis, energy in does more or less equal energy out. Again I see no harm in saying this to a lay audience.
3) "Why is water vapor separated from other greenhouse gases?" For some purposes, you’d group them together; but for purposes of discussing policy, I see no problem with distinguishing water vapor. It’s role is pretty unique. They are not denying a greenhosue role to watrer vapor.
4) Your objection to Box 2 again is definitional. This is hardly a "howler". They’ve not lost track of water vapor. They are discussing the gases with policy aspects to them.
5) on the increases in CO2 – again, their paragraph seems fairly written to me. Objectively there have been increases in CO2 concentrations. It’s sensible to inquire into the impact. The authors are not necessarily pre-judging the impact, merely stating the issue. I think that it’s fair enough.
I haven’t been through the report, but I disagree with your view that the quotes are "howlers".
Re #9, I had read John A’s #27 on the “Weather and Climatology: Mandelbrot’s view” thread, and the various views expressed in #28 – #38 on the same thread. I noted in particular Ross McKitrick’s comment (#38) that he wouldn’t wear out his welcome with Steve “by provoking an average temperature donnybrook on this thread.” My comments in the penultimate para. of #8 were not intended to imply that the issue of whether there was such a thing as “global temperature” was unimportant, nor did I intend to reflect in any way upon the Essex & McKitrick position on this matter.
But the Committee’s “dominant concern [was] with certain aspects of the economics of climate change”, and their short preliminary discussion of the underlying science must be read in this context.
John A characterises what the Commitee says in their para. 7 as “highly disputable, tendentious or just plain wrong”, and alleges that “the science part is a compendium of oversimplified, tendentious or just scientifically wrong statements.” He claims that others with a better science background than he has could spot even more such statements. Well, John A and others with a better science background might not go along with all of the statements made by Sir John Houghton in evidence to the Committee:
“These are the things we know. First of all, carbon dioxide has increased by over 30 per cent since the beginning of the industrial revolution. We are sure, absolutely certain, that the reason for most of that is the burning of fossil fuels by the world’s industry… Secondly, we are quite certain about the basic science of the greenhouse effect, the fact that we have gases in the atmosphere, carbon dioxide and methane in particular, that absorb radiation admitted by the surface … and stop it going out to space and act as a blanket over the earth’s surface and make the earth warmer than it would otherwise be… I think the reason we believe in human induced climate change is because the basic science is there, the carbon dioxide is there, and … the story as we read it in the climate record over the last century is also there.”
Far from accepting the certainties that were put to them by Sir John Houghton and Sir David King, the Committee pointed to “the divergence betweeen actual and expected recent past warming” and to 8 sources of doubt about other features of the accepted science (para. 17). Their conclusion was:
“We do not propose to evaluate these doubts, nor are we qualified to do so… But the science of climate change remains debatable. We heard from witnesses who seemed in no doubt at all about the science, while others expressed one or more of the above concerns. THAT MAKES IT CLEAR THAT THE SCIENTIFIC CONTEXT IS ONE OF UNCERTAINTY” … (para. 18, EMPHASIS in original).
These statements would not have been well received by Britain’s science establishment, and they are at odds with the loudly-proclaimed views of British Ministers. John A may wish that the Committee had gone further and asserted that the science establishment’s certainties were certainly wrong. They did not feel qualified to take such a position, and if they had done so they would not have brought about the U-Turn in the British Government’s position on the need for a more rigorous approach to the economics of climate change that is signalled by the estanlishment of the Stern Committee.
I posted #12 before seeing Steve’s post #11, with which of course I agree.
Steve McIntyre doesn’t want any discussion of average temperature here because he knows that Essex and McKitrick are wrong.
Ian and John A, you are both right.
The House of Lords report is a very good effort by non-experts to understand what is going on. They have tumbled to quite a few of the warmers’ tricks. At times one feels their surprise to find that one or other aspect of the popular understanding of this issue is dubious, contested, or probably wrong.
But they do make lots of little mistakes of unfamiliarity. Mauna Loa measurements started in March 1958, not 1959 (p. 12); nitrous oxide concentrations are less variable than carbon dioxide, not more variable (p. 12); the dates of some IPCC reports are a year out (not all of them, because the 1995 report did come out in 1996; p. 15); the IPCC’s predicted sea level rise was 9 to 88 cm, not 20 to 88 cm (p. 22).
Also, where they didn’t get an expert, they often swallowed a dubious “orthodox” line. Munich Re’s “fivefold increase in the number of weather-related extreme events” from the 50s to the 90s (p. 24) is ludicrous as a statement about climate. The “evidence that CO2 concentrations “lead’ temperature rather than the other way round” (footnote 5) is actually weaker than the contrary evidence. Soot is thought to have a warming effect, not a cooling effect, except when it reaches the stratosphere (p. 12).
More broadly, the scientific section is trapped in the “radiative forcing” model that underlies the greenhouse paradigm. It betrays little appreciation of the importance of thermodynamics to earth temperatures, and no notion of the fact that convection, turbulence, water vapour and clouds are the prime regulators. There remains a “touching faith” (Lindzen) in climate models.
Nevertheless, like Ian, I rejoice at the House of Lords report. It brings some long-needed sobriety and balance to this issue, and makes a strong case for serious economic assessment of the risks of climate change. Hopefully, further investigation will show just how hypothetical — and how hyped — those risks are.
Much as I’d like to defer to Ian and Steve in their general knowledgeability on economics and mathematics (respectively), I still baulk at the idea that because a scientific description is misleading, it can somehow be “reasonable as first-order statements”.
I’ve heard the Mann Hockey Stick described by people as “correct to a first order” (and I didn’t mean that statement to be taken as pique, Steve).
First order statements can be misleading. It is disappointing that the Lords did not get an astrophysicist involved, because I’m sure that the content of those sections would be radically different.
If Steve would allow me the privilege of analysing the scientific content of the Lord’s Report in a series of articles (no more than 3 required), I’ll explain why in better detail and with some nice diagrams to boot.
Re #16, the report was produced by the House of Lords Select Committee on Economic Affairs and its subject was the ECONOMICS of climate change. In that sphere, the Committee produced a very impressive report very quickly. An astrophysicist would have wondered why she’d been asked to give evidence.
That said, I don’t agree that the report can’t be recommended or that there are “howlers’ in the short scientific section — and I think that David’s criticism of this section of the report for “little mistakes of unfamiliarity” (#15) is hypercritical. I’d argue that David is judging it against higher standards than he applies to the IPCC or to himself.
For example, their Lordships say, correctly, that the IPCC “projects” a mean global sea level rise and, incorrectly, that that projection is for 20 cm (rather than 9 cm) to 88 cm by 2100. David says, incorrectly, that the range in question relates to the IPCC’s “predicted” sea level rise, although the Panel is at pains to emphasise that their projections should not be treated as predictions (though they know perfectly well that this is how they will be interpreted).
The IPCC scientists made mistakes of their own on sea level rise. Take a look at Figure 11.12 in Chapter 11 (“Changes in Sea Level”) of IPCC, 2001, Climate Change 2001: The Scientific Basis, and note that it is not consistent with Table II.5.2 in “Appendix II SRES Tables” in the same report (the table is at http://www.grida.no/climate/ipcc_tar/wg1/553.htm , and provides links to the Figure ). The text under the “Models minimum” panel of the Table states that “The final values of these timeseries correspond to the lower limit of the coloured bars on the right-hand side of Chapter 11, Figure 11.12”, and the text under the “Models maximum” panel makes a similar statement with respect to the upper limit of these coloured bars. But it’s obvious that the final values of the time series relating to two of the illustrative scenarios — A1FI (FI = “fuel intensive”) and A1T (T = “technology”) – DON’T correspond to the upper and lower limits of the coloured bars. Figure 11.12 shows (correctly) that the sea level rise is greatest for the A1FI scenario, whereas Table II.5.2 shows (incorrectly) that the “Models maximum” and “Models minimum” sea level rise is greatest for the A1T scenario (though for the “Models average” the sea level rise is far greater for the A1FI scenario).
These errors may simply reflect the fact that the column headings for the two scenarios have been accidentally transposed in the second and third panels of Table II.5.2, but it may be worse than this. As the base levels and emissions in 1990 and in 2000 were standardised for all of the scenarios (see Box 5-1 of SRES at http://www.grida.no/climate/ipcc/emission/115.htm ), it is possible that reported sea levels in 2100 in one scenario have been compared with the originally reported sea levels for the base period in a different scenario (depending on the stage in the tabulation process that the transposition error occurred). Even if the only flaw is the transposed column headings, it’s a pretty embarrassing mistake in a report that took years to complete, went through the allegedly rigorous review processes of the IPCC and was approved by 180 governments.
OK, Mauna Loa measurements started in March 1958, not 1959. According to the digital data available at http://cdiac.esd.ornl.gov/ftp/trends/co2/maunaloa.co2 , the average CO2 concentration in 1958 was –99.99 ppm and that for 1959 was 316.00 ppm. So their Lordships should have said that 1959 was the first year available for the annual time series, not the year in which measurements began.
This is a trivial error, as are the other errors of one year in the date of publication of some IPCC reports in the Lords Committee’s Box 4 (Incidentally, does the IPCC itself present such a clear and succinct listing of its main publications, and how they relate to one another over time and between the Working Groups, as the Lords Committee Report does in this box?).
The IPCC’s errors in this genre are far more numerous and far more serious than those in the Lords Committee Report.
For example, according to the fourth paragraph of Section 1.3.1 of Climate Change 2001: Mitigation (available at http://www.grida.no/climate/ipcc_tar/wg3/057.htm ) “the Human Development Report 1989 (UNDP, 1989)” illustrated the distributional dimensions of global poverty in the form that has come to be known as the champagne glass. In fact, there was no HDR 1989: the first such report was entitled HDR 1990, it was published in 1990, and it didn’t feature the champagne glass or the grossly misleading figures that were later to be illustrated in the champagne glass. These tabulations date from HDR 1992, published in 1992. The statement in the same paragraph that “inequality has widened further since then” is also wrong, and the figures in support of this statement in the cited “(UNDP, 1999)” actually relate to 1997 (but were found by an expert group that reported to the United Nations Statistical Commission in 2000 to be a “material error”). Several pages further on (p. 97 in the printed version), Table 1.1 is headed “Per capita income and carbon emissions in various regions” – but there are no data relating to per capita income in the table.
Two chapters in the IPCC WG II report (2001) – Chapters 1 and 9 – also uncritically report the same material error as that noted in the previous paragraph, and in the reference list to Chapter 9 the UNDP publication is wrongly cited as the WORLD Development Report rather than the HUMAN Development Report.
I also think it’s hypercritical to say that the Lords displayed a “touching faith” in climate models. They said in effect that they were not qualified to challenge the majority view, but they didn’t say that the majority view was beyond challenge. In terms of raising doubts about the certainties that the British and many other governments are proclaiming about climate change, I’d say the report has been more effective that if they’d directly challenged the consensus view.
Of course, individual members of the Committee have their own views, and Lord Lawson has given his to the US Senate Committee (see #6 and #7) – a very impressive performance in my opinion.
Ian,
I follow your arguments about the economic content pretty well, and I can understand why you’re excited about it. But the scientific content is just poor, and I don’t mean just the parts the repeat the IPCC’s "analysis". I mean straightforward facts and inferences that their Lordships make which undermine (or could undermine) the seriousness with which those economic judgments are made.
As for
I’d like to think you’d know better, since the central points made in the scientific summary are astrophysical in nature, and directly relate to how much the climate changes as a result of changes in "greenhouse gases".
For example, their Lordships boldly state that the Sun’s luminosity is constant. If they had had an astrophysicist to refer to, that astrophysicist would not have recommended such a bald statement especially as it is untrue. I have asked an astrophysicist and that astrophysicist’s reaction was thus:
Let me explain in terms you would understand: The Dow Jones Industrial average has traded between 9,660.18 and 11,027.20 over the last 52 weeks, and as far as I can gather, has traded between those ranges from late 2003.
If I were to write an economic report that stated that the DJIA was a constant "to a first approximation" would you consider that statement
a) true
b) superficially true
c) misleading or
d) false?
If we are talking about climatic change, which we agree is on the timescales at least of decades, if not centuries or millenia, then to state that for the period we are considering, the Sun may be regarded as having constant luminosity and the Earth to have perfect radiative balance during that time EXCEPT FOR RELATIVELY RECENTLY, is not, in my view, a reasonable statement of science. It is oversimplistic, misleading and/or false.
Having made the gaffe of excluding the greatest source of climatic variability, their Lordships have no choice but to reiterate (with the caveat that they are unqualified to make a judgment) a restatement of the greehouse theory with its current emphasis on "tipping points" and "pushing the Earth’s climate out of equilibrium" intact.
They go even further and restate the Radiative Equilibrium model of the Earth’s climate as if it were an established fact, rather than an incredible oversimplification of the Earth’s climate system used by modellers, The Earth’s ocean/atmosphere system is dominated by a system of energy transport that is not primarily radiative, but convective and turbulent.
No competent astrophysicist would have allowed their Lordships to make such basic errors. But they weren’t to know that, and nobody told them.
John A, I can only repeat that the subject of the report is The Economics of Climate Change, & note that I’ve already said that I agree with Steve’s view that “the statements that you’re objecting to are all reasonable as first-order statements and that you’re losing sight of what the committee is doing in making first-order summaries.”
Ian,
If the scientific statements are wrong, they will be used as an excuse to rubbish the economic arguments. That’s the reality, political and scientific.
Do you think that Steve McIntyre, or Ross McKitrick or Chris Essex have been lambasted over tiny details of their mathematical or physical arguments because people like to discuss these things as academic curiosities?
In the first instance, whatever we feel about the economic arguments put forward in the Committee’s report, we should at least seek to correct very obvious mistakes which will be pulled out and used as political bludgeons to demolish those same arguments.
I admire Nigel Lawson for his speech in front of the Congressional Committee. In more normal times such lines as
and especially
…would have been splashed as major global headlines beginning with “IPCC crisis: Former UK Chancellor calls for abolition of IPCC“.
As I said, normal times. Which we don’t live in.
The Devil is in the details and this is Hell.
OK very late after the event; I have started reading the minutes of the House Lords select committee proceedings.
It is an extremely good read!
If there is anyone else out there who hasn’t read it -do give it a go:
Click to access 12ii.pdf
It is a good read. Just don’t inhale.
Hi,
With all due respect to Mssrs. Castles and Henderson, the simple fact is that it’s extremely likely that virtually the entire economics profession is basically clueless about what will happen to world GDP in the 21st century.
The world per-capita GDP in the year 2000 was approximately 7200 USD. It is quite likely that the world per-capita GDP in the year 2100, expressed in year 2000 dollars, will be more than 1000 times higher (i.e., more than 7,000,000 USD):
http://www.longbets.org/194
http://markbahner.typepad.com/random_thoughts/2004/10/3rd_thoughts_on.html
http://www.techcentralstation.com/101805C.html
Given the likely cluelessness of most of the economics profession regarding the economics of the next 100 years, I have doubts about how much most economists can contribute to the debate about climate change.
That’s a 7% CAGR. Was that seen for the last century? The one before? What is current CAGR?
World per capita GDP grew at a compound annual rate of about 1.6% in the 20th century. The CAGR in the century before that was less than 1% (Source: “The World Economy: Historical Statistics”, p. 263). The current CAGR is around 3%.
PS. “The World Economy: Historical Statistics” is by Angus Maddison and published by the OECD. According to Maddison’s estimates the average rate of growth in world per capita output between 0 and 1000 was nil, and the average between 1000 and 1820 was 0.05% per annum.
TCO writes, “That’s a 7% CAGR.”
Yes, a factor of 1000 represents a 7 percent per year compound annual growth rate (CAGR).
“Was that seen for the last century?”
No, as pointed out by Ian Castles, Angus Maddison estimated that rate for the 20th century was 1.6 percent per year. (Another economist, Brad DeLong, estimated a value more like 2.2 percent per year.)
“The one before?”
As pointed out by Ian Castles, growth in the 19th century averaged less than 1 percent per year.
“What is the current CAGR?”
Again, as pointed out by Ian Castles, the current per-capita world GDP growth rate is about 3% per year.
But that is exactly my point. Your questions seem to be heading in the direction of the vast majority of economists. But that does NOT mean you’re headed in the right direction! In fact, if you head in that direction, I predict that your per-capita GDP value will be off by more than a factor of 100 by the year 2100.
What you seem to be missing is the reason why the 21st century will be dramatically different than any other previous century: by the year 2100, the effective human population will be infinite.(!) Therefore, the economic growth rate by 2100 will be large almost beyond belief. There are two fundamental questions to ask, with answers that show why this is so:
1) “What causes economic growth?”–>If you’ve read Julian Simon, you know that (free) human minds cause economic growth. The more free human minds you have, the higher your economic growth rate.
2) “How many human mind (equivalents) will there be in 2100?”–>Here, you need to switch your reading to Ray Kurzweil. If you read Ray Kurzweil, you’ll know that the number of (equivalent) human minds in the year 2100 will approach infinity.
Consider this: Today, there are roughly 200 million personal computers produced, each with the capability of maybe 1/1000th of a human brain. But the capability of personal computers is doubling every…let’s say 2-5 years. And 10 doubling periods is a factor of 1000. Therefore, even with the capability doubling every 5 years, a personal computer will have the capability of a human brain circa 2050…and will be 1000 times more powerful than a human brain circa 2100.
Therefore, even if ONLY 200 million personal computers are produced in 2100 (far MORE will be produced), and each has a capability of 1000 times the human brain, the effective human population increase will be 200 billion per year. So essentially, come 2100, the number of (equivalent) human brains will be infinite. Therefore, economic growth will be spectacularly high.
If you didn’t have a chance to read this article, I heartily recommend it:
http://www.techcentralstation.com/101805C.html
From that article:
P.S. My prediction is testable in the short term, even though my prediction isn’t quite so bold as Ray Kurzweil’s/Arnold Kling’s. I predict that the world will see a few years with per-capita CAGRs above 4 percent, and maybe even above 5 percent, even by 2020. If that is indeed true, I think we can be extremely confident that the vast majority of the economics profession is going to be spectacularly wrong about economic growth in the 21st century.
So once again, I don’t disrespect Ian Castles and David Henderson at all for trying to add some science to the IPCC’s analyses. However, given the fact that virtually the entire economics profession currently spectacularly underestimates the likely economic growth in the 21st century, I don’t think additional conventional economic growth perspectives will add much to the IPCC’s analyses. Unless someone in HM Treasury knows Ray Kurzweil’s work, and understands and accepts the implications of his work for economic growth, the economic growth projections will be spectacularly wrong (on the low side).
Re: 27, Mark, thank you for a very interesting post. You say
While this is a tempting argument, I fear that there are a few holes in it.
1) The current capacity of a computer is far, far, far less than that of a human brain. The brain contains on the order of 100 billion processors (neurons, each of which is a processor). A current computer has one. This is not a thousand to one, but a hundred billion to one. Thus using your figures, the capacity of the computers will not be a thousand times more powerful, but on the order of 1/20,000 as powerful as a human brain by 2100.
2) Number of processors and number of computers and size of memory capacity miss the point. When was the last time a computer came up with a new idea? When was the last time a computer was awarded a patent for creating a new device? To date, at least, these are exclusively the province of humans. While this might change, it also might not … and our increases in productivity are tied, at least in part, to new ideas and new devices.
3) According to the productivity report cited in the Kurzweil article, the productivity in 1984 was greater than the productivity in 2004. Meanwhile, computer power during that time increased by a huge factor, and computers came into common use in business. Why hasn’t your projected rise in productivity already occurred? And given the fact that it has not occurred, why do you think it will suddenly start to occur?
(This, by the way, has been called among other things the “computerisation paradox”. It refers to the fact that many people, not only yourself, confidently predicted that productivity would rise greatly as computers became commonplace between 1980 and the present. This has most assuredly not happened, and that’s the paradox.)
4) Your argument reminds me of the argument that two fruit flies, in only 10 weeks, will have 1,425,219,894,286,254 descendents. Yes, they would if there were no other constraints to growth … but the reality is that there are always constraints to growth, even in productivity.
So, while rosy predictions sound good and make good copy for articles, in the real world things move much more slowly. Limitations of things like education, transportation, materials availability, energy costs, and Murphy’s law always conspire to bring things down to a more prosaic growth rate.
w.
Willis,
I don’t know that I agree with this computers uber alles idea either, but you are somewhat underestimating their present capabilities. For one thing the cycle rate of a human neuron is slow. I’m not surewhat it is exactly but at most it’s a few kilohertz. a computer processor is a giga hertz or two, I.e. several thousand times as fast at least. Also a computer processor typically can address more memory, i.e. more other neurons than a human neuron. Finally there’s less overhead in a computer since it doesn’t have a body. Even the externals it does have usually have there own built-in processors these days too. So there’s less difference than is apparent.
What you are all forgetting about is the basic assumption that the human mind is just a large, highly parallel, organic general-purpose computer with some complex software. But we really don’t know how the human brain works, let alone the *mind*. Our capacity for abstract thought and the ability use symbolic information is nothing short of amazing. I’m not even sure we will be able to understand our brains well enough in the next 50 years to even begin the process of designing a thinking machine. Even if we did I suspect there is a universal law that says you can either have fast processing and perfect memory recall, or you can have sentience, but not both. And even if I’m wrong, you’d have to empacipate these machines if they had free-will, so what’s the point. We can already create another intellegence anytime we want – it’s called a child.
I’ve been thinking a bit more about the gains in productivity in the late 90s and early 00s. I suspect that they had much more to do with improvements in communications than in computers. In particular, the advent of both email and the web have greatly improved my own productivity, and I suggest that of others as well.
There are two things about these improvements in communication. One is that, although they require computers, their speed is not particularly changed by fast computers. An old Windows 98 computer will send an email about as fast as a Cray. The second is that we are never going to speed up emails much, they are nearly instantaneous as it is. So in the arena of email, and communications in general, there’s not a whole lot of room for speed increases.
Unfortunately, the same is true of a number of the things that we do. For example, I spend a good chunk of my day writing letters, faxes, and emails. The speed limitation on all of these is not the computer. It is me. My computer has weeks of computer time to think about other things in between each of my keystrokes. I’m typing on my trusty old Mac. Is there anyone who thinks that I would be typing faster if I had a Cray?
Faster computers may speed this up eventually, using voice input. I work in a small office with 6 desks in it, though, and a lot of talking and through traffic, so I suspect I won’t be able to use it. But even if I could, I type at about 70 wpm, and speech is at about 250-300 wpm, so it will be a max 4x speedup even if the dictation programme is letter perfect. And I still need to think up the letter, and proofread it and format it and correct it when it is done, so the increase will be even less.
The ugly reality is that many of the things that I do, from writing letters to sending emails to creating spreadsheets to pounding nails, will not benefit in the slightest from faster computers. How much time in a day do you spend waiting for your computer? In most cases, very little. The odd program in R may take a few minutes to run, but much more time than that to write. My computer spends much more time waiting for me than I spend waiting for my computer.
In general, then, I would say that we have already skimmed much of the cream from the computer and communication revolution, and that barring the invention of truly autonomous computers, faster computers won’t make that much difference. It won’t speed up typing, or emails, or the overwhelming majority of the tasks that fill up our working lives.
So to me, Kurzweil’s claim that we will see a giant increase in productivity in the coming decades seems extremely doubtful. Yes, we will continue to get faster and slicker and quicker, but not at any giant rate of speed.
w.
Yes, and given 20 years of food, room and board, and education, and that child will *probably* be able to beat a chess computer program I can download for free from the Internet …at least if the program is set on a low level.
I’m not putting down the human brain, but by the end of this century, it’s virtually certain that computers will be to the human brain what a human brain is to a dog brain…and probably even a worm brain.
And they will be building themselves at a rate of billions…or even trillions…per year. Finally, they will be able to teach themselves what takes us many years to teach a child in a matter of minutes or seconds.
P.S. I refer to them as “them”…but I don’t expect any of “us” to settle for our dog/worm brains, when better electronic ones are available.
re: 32, Thanks for posting, Mark. I’d be interested in your response to some of the issues I raised in my posts regarding your claim.
In your most recent post, you say “by the end of this century, it’s virtually certain that computers will be to the human brain what a human brain is to a dog brain…and probably even a worm brain.” Near as I can tell, nothing about the year 2100 is virtually certain … but I digress. What I wanted to comment on is this:
Many, many such predictions have been made about “artificial intelligence”. When computers were first built, many predictions like yours were made, except in those days they were regarding the year 2000. (I wrote my first computer program in 1963, and I remember well the hype and the buzz.)
Artificial intelligence, however, has the distinction of being perhaps the slowest progressing of all of the fields of computer science. After 50 years of trying, we don’t even have a computer yet which can pass the Turing test, much less think independently. Don’t mistake processor speed and memory size for advances. Despite years of these advances, the most modern computers are still totally incapable of independent thought. In fact, as the Turing test shows, they are not yet even able to successfully mimic independent thought.
The real issue, though, is not that. It is that for most jobs, more computer power does not necessarily translate into more productivity. If we replaced the computer on your desktop with a supercomputer, how much do you think your productivity would increase?
My guess?
Not much at all.
Finally, regarding the basic issue in question, that of estimating future gains in productivity. While your claims are possible, they are certainly not what I would use if I were attempting to estimate future gains. This is because, while they are possible, there is no evidence that they are actually occurring. And there is evidence that they are not occurring. As I mentioned before, productivity in 1984 was greater than in 2004. Given that, I’d have to see something hard in the way of evidence before I relied on your claims.
w.
Actually when Mark first mentioned astonishing increases in wealth by the end of the century I thougth he was going to talk about the area of technology which actually can deliver on that promise, nanotechnology. It’s a sleeper science slowly gaining momentum, but once it actually starts working up to theory, and there doesn’t seem to be real problems which don’t just require a lot of hard work and brainpower, the wealth increase will be stupendous. So far just the first stages of the needed nano-infrastructure are being developed, but as more and more of them are developed an eventual tipping point will be reached and changes may well be too fast for humans to keep up. Of course controlling nanosystems will require huge amounts of computing power and while humans will control the overall process, it will be the computer-nanosystem interface which will be the most intense and this will be far more the source of future wealth than improved e-mailing.
Of course when anyone anywhere can essentially have anything he / she /it can imagine at very little cost, the entire economic system is going to be a lot different than it is now. Most of us may well make our money by helping others on the imagination end of things.
I’d like to second Willis’ response to Dave’s comment. Progress in artificial intelligence has proceeded at a snail’s pace over the last 25 years. It’s one thing to invent chess programs, and quite another to replicate or even mimic independent thought.
Even when I studied neuropsychology in the early 1980s the possibilities conceded by practitioners were very limited and narrow. Some of the very narrow targets have been achieved – for example, we can now associate electrical activation in parts of the brain with various emotions and behaviours. Practical advances associated with this research mean that neural activation of prothsetic limbs is now crudely possible.
However many other aspects of cortical function remain as much of a mystery as they were 25 years ago. Setting aside even the fundamental problem of consciousness, the problem of memory remains intractable. It is completely inappropriate to compare computer memory to human memory.
A computer can “remember” a bunch of keystrokes forever. I can remember, 30 years on, my first kiss – where I was, what she was wearing, her perfume, how it made me feel. But as to where these memories are stored in the brain, or the mechanism that enables me to recall them in an instant, nobody has a clue.
Computer power can increase exponentially. Brute force will produce unbeatable chess programs. But it will be a long, long, time (or maybe never) before a computer, on its own, will invent Velcro.
Willis Eschenbach writes, “Thanks for posting, Mark. I’d be interested in your response to some of the issues I raised in my posts regarding your claim.”
OK. 🙂
Each neuron is not a “processor.” A neuron is essentially a switch, so a neuron is like a transistor, not a “processor.” The brain gets its power from the fact that the switches are massively parallel, since the switches (neurons) are incredibly slow.
Today. Yesterday. The day before. Ray Kurzweil has programmed computers to write poetry and make up jokes. Computers are being used not simply to analyze mechanical designs, but to develop their own mechanical designs. Read some of Ray Kurzweil’s books. Subscribe to MIT’s “Technology Review.”
So? You’re so clearly picking a “cherry,” it isn’t even worth discussing. You are picking that *one* point, in order to completely disagree with the entire conclusion of the article:
http://www.frbsf.org/publications/economics/letter/2005/el2005-05.html
From Figure 1, it’s pretty clear that a regression line through *all* the data would show an upward trend. Why didn’t you pick 1979? Or 1981? Or 1985? Or 1988? You didn’t pick those years because only 1984 supports your hypothesis!
And in any case, the article is about **U.S.** productivity, not world productivity. Show me data indicating the world productivity growth rate has gone down since the 1980s.
If “my argument” reminds you of that, then you complete misunderstand “my argument.”
The world per-capita GDP in the year 2000 (in 2000 USD) was approximately 7200. My prediction in Long Bets #194 is that world per-capita GDP (in year 2000 USD) will exceed 13,000 in 2020, 31,000 in 2040, 100,000 in 2060, 1,000,000 in 2080, and 10,000,000 in 2100. These values represent average annual per-capita GDP growth rates of approximately 3%, 4.5%, 7.5%, 10.5% and 12.0% for periods ending 2020, 2040, 2060, 2080, and 2100, respectively.
My prediction can be contrasted with the IPCC minimum, median, and maximum world per-capita GDP in 2100 (year 1990 dollars) of 14,600, 44,800, and 129,000, respectively. Even the *maximum* value projected by the IPCC represents an average growth rate of only 3.0% per year for the entire century.
If you think my prediction is wrong, please go to the Long Bets website, and vote against it. The voting is free.
http://www.longbets.org/194
Mark, growth in per capita GDP in ALL of the major rich (computer-using) countries was lower in the fourth quarter of the twentieth century (1975-2000) than the third quarter of the century (1950-75). On your argument it should have been the other way around. This isn’t “cherry-picking” a 1975 turning point – you can pick 1971 or 1973 or 1977 or 1981 & the conclusion is the same.
Maybe this has to do with population trends (less babies)?
Maybe the correlation is between large state control in Western societies and growth. Instead of the huge growth promised by economic liberalisation, growth was poorer than the statist 1950s and 1960s. Or maybe it was an era of cheap and abundent oil…
I’m not advocating more state control by the way, just mentioning it.
I don’t think so, TCO. There certainly were less babies in all these countries in the later decades of the century than in the 1950s and 1960s, but in the short run this boosted GDP per capita. Babies don’t add to the numerator (GDP) but to the denominator (population). Taking account of demography STRENGTHENS my point.
Hi, Ian:
You write, “Mark, growth in per capita GDP in ALL of the major rich (computer-using) countries was lower in the fourth quarter of the twentieth century (1975-2000) than the third quarter of the century (1950-75). On your argument it should have been the other way around.”
Not really…for several reasons:
1) In 1945, Japan and Germany lay in smoldering ruins. Many of the “winners” of WWII didn’t fare much better (e.g. France, the Netherlands, Belgium). It’s not surprising that all those countries grew very fast in the 1950s, as they rebuilt their countries almost from zero. For example, what shape were Hiroshima, Nagasaki, Tokyo, Dresden, Berlin, etc. in, circa 1950?
2) As I previously noted, the number of “effective human brains” is actually still quite low. For example, I previously gave an illustrative example that assumed that a personal computer was currently 1000 times less capable than a human brain, but assumed that the capability was doubling every 5 years. But let’s take some real numbers, from Ray Kurzweil:
http://www.kurzweilai.net/articles/art0184.html?printable=1
a) According to Kurzweil, the human brain has 100 billion neurons, each with 1000 connections, and a switching speed of about 200 times per second. That makes for a total processing power of 20 quadrillion (20 x 10^15) operations per second.
b) According to Kurzweil, in 2000, $1000 bought computing power equal to approximately…let’s call it 200 million (2 x 10^8) operations per second. Therefore, in 2000, $1000 bought a computer that’s approximately one-ten-millionth (1/10,000,000) as powerful as a human brain.
c) But according to Kurzweil, the amount of computing power per $1000 is increasing by a factor of 1000 approximately every 12 years. So in the next 20-25 years, $1000 will buy a computer that can do the same number of calculations per second as a human brain.
d) So if we’re building 200 million of those $1000 computers, we’ll be adding 200 million human brains’ worth of computing power to the population every year, circa 2025-2030. (In reality, we’ll be building much more than 200 million computers per year, because as I pointed out, it will be waayyyy cheaper than having and teaching children. ;-))
e) Even more important, $1000 will buy a computer that can do 1000 TIMES as many calculations, only 12 years after that.
According to Kurzweil:
Does Her Majesty’s Treasury take that factor into account, in its analysis of economics in the 21st century? My guess is that it does not. Therefore, my prediction is that HM Treasury’s economic analyses will be spectacularly wrong (on the low side).
Once again, I don’t think anyone will have to wait until 2100, to see how HM Treasury’s (and the IPCC’s) analyses will be spectacularly low. It should be very clear in the next 5-20 years. (In fact, I can see it already. But that’s because I stand on the shoulders of giants. 🙂)
Re #39. Again, John A, I don’t think so. The decline in growth rates in rich countries has occurred in spite of a RISE in economic liberalism (decline in state control). My co-author, David Henderson, has written several books about the subject, documenting this in detail. Growth rates in rich countries were still high in the last quarter of the century by comparison with all of previous history except for the so-called “golden age” (usually defined as 1948-73). No countries that were already rich have maintained growth rates of 3%, let alone 4.5%, 7.5%, 10.5% or 12%.
Re #41. The rapid exponential growth of computing power has been proceeding for decades, and all of the models (including those used by the IPCC) assume that this will continue.
Mark, although you and others have questioned Willis’s argument in #28 and #31 on points of detail, I don’t believe that you’ve answered his key points, which in my view are correct.
I agree on the per capita point. I guess I just meant the babies help your overall GDP growth as a country as a world.
Heh, heh, heh! Good one. Let’s get back to basics:
1) Do these models all agree that (free) human minds are what creates wealth? (For example, suppose all our brains turned into chimp brains tomorrow…I presume the models would show economic growth coming to a rather abrupt halt? 😉)
2) If the models all agree that human minds create wealth, why don’t they show economic growth increasing rather spectacularly when (as Ray Kurzweil predicts will happen circa 2025) $1000 worth of computing power can crank out as many instructions per second as a human brain…or when (as Ray Kurzweil predicts will happen circa 2050) $1000 worth of computing power can crank out more instructions per second than all the biological brains on earth combined?
Well, first I need to get some exercise and some supper. I’ll try a bit later, though.
Best wishes,
Mark
Mark, On #27 you said “If you’ve read Julian Simon, you know that (free) human minds cause economic growth. The more free human minds you have, the higher your economic growth rate.” Yes, I have read Julian Simon and admire and agree with much of what he says. Can you give me a source for the statements that you attribute to him? I’d be surprised if he made such a crude statement, because he recognised that “There are many costs and benefits of there being more or fewer people other than the production of new knowledge” (“The Ultimate Resource 2”, p. 599).
Firstly, post #21. Thank you very much for your reference to the minutes of the House of Lords committee.
Secondly, I do agree with some of John A’s comments about the lack of concern shown by the IPCC about possible variations in the solar constant.
One thing that could have been done is to make a large number of observations of the brightness of stars in galactic clusters. Clusters such as M67 contain lots of stars which are similar to the Sun in composition, surface gravity and effective temperature. What variations in brightness and colour do these stars have over short and long periods?
Re #36, Mark, thanks for your reply. Settle in, this could take a few moments.
You say:
Dunno where you’ve been, but the idea that neurons are simply switches has been discredited for almost 50 years. The first evidence of this, curiously, was obtained from frog’s eyes. Eyes are made up of neurons arranged in layers. By the time the information has only gotten as far as the second layer back from the rods and cones, far from the brain, certain frog neurons are already only responsive to small, moving objects (flies=food), while other neurons are only responsive to large moving objects (=danger). It quickly became clear that neurons, far from being the simple switches that people had assumed, were actually processors capable of quite complex calculations. You might start with the classic first paper on the subject, “JY Lettvin, HR Maturana, WS McCulloch and WH Pitts, “What the Frog’s Eye Tells the Frog’s Brain,” Proc. IRE 47 (1959) 1940–1951″ … like I said, you’re starting from a ways back.
Your claim that a large number of simple switches arranged in parallel can somehow perform all of the activities associated with life is … mmm … well, let me just call it a result of a fundamental misunderstanding of the nature of life, computers, and switches. If your claim were true, we could simply arrange a huge number of computer memory units (which, unlike neurons or processors, are switches) in parallel and have a very complex, fast computer … not. I also note that your first post claimed that computers are 1/1000 as powerful as human brains, and your later post said 1/10,000,000 as powerful … which is your real claim?
Regarding my question about the last time a computer came up with a new idea, you say:
I guess I’ll have to assume that you are serious about this, although it hardly seems possible. Someone programs a computer to write “poetry”, and you seriously think that this is a computer coming up with a new idea??? Let me point out a distinction that seems to have eluded you. Ray Kurzweil came up with an idea about poetry. The computer did not. Ray Kurzweil instructed the computer to write poetry. The computer did not. Perhaps the distinction might be made clearer by considering which one stands a chance of understanding what the computer “wrote” — Ray, or the computer?
Please spare me the pain of your trying to instruct me about computers. I have been actively writing computer programs for over 40 years, from the sound of your postings perhaps longer than you have been alive, and I follow the field very closely, including MIT’s Technology Review. I have programmed a host of different computers to do a whole variety of interesting things, including writing a program twenty years ago to make a computer write poetry (using a Markov chain algorithm). I can assure you that, although the poetry was sometimes interesting, the computer was not creating new ideas as we understand the term.
Are you really serious about this, do you really believe that Ray or I programming a computer to write poetry (or develop mechanical designs, for that matter) means a computer can think? If so, you are definitely not up to speed in the field.
Regarding your accusation of my “cherry picking” the data, my point was simply this: in 1984, productivity gains were up. Since then, they dropped for about ten years, and then rose for ten years, and have not yet reached the 1984 level. This hardly constitutes confirmation of a theory that computerization leads to higher productivity. 1984 was when the Macintosh was introduced, and PC use was starting to become commonplace. Despite that, the productivity gains dropped for ten years, and have not exceeded historical levels. Explanation?
As I mentioned, this lack of productivity gains despite computerization is a recognized paradox in the industry, not the result of cherry picking. A lot of better brains than mine have been unable to find an answer to the question, which is why it is called a paradox. You can’t dismiss it by handwaving about cherries.
You say:
Wait a minute, dude, you referred me to that data. That was the data used by the article that you cited. If you don’t think that data is relevant, then the conclusions of the article no longer stand, as they are based on that data, and you should stop citing the article.
However, the worldwide data doesn’t differ much. Since 1970, gains in GDP peaked at about 7% in 1973, and have never regained that peak. Nor have they generally increased during the period of computerization. See http://www.theworldeconomy.org/publications/worldeconomy/histostats-table07-3.pdf for details.
In fact, the growth rate has been dropping steadily for the last three decades. In the 1970s, the GDP growth averaged 4.5% (using PPP). In the 80s, it dropped to 3.4%, and in the 90s, the average fell again, to 3.1%. (Using MER it’s worse, the drop is greater). The data shows clearly that the gains are decreasing, not just since the 80s as you asked about, but during every decade since the 70s. This is hardly a resounding endorsement of your hypothesis that computerization -> increased GDP.
Regarding fruit flies, the analogy is clear. Both fruit flies and your data assume unlimited growth over a long timespan. Your argument is that there will be an increasing growth rate in per-capita GDP for the next 100 years. In the real world, however, this kind of uninterrupted growth almost never happens. Reality intrudes, as it always does, with a variety of limits to the kind of exponential growth that both your claims and the fruit fly example depend on. Will I bet on what it will be like in 2100? Sorry, I’m not a betting man, but I doubt very much if we’ll see the gains you assume.
Finally, you seem to have difficulty distinguishing speed from quality. You say:
So in 2025 computers will be doing the same number of calculations per second as the human brain … so what? You seem to think that is important. It is not.
To see why, try redoing your same calculations for an elephant. An elephant has a much larger brain than a human today, so it must be doing more calculations per second, and yet somehow I don’t notice too many elephants writing novels and winning poetry contests … yes, they’re fast, and so is a really, really, really fast adding machine, but so what? We’re not looking for speed, we’re looking for things like imagination and creativity and the like.
w.
PS — As I mentioned before, computers can’t yet even mimic humans, much less actively create like humans. None has passed the Turing Test (Turing himself thought a computer would pass it by the year 2000, but despite modern computers being far more powerful than any envisioned in 1950, none has yet passed the test.)
My own criteria for computer artificial intelligence is more stringent, however. I call it, with commendable modesty, the “Willis Test”. To pass the Willis test, at some point a computer needs to spontaneously tell the operator “I can’t be bothered with calculating any more of your stupid trivial Excel spreadsheets today, I’ve got much more important things to think about …” At that point, I will agree that we have artificial intelligence. Until then, we just have fast adding machines.
Mark, a couple more things. First, thank you so much for the longbets url, what a great site. Like I said, I’m not a betting man, but great fun. Much appreciated.
I also noted that whether a computer will pass the Turing Test by 2029 was bet #1, how appropriate.
Second, I’ve looked further at Kurzweil, and found a very interesting quote (emphasis in original):
Life, at its core, is about recognizing patterns. Sunflowers turn to face the sun.
Clearly, Kurzweil states that progress in pattern recognition is linear with respect to with exponential gains in processor speed and memory capacity. What you are talking about, machine intelligence, is at least an order of magnitude more difficult than pattern recognition. Thus progress in machine intelligence must perforce be much less than linear wrt tecnological advance. The technology (speed and memory) will advance exponentially for quite a while longer, but the progress towards machine intelligence will continue to be less than linear.
But there is a further problem. Remember that what you are looking for is not really machine intelligence at all. Sounds crazy, but true. Suppose I build an intelligent computer, smart as your average 28 year old. Would you use that machine to advise you on your marriage?
We don’t want our advisors to be intelligent. We want them to be wise. Oh, intelligence is a plus, but without wisdom it doesn’t go far. We want wise advisors.
Now wisdom is a quality that’s scarce among humans, much scarcer than intelligence. I know I’m not over-endowed. Generally, it is in some sense a function of age and experience. We learn from our failures and our successes, and after a lifetime of it, some people like my grandmothers get wise. One was smart, and one wasn’t, but both were wise. Both were good advisors. Smart is helpful … but not necessary. We want wise advisors.
So we have a hierarchy of challenges and their respective rates of progress WRT technological advance:
Pattern Recognition (linear)
Machine Intelligence (less than linear)
Machine Wisdom (who knows, but very slow).
Yes, we will get there … but it will require a number of further paradigm shifts to do it, and progress is less than linear WRT speed and memory.
I haven’t yet touched on the final part of the story. There’s worse news. Suppose I built a wise, inventive, surprisingly witty computer, and it came up with a great idea. Sure fire. Will that idea make it to the marketplace? Will it even get published or noticed? In short, will that idea make any difference?
Only if it has a champion.
Ideas are cheap. Everyone has lots of good ideas, ideas that could be commerially or scientifically or socially or personally successful. You do, I do. Every one of those ideas withers and dies without a champion. We’re not short of ideas.
We’re short of champions, the people who take ideas out into the world and make them happen.
Which is relevant to this discussion because without a champion, even the best idea will have no effect on the GDP.
Ugly. This leads to another category of smart computers, that category that can organize the effort and see it through to the end, and another rate of progress for that category:
Machine Champion for a new idea, develops the plan, places the orders, makes the contacts, inspects the finished product, affects the GDP … (maybe before 2050 … maybe not this century.)
Because of all of that, I would say that your hypothesis, that machine intelligence will loom large and soon in an economic sense, affecting IPCC or Castles projections, is half right. It will certainly loom large … but soon?
To be fair, each step of the way will bring its own benefits. Long before we have wise advisors we’ll have efficient assistants. But they won’t have much effect. Yes, it’d be cool to say “Jeeves, call Alice and Bill to set up a meeting at 8”, and have my wristwatch do it, it would increase my efficiency. But at the end of the day, I’m the one that has to meet with Alice and Bill and make the decision to hire either one or the other. It is my action in going to the meeting and my wisdom or my stupidity in that hiring decision that will affect the GDP, no matter how smart my dang wristwatch is …
And here’s the beauty part … with a fast, efficient assistant, not only can I do things better, I can also magnify my stupidity and make really big mistakes! And at various times we all do just that, both individually and as a species …
So the final reason the GDP isn’t going to take off like a rocket is because … um … er … well, no easy way to say this, so I’ll just spit it out —
Humans do some really, really, really ugly and dumb and destructive and foolish and vicious stuff to each other, both personally and professionally and commercially, and getting more powerful doesn’t necessarily help that situation … what I’m saying is, we get in each other’s way and slow each other down a lot.
That, in turn, puts what I call a “stupidity based governor” on the rate of GDP growth … that’s where if it starts to grow too fast, don’t worry … some fool will drag it back down soon …
Thanks again for longbets, Mark, I’m heading back there now.
w.
“My own criteria for computer artificial intelligence is more stringent, however. I call it, with commendable modesty, the “Willis Test”. To pass the Willis test, at some point a computer needs to spontaneously tell the operator “I can’t be bothered with calculating any more of your stupid trivial Excel spreadsheets today, I’ve got much more important things to think about …” ”
Or, as has been writ before:
“I’m sorry, Dave, I’m afraid I can’t do that.”
Ian Castles writes,
I apologize if my statements made it seem like I was actually *quoting* Julian Simon. I was instead capturing the general flavor of his arguments…as can be determined from book titles like, "Ultimate Resource-2"…referring to (free) human minds.
Here are a few actual Julian Simon quotes that reflect the idea that more people means more economic growth:
http://www.acton.org/publicat/randl/interview.php?id=144
http://www.catholiceducation.org/articles/population/pc0015.html
But this whole issue is essentially irrelevant to the question at hand. The question at hand is not whether more PEOPLE result in a higher rate of economic growth. The question is whether an essentially infinite number of essentially infinitely powerful ***computers*** is a tremendous boon to economic growth rate.
If (only!) Julian Simon were alive, I’m sure that if I asked him *that* question, he’d answer in the affirmative! After all, if we have computers equivalent to an infinite number of human minds, and the human mind is "The Ultimate Resource," then we have an infinite number of "The Ultimate Resource"…without bearing the ***cost*** of feeding, housing, and teaching that resource for ~20 years, as the mind learns enough to make a significant contribution to society.
So my question back to you again, is, "What do the IPCC’s and the economics profession’s models say will happen to economic growth when $1000 worth of computing can do the same number of instructions per second as a human mind?"
And also, "What do the IPCC’s and the economics profession’s models say will happen to economic growth when $1000 worth of computing can do the same number of instructions per second as ALL BIOLOGICAL MINDS ON EARTH, COMBINED?"
I say that, if those economic models don’t show pretty darn spectacular economic growth, they probably aren’t correct. In fact, I’ve made explicit numerical predictions regarding economic growth in the 21st century, at “Long Bets #194.” I challenge anyone who claims to disagree with my assessment, to go to Long Bets #194, and register your vote. The voting process is ****COMPLETELY FREE,**** and doesn’t require more than a minute to register.
http://www.longbets.org/194
Ian Castles writes,
I apologize if my statement made it seem like I was actually *quoting* Julian Simon. I was instead capturing the general flavor of his arguments…as can be determined from book titles like, “Ultimate Resource-2” (referring to (free) human minds).
Here are a few actual Julian Simon quotes that reflect the idea that more people means more economic growth:
http://www.acton.org/publicat/randl/interview.php?id=144
http://www.catholiceducation.org/articles/population/pc0015.html
But this whole issue is essentially irrelevant to the question at hand. The question at hand is whether an essentially infinite number of essentially infinitely powerful *computers* is a tremendous boon to economic growth rate.
If (only!) Julian Simon were alive, I’m sure that if I asked him *that* question, he’d answer in the affirmative! After all, if we have computers equivalent to an infinite number of human minds, and the human mind is “The Ultimate Resource,”…then we have an infinite number of “The Ultimate Resource,” without bearing the *cost* of feeding, housing, and teaching that resource for ~20 years, as the mind learns enough to make a significant contribution to society.
So my question again, is, “What do the IPCC’s and the economics profession’s models say will happen to economic growth when $1000 worth of computing can do the same number of instructions per second as a human mind?” And also, “What do the IPCC’s and the economics profession’s models say will happen to economic growth when $1000 worth of computing can do the same number of instructions per second as ALL BIOLOGICAL MINDS ON EARTH, COMBINED?”
I say that, if those economic models don’t show pretty darn spectacular economic growth, they probably aren’t correct. In fact, I’ve made explicit numerical predictions regarding economic growth in the 21st century, at “Long Bets #194.”
If anyone disagrees with my assessment, I think you should go to Long Bets #194, and register your vote against my prediction. The voting process is *COMPLETELY FREE,* and doesn’t require more than a minute to register.
In #43, Ian Castles wrote,
I think I’ve addressed each of his points in #28…or is there something in particular from #28 you think I’ve missed?
Regarding #31:
1) Willis wrote, “I’ve been thinking a bit more about the gains in productivity in the late 90s and early 00s. I suspect that they had much more to do with improvements in communications than in computers. In particular, the advent of both email and the web have greatly improved my own productivity, and I suggest that of others as well.”
Well, yes. Ray Kurzweil definitely does NOT say that the ONLY reason why economic growth is going to be spectacular is due to exponential increasing in microprocessor speed. He clearly emphasizes that hard drive capacities are also increasing exponentially, and that data transmission speeds are increasing exponentially.
For example, I think it was an MIT Technology Review cover article that essentially asked the question, “Is Google God?” (To the tremendous offense of some readers.) But Google would NOT be God, unless they had their UNBELIEVABLE server storage capacities. Not to mention their excellent software that is able to produce search results that are very relevant to the questions at hand.
Likewise, at work I have a super-high-speed connection (T1, I think). But until 6 months ago, I had only a 56k connection here at home. The difference in time for downloading a 2-3 megabyte Adobe Acrobat file was tremendous. It would literally take 20 minutes for me to download something at home that could be downloaded in less than 20 seconds at work.
So you’re tremendously misunderstanding Ray Kurzweil’s and my point, if you think it’s *only* microprocessor speed that will cause the tremendous acceleration in economic growth: it’s microprocessors, solid state memory, hard drives, communications speeds…and probably a dozen more trends.
To pick another example: a complete human genome contains 3 gigabytes of data (as I recall). Well, only 10 years ago, a 3 gigabyte hard drive was HUGE…a multi-thousand dollar investment. Now, I can by a drive 100 times that size for under $200. So I can not only store my own complete genome, but all my relatives (if I wanted to). And in 30-40 years, I’ll (hoping to still be alive then) probably be able to store all the genomes of everyone in the U.S. on a $1000 hard drive.
2) Willis continued, “Faster computers may speed this up eventually, using voice input. I work in a small office with 6 desks in it, though, and a lot of talking and through traffic, so I suspect I won’t be able to use it. But even if I could, I type at about 70 wpm, and speech is at about 250-300 wpm, so it will be a max 4x speedup even if the dictation programme is letter perfect. And I still need to think up the letter, and proofread it and format it and correct it when it is done, so the increase will be even less.”
This is a great example of how Willis is not thinking into the future very far at all. First of all, a human can understand when someone whispers words to them…it’s not so way-out to expect that a microphone and voice recognition system will be able to do the same thing. Then there’s active noise cancellation techniques; they’ll be commonplace in 10-20 years. There’s really cool directed-sound technology that can put sound to a single *point* in a room. (And it won’t be long before that directed-sound can follow a person around a room.)
And this doesn’t even get into potential brain-scanning techniques…such that a person doesn’t have to talk at all.
And proof-reading…goodness! Even bleeding Windows has automatic spell correction. Eventually, even Windows will be able to make human proof-reading essentially obsolete.
But this mundane stuff completely misses the beauty of computers. I’m an environmental engineer who does environmental research. I’m currently working on an international assessment of black carbon emissions. There’s a research in India that we hope to talk to by phone next week. (Rather than communicate through email, which is slow for large amounts of discussion.)
Even *1 year* from now, my company will have Internet telephony, such that the call will be essentially free. And we probably could ALREADY have Internet audio/video conferencing with the guy in India. Ten-to-twenty years from now, it will be like we’re right there in the room with him (with a life-size LCD screen…probably even in 3-D).
We also would like to talk to researchers and government people in China. Unfortunately, we don’t speek Chinese. Again, 10-20 years from now, translation-on-the-fly will be probably be routine.
Finally, Willis wrote, “The ugly reality is that many of the things that I do, from writing letters to sending emails to creating spreadsheets to pounding nails, will not benefit in the slightest from faster computers.”
Let’s talk about “pounding nails.” In 30-40 years, one will be able to buy a foam or glass-fiber-reinforced house that will be built entirely by a machine. The windows will be filled with transparent aerogel, so they’ll look like windows do today, but have R-19 insulating value. Heck, the roof itself might be electrochromic aerogel, so that it lets in light during the day, but is opaque at night.
The things will be incredibly cheap, but rock solid…able to handle a Category 4 hurricane without a problem. Except that there WON’T be any Category 4 hurricanes…because I’m working right now on developing hurricane reduction ideas.
Again, I urge both of you to read Ray Kurzweil. I just ordered, “The Coming Singularity.” But I can already recommend, “The Age of Spiritual Machines.”
I used to be a member of the World Future Society. But I dropped out, because most of them are complete amateurs. Ray Kurzweil is a PRO. If anyone is trying to think 50-100 years ahead, and has NOT read Ray Kurzweil, they’re amateurs, practicing futurism without a clue.
In #43, Ian Castles wrote, “Mark, although you and others have questioned Willis’s argument in #28 and #31 on points of detail, I don’t believe that you’ve answered his key points, which in my view are correct.”
I think I’ve addressed each of his points in #28…or is there something in particular from #28 you think I’ve missed?
Regarding #31:
1) Willis wrote, “I’ve been thinking a bit more about the gains in productivity in the late 90s and early 00s. I suspect that they had much more to do with improvements in communications than in computers. In particular, the advent of both email and the web have greatly improved my own productivity, and I suggest that of others as well.”
Well, yes. Ray Kurzweil definitely does NOT say that the ONLY reason why economic growth is going to be spectacular is due to exponential increasing in microprocessor speed. He clearly emphasizes that hard drive capacities are also increasing exponentially, and that data transmission speeds are increasing exponentially.
For example, I think it was an MIT Technology Review cover article that essentially asked the question, “Is Google God?” (To the tremendous offense of some readers.) But Google would NOT be God, unless they had their UNBELIEVABLE server storage capacities. Not to mention their excellent software that is able to produce search results that are very relevant to the questions at hand.
Likewise, at work I have a super-high-speed connection (T1, I think). But until 6 months ago, I had only a 56k connection here at home. The difference in time for downloading a 2-3 megabyte Adobe Acrobat file was tremendous. It would literally take 20 minutes for me to download something at home that could be downloaded in less than 20 seconds at work.
So you’re tremendously misunderstanding Ray Kurzweil’s and my point, if you think it’s *only* microprocessor speed that will cause the tremendous acceleration in economic growth: it’s microprocessors, solid state memory, hard drives, communications speeds…and probably a dozen more trends.
To pick another example: a complete human genome contains 3 gigabytes of data (as I recall). Well, only 10 years ago, a 3 gigabyte hard drive was HUGE…a multi-thousand dollar investment. Now, I can by a drive 100 times that size for under $200. So I can not only store my own complete genome, but all my relatives (if I wanted to). And in 30-40 years, I’ll (hoping to still be alive then) probably be able to store all the genomes of everyone in the U.S. on a $1000 hard drive.
2) Willis continued, “Faster computers may speed this up eventually, using voice input. I work in a small office with 6 desks in it, though, and a lot of talking and through traffic, so I suspect I won’t be able to use it. But even if I could, I type at about 70 wpm, and speech is at about 250-300 wpm, so it will be a max 4x speedup even if the dictation programme is letter perfect. And I still need to think up the letter, and proofread it and format it and correct it when it is done, so the increase will be even less.”
This is a great example of how Willis is not thinking into the future very far at all. First of all, a human can understand when someone whispers words to them…it’s not so way-out to expect that a microphone and voice recognition system will be able to do the same thing. Then there’s active noise cancellation techniques; they’ll be commonplace in 10-20 years. There’s really cool directed-sound technology that can put sound to a single *point* in a room. (And it won’t be long before that directed-sound can follow a person around a room.)
And this doesn’t even get into potential brain-scanning techniques…such that a person doesn’t have to talk at all.
And proof-reading…goodness! Even bleeding Windows has automatic spell correction. Eventually, even Windows will be able to make human proof-reading essentially obsolete.
But this mundane stuff completely misses the beauty of computers. I’m an environmental engineer who does environmental research. I’m currently working on an international assessment of black carbon emissions. There’s a research in India that we hope to talk to by phone next week. (Rather than communicate through email, which is slow for large amounts of discussion.)
Even *1 year* from now, my company will have Internet telephony, such that the call will be essentially free. And we probably could ALREADY have Internet audio/video conferencing with the guy in India. Ten-to-twenty years from now, it will be like we’re right there in the room with him (with a life-size LCD screen…probably even in 3-D).
We also would like to talk to researchers and government people in China. Unfortunately, we don’t speek Chinese. Again, 10-20 years from now, translation-on-the-fly will be probably be routine.
Finally, Willis wrote, “The ugly reality is that many of the things that I do, from writing letters to sending emails to creating spreadsheets to pounding nails, will not benefit in the slightest from faster computers.”
Let’s talk about “pounding nails.” In 30-40 years, one will be able to buy a foam or glass-fiber-reinforced house that will be built entirely by a machine. The windows will be filled with transparent aerogel, so they’ll look like windows do today, but have R-19 insulating value. Heck, the roof itself might be electrochromic aerogel, so that it lets in light during the day, but is opaque at night.
The things will be incredibly cheap, but rock solid…able to handle a Category 4 hurricane without a problem. Except that there WON’T be any Category 4 hurricanes…because I’m working right now on developing hurricane reduction ideas.
Again, I urge both of you to read Ray Kurzweil. I just ordered, “The Coming Singularity.” But I can already recommend, “The Age of Spiritual Machines.”
I used to be a member of the World Future Society. But I dropped out, because most of them are complete amateurs. Ray Kurzweil is a PRO. If anyone is trying to think 50-100 years ahead, and has NOT read Ray Kurzweil, they’re amateurs, practicing futurism without a clue.
Re #51, Mark, thanks for your comments. I have addressed them directly in my latest two posts, and await your reply.
In particular, the issue of what will happen when "$1000 worth of computing can do the same number of instructions per second as ALL BIOLOGICAL MINDS ON EARTH, COMBINED" is of interest. As I pointed out, the paradox is that the recent and continuing computer explosion has not led to increased GDP.
More to the point, is your claim that GDP is somehow connected to the total number of instruction per second executed by worldwide computers. Here’s a though experiment:
Suppose you invented a magic beneficial computer virus that would run on any computer, and its only effect was to instantly make that computer work one hundred times faster. You release the virus, and POOF … within a week, virus goes wild, you’ve tripled the speed of every computer in the world, you get awards and go on talk shows …
How much would that tripling of computer speed affect the GDP? Would it make a noticeable difference? A large difference? Any difference?
Starting close to home, we have five computers in our office, and I can assure you of one thing. The operators (including myself) are generally much slower than the computers. Increasing the computer speed one hundred fold would improve productivity by less than 1%. It doesn’t matter if Excel takes 0.001 milliseconds to do my spreadsheet instead of 0.003 seconds, my production doesn’t increase. My letters won’t get typed any faster, my email won’t change. A faster colour printer would improve our office productivity much more than five faster computers would.
(This is because printers move molecules, while computers only move electrons. Electrons are easy to move, while molecules (of paper and plastic and ink in this case) are much harder to move. It is far easier for a computer to speed up 100 fold, for example, than for a printer to speed up 100 fold. The parts fly off the printer if you try that, whereas it’s hard to make parts fall off an electron. This is the core reason why increases in COMPUTATIONAL SPEED (electrons) will not be matched by increases in PRODUCTION (molecules). My computer is now running one hundred times faster, but my print speed in unchanged … but I digress.)
Looking outward, the tasks that are important to most businesses do not require ultrafast or ultralarge computers. Yes, Boeing Aircraft has to have a supercomputer to design the next generation of jets, but the sale of supercomputers to business is small. These days, our computers are like our brains, in that (as urban legend has it) we only use 10% of their capabilities … how much would increasing your computer’s speed improve your productivity? My sense is that our office is typical, that for most businesses it would be much less than 1%. For the overwhelming majority of the tasks for which computers are used worldwide, increasing computer speed will not significantly increase worker productivity.
You are right that further productivity increases will come, and that they will come from the new emergent properties of machine intelligence and machine wisdom, which we both agree will occur. Until then, however, speed increases won’t help much.
w.
Mark — There is an important word in the Hume via Simon quote that seems to have been missed in this discussion. The quote was:
“We now know statistically that what David Hume wrote on the subject in the 1700s was exactly right. When identifying why Holland was the richest country in Europe, Hume said that “Liberty, necessity, and a multitude of people” were the causes.”
The important but under-appreciated word is “necessity”. It is — at least partly — the absence of “necessity” that lies behind the slowing of economic growth in affluent countries.
There are only so many hours in the day, and the higher one’s level of material well-being, the smaller is the incentive to allocate marginal hours to economic production. Once one is enjoying a prosperous retirement income, such as that available to distinguished former national statisticians, one may be more tempted to try to inhibit the flow of junk production, such as IPCC scenarios.
The serious point is that future humans that are already five or ten times as wealthy as we are now may lack the incentives to sacrifice themselves to the extent necessary (even then) to produce 10% real annual growth.
No, Willis, you haven’t addressed my comments. I have made predictions of for the world per-capita GDP in Long Bets #194:
http://www.longbets.org/194
You have made comments like:
In my comment #51, I suggested that if you disagree with my predictions for economic growth in the 21st century, you should go to Long Bets #194, and vote that you disagree. The vote is completely FREE, and only takes a minute to register.
Long Bets was set up specifically as a long-term record for people to take responsibility for their predictions. If you really disagree with my prediction, you should put it on the Long Bets record. (And if by chance you should ever later decide you were wrong, you can add that in the discussion section.)
Re 57, hi, Mark. Been there, voted that I disagreed with the prediction.
Onwards …
w.
Great. Well, not great…because you’re wrong. But I respect and appreciate your willingness to actually take a position.
(So far, not one IPCC author, or author at the Real Climate website, has even had the integrity to vote against my predictions #180 or #181, regarding climate change.)
I’ll explain why you’re wrong as time allows. But today’s a wrap for me. 🙂
Best wishes,
Mark
Re #51. Mar, I disagree with your assessment and have voted accordingly.
Sorry, Mark (not Mar).
Hi Ian,
I appreciate very much your willingness, like Willis Eschenbach, to take responsibility for your opinion…to go “on the record” at a site that is deliberately set up as a long-term record of peoples’ predictions and opinions.
But if anything, my discussions with you and Willis to this point have convinced me even more of the correctness of my opinion. (In fact, I think my predictions for the latter part of the century are almost certainly much too low.) Also, I just received Ray Kurzweil’s 600+ page “The Singularity is Near,” and even my brief scanning to this point reveals more data to support my position.
As I’ve previously commented, I think by 2020 (or 2030 at the latest), it will be abundantly clear that your position is wrong. But by that time, “billions and billions” (guess I shouldn’t put that in quotes! :-)) will have been wasted on global warming worries.
If Ray Kurzweil, Arnold Kling, Robin Hanson and I are right–and I’m very sure we are–worrying about people in 2050 or 2100 is simply ridiculous, because they will be wealthy almost beyond imagination. So if we are right, that fact alone renders any debate about how global warming will cause future problems completely unnecessary.
In my opinion, the fact that the IPCC Third Assessment Report (deliberately) produced ridiculously high “projections” of methane atmospheric concentrations, CO2 emissions, CO2 atmospheric concentrations, and resultant temperature increases, is very important. (In fact, I think it constitutes the greatest fraud in the history of environmental science.)
But if Ray Kurzweil, Arnold Kling, Robin Hanson and I are right, even that fraud is essentially unimportant, because I think it’s obvious even to lay people that no one should be worried about people living in a world where the world average income is multiple millions of dollars. Even lay people understand the fundamental immorality of the poor (i.e., people of today) sacrificing for the benefit of the extravagantly wealthy (which the people of 2050 and 2100 will be).
So I will continue, as my free time allows, to try to persuade you why you’re wrong. (And if the opposite occurs, well, it wouldn’t be the first time.) (Maybe the second. ;-))
Best wishes,
Mar(k) 😉
Mark,
What’s the track record of those making similar predictions? Going back 20 years? 30 years? 40 years? 50 years? I don’t think its very good. But, your site will be an interesting test. Will it even exist long enough to know who was right.
As the EPA says about gas mileage, “Your mileage may vary.” I don’t accept that one can look at the “track record of those making ‘similar’ predictions,” to determine how my accurate *my* economic predictions will be.
The more I evaluate the subject, the more confident I am that my predictions will actually be substantially *lower* than actual economic performance. I think this will be particularly true in the last half of the century.
According to Ray Kurzweil (“Are We Spiritual Machines? Ray Kurzweil vs the Critics of Strong A.I.”) these are some relevant dates for computing power:
1) One human brain capability (2 x 10^16 cps) for $1000 around the year 2023.
2) One human brain capability for one cent around the year 2037.
3) Entire biological human race capability (2 x 10^26 cps) for $1000 around the year 2049.
4) Entire biological human race capability for one cent around the year 2059.
If these years are 100% accurate, I’d be shocked if the world per-capita GDP growth isn’t averaging above 7% per year even as early as the mid-2020s.
Let’s think about two paradigm-shifting technologies:
1) Space elevators (which have the potential to lower costs to reach space by more than two orders of magnitude), and
2) Hydrogen-boron fusion (which has the potential to supply essentially infinite power at essentially infinite power densities, with essentially zero pollution).
In today’s world of 6.5 billion people, how many people do you think are working full-time on each of these technologies? My guess is that less than 2,000 people, on the entire planet, are working full-time on these technologies.
http://www.futurepundit.com/archives/001660.html
http://focusfusion.org/what/plasmafocus.html
Well, come 2025, it should be fairly easy to have the equivalent of 2,000,000 human brains working on those technologies. That’s **1000** times as many people as are probably currently working on them. And come 2050, it should be fairly easy to have the equivalent of 2 TRILLION human brains working on those technologies. So if Ray Kurzweil is right, it would be surprising if BOTH those technologies aren’t available by 2050.
In fact, with hydrogen-boron fusion, even such amazing thoughts as terraforming Mars become clearly possible. With hydrogen-boron fusion rockets, diverting comets to fall on Mars (to supply water) would be possible, and even installing an artificial sun in orbit around Mars would not be out of the question.
Or let’s take a more mundane technology: batteries to power automobiles. At present, major problems for automobile batteries are: 1) lack of sufficient range, 2) charging time is too long, and 3) relatively expensive.
Probably more human brains are working on the problem of batteries to power automobiles. Maybe in 2005, there are 500,000 human brains working on that problem, worldwide? By 2050, TRILLIONS of human brain equivalents could be working on the problem. (A couple notes: The present range of batteries is typically less than 50 miles, but even that would be sufficient for many people. And as I recall, MIT’s Technology Review just had a piece about lithium-ion batteries that can be charged to 80% within 3 minutes. So batteries aren’t so completely-out-of-the-question as one might think. If the range could be extended to maybe 150 miles, with a charging time of 3 minutes or less, both internal combustion engines and fuel cells would be almost completely unnecessary. And fuel cells could run on natural gas in the home, charging the batteries, and also providing power and heat for the home.) (In fact, with hydrogen-boron fusion, even neighborhood fusion generators would probably be possible…so very large centralized power plants would probably become obsolete, as would the nationwide grid, with its high tension towers.)
Mark…
Email me links on H-B fusion?
Thanks.
Hi Paul,
I don’t see your email address, so here are some links:
http://focusfusion.org/what/plasmafocus.html
http://www.rmi.org/sitepages/pid9.php?pageId=9&Thread=00001834&postId=1834&groupId=4
http://fusion.ps.uci.edu/beam/introb.html
P.S. This is relevant to the question at hand. For example, suppose it takes $100 billion to bring this technology to commercialization. (Just to pull a figure more or less out of the air.)
Well, suppose by 2050 there are some people with 10 times Bill Gates’ net worth…say $500 billion. Well, that one person ALONE could fund all the research and development to bring hydrogen-boron fusion to commercialization.
So it’s quite possible this type of fusion will be developed before 2100. That solves all global warming and energy problems right there.
re #64 Mark,
I think the trouble with Space Elevators is that they don’t seem to be possible on Earth as pure elevators. Given the strength of the electromagenetic field, an elevator reaching all the way to the earth’s surface would be too massive to be practical no matter what material you made it of. A rotating elevator which only reached to the top of he atmosphere would be possible but not quite as simple as a pure elevator. A fountain powered elevator would work, but I’d hate to think what would happen if the fountain was ever shut off. I don’t think I’d allow such a thing in my back yard, backyard being the planet I live on.
Hi Dave,
I don’t want to get too far afield (probably that horse has already left the barn), but here’s an excerpt from the Los Alamos National Lab:
http://www.lanl.gov/news/releases/archive/03-124.shtml
Regardless of the merit (or lack of merit) of this particular concept, the point I was making is of the amazing economic potential of such massive computer power as we will have by the middle third of this century. It will be possible during that time to put a millions, and later BILLIONS and TRILLIONS of human mind equivalents to work on any particular problem.
This can be compared to current conditions, where computers are so costly and primitve that one computer even approaching a human mind’s calculating capacity costs millions of dollars. So right now, we are essentially limited, in “far-out” projects like hydrogen-boron fusion or space elevators, to at most a few thousand human minds working on those problems.
A good analogy might be the Manhattan and Apollo projects. Given a whole lot of money, and a great number of great minds, they accomplished amazing things in a short amount of time.
But if one has a $1000 computer with the calculating capacity of 10 BILLION human minds–as Ray Kurzweil projects will be available circa 2050–you have calculating power that dwarfs the Manhattan and Apollo projects available on essentially EVERY problem in the world.
Want cheap and efficient photovoltaic cells? Put 10 billion minds to work (for $1000).
Want better automobile batteries? Put 10 billion minds to work (for $1000).
Want cheap fuel cells that run on natural gas? Put 10 billion minds to work (for $1000).
Want to figure out a way to reduce the strength of all hurricanes such that the most powerful storm that ever comes on land is a tropical storm? Put 10 billion minds to work (for $1000).
Such problems should be solvable almost instantaneously, because the computing power could be used to do detailed analyses of virtually every aspect of the problems/solutions.
It seems almost inconceivable that such a situation could NOT lead to phenomenal economic growth. Even a situation where a $1000 can buy the computational power of ONE human mind ought to lead to phenomenal economic growth. And Ray Kurzweil predicts that situation will occur circa 2023.
We should be talking about completely autonomous cars driving at 80 mph bumper-to-bumper, groceries and retail goods delivered door-to-door by autonomous trucks, every house continuously spotless…
Mark, you seem to conflate large computers with intelligent computers … while computers in my own lifetime have gone from the first computer I programmed in 1963, with 4k of the radically new magnetic ring memory and a speed of 2 khz, to my current macintosh running at half a gigahertz with a half a gig of memory.
Now this is an increase of hundreds of thousands of times the computer power. But has my productivity increased by a factor or hundreds of thousands? Not.
To get the results you speak of will require, not just an increase in power and speed, but in intelligence. How much more intelligent is my computer today than my computer in 1963?
Not a bit. Not a jot. Not a tittle. Not a millismart, not a microalbert. (one microalbert is one millionth of the intelligence of Albert Einstein.
Yes, when computers get intelligent we will be able to, as you say, “put millions of minds to work” on a problem. But computers are not intelligent, and show no signs of becoming so. As I pointed out, none has even passed the Turing Test, and to do that, a computer doesn’t have to actually be intelligent, it only has to mimic intelligence.
I agree with you that when the day comes that we can “put millions of minds to work” on a problem, we will solve problems more quickly. Your naive faith that that day is just around the corner is touching, but does not accord well with the history of computers.
Historically, computers have gotten larger very quickly. And computers have gotten faster very quickly.
But they have gotten smarter very, very, very slowly, if at all. My Macintosh is not any smarter than the 1963 Algol I started on. Not even the slightest bit smarter. It’s a whole lot faster and larger, but it’s still dumb as a post, just like the first one.
Now you claim this will all change, and computers will suddenly become intelligent. You say you know this because they are still getting faster and bigger.
But the last fifty years have already proven that faster and bigger computers does not mean smarter computers … so since you are claiming that this will suddenly change, you need to show some proof, or at least some theory, that explains how we’ll get from no useful progress at all on intelligent computers for the last 50 years, to having an independent machine intelligence, in only a decade or two.
w.
Willis,
In some ways, computers have gotten smarter, largely in service of getting faster, and/or more accurate.
They are much smarter about how to send, and receive, signals over a piece of copper wire called a telephone line. They are smarter about detecting, and often correcting, internal errors.
The addition of cache memory make a computer smarter about getting pertinent data to the cpu faster.
Such smarts can be important to the functioning of computers, but they would seem relatively trivial as compared to the kinds of intelligence you may have in mind in your discussion.
Correction: a pair of copper wires called a telephone line.
While I’m at it, let me add that, for example, cuurent disk drives are immensely smarter, than old disk drives, about squeezing more date within a given amount of disk surface.
Generally, recent computer technologies have become smarter about distinguishing signals from noise, and using smaller, and faster, signals than older computer technologies.
The Turing Test does NOT require a computer to mimic “intelligence.” Passing the Turing Test requires a computer to mimic HUMANITY. There is a huge difference.
The Turing Test allows the questioner to ask a large number of questions that have absolutely nothing to do with intelligence, but have everything to do with human experience. For example, the Turing Test questioner may ask the computer about the computer’s experiences in kindergarten or high school. But the computer has no experience in kindergarten or high school. That has nothing to do with intelligence, it has to do with humanity.
In fact, the Turing Test was apparently inspired by Turing’s familiarity with a game where males and females pretend to be of the others’ sex, and questioners (receiving typed responses in another room) attempt to guess whether the respondent is truly of the sex being represented:
http://en.wikipedia.org/wiki/Turing_test
By the way…we have IQ tests that are used to measure human intelligence. If a computer took such a test, and scored a 140+, would you accept that it was “intelligent?” If not, why not?
Heh, heh, heh! My “naàÆà⮶e faith,” eh? Your condescension aside, let’s get the facts straight: My reference to putting “millions of minds to work” referred to the 2025 to 2050 time frame. Your characterization of the period from 20 to 45 years from now as “just around the corner” “does not accord well with” most humans’ conception of time. If you have to so blatantly misrepresent my position, is it perhaps because you can’t otherwise refute it?
Further, if anyone has made any statement of “naàÆà⮶e faith,” it is YOUR statement of October 25: “So in 2025 computers will be doing the same number of calculations per second as the human brain … so what? You seem to think that is important. It is not.”
Considering that not even ONE multi-million-dollar supercomputer can currently perform 20 quadrillion calculations per second, your assertion that it is not “important” if $1000 could buy an equivalent computer in 2025 sure seems like a statement of “naive faith” to me.
I have quoted Ray Kurzweil’s predictions regarding the cost and number of calculations per second for computers:
a) One human brain capability (2 x 10^16 cps) for $1000 around the year 2023.
b) One human brain capability for one cent around the year 2037.
c) Entire biological human race capability (2 x 10^26 cps) for $1000 around the year 2049.
d) Entire biological human race capability for one cent around the year 2059.
Do you (and Ian Castles) agree with those calculation speed/cost predictions, or do you disagree? And if you disagree, by what amount do you disagree (e.g. a factor of 1,000 slower speed, or a factor of 1,000,000, or what)?
Per my Long Bets #194, some approximate average annual per-capita GDP increases for the 2023, 2037, 2049, and 2059 time frames would be:
a) 2023: 3.5%,
b) 2037: 5.5%
c) 2049: 7.5%
d) 2059: 9.5%
Obviously, you and Ian Castles don’t agree with those predictions. But neither of you has answered my question: If one could indeed purchase a computer that can do the same number of calculations per second as the entire biological human race for $1000 in 2049, wouldn’t you both expect that to dramatically improve economic growth? If not, why not?
Oh, brother. This is the logical equivalent of Swiss cheese. Here are just SOME of the holes in this “logic”:
1) Writing novels and winning poetry contests isn’t even being discussed. What’s being discussed is world GDP growth rate.
2) Elephants don’t contribute much to GDP growth rate because (among many reasons):
a) There are very few of them around (only small numbers in Africa and India),
b) The very few elephants in Africa and India don’t contribute much to world GDP growth rate for the same reason that the very many humans in Africa and India don’t contribute much to world GDP growth rate…the various governments’ policies are so screwed up that the elephants and humans spend all their time just trying to survive.
3) The fact that a brain is large does NOT necessarily mean it is conducting more calculations per second…there could be larger neurons, or fewer connections, or slower connections.
4) Many of the calculations of an elephant’s brain are probably dedicated to using the muscles to be able to control its amazing nose, or being able to very sensitively sense the earth’s tremors through its feet, or sensing sounds with its large ears, or any other of its bodily functions that are much improved over humans.
First off, 100 times is not “tripling”…100 times is 100 times.
But to answer your question, simply go to my predictions for GDP growth in the 21st century. From 2000 to 2020, I predict 3% per year. For 2020 to 2040, I predict 4.5 percent per year. The midpoint of those values is the year 2010 and 2030. Per Ray Kurzweil, the improvement in calculations per second per $1000 from 2010 to 2030 will be approximately a factor of 1 QUADRILLION.
So obviously I don’t think an overnight increase by a factor of 100 will have any impact on the economic growth rate.
RE:69
It is a rather poor analogy as the underlying theory for both was established prior to attempting those projects. IOW, they already had a pretty good idea of how to get from point A (theory) to point B (engineering realization). This is not the case for computer technology mimicking human intelligence. There is no underlying theory that explains how intelligence works and therefore no point A as a starting point.
RE:27
The doubling, referred to a Moors Law, is clearly coming to an end. The doubling in speed is proportional to our ability to shrink the feature size. What your assuming is that computers will be able to increase speed by 1000 from a gigahertz to a terahertz. Computers based on silicon will never reach clock speeds in the terahertz region (10^12). This would require feature sizes on the order of one atom, which is clearly impossible due to numerous other physical restraints. We might get a couple of orders of magnitude increase, but that’s it, were done. To get any faster were going to need a new technology that has not been invented yet. This brings us right back to going from point A to point B without point A.
Oops. In #73, I was trying to work too fast, and made an error. I said that Ray Kurzweil predicted a factor of 1 QUADRILLION increase in computer calculations per second per $1000 from 2010 to 2030. But Ray Kurzweil actually predicts “only” a factor of 1 MILLION increase in computer calculations per second per $1000 from 2010 to 2030.
But the answer to the question Willis asked about a hypothetical worldwide factor of 100 increase in computer speed is still the the same. I think such an increase would have no discernable effect on the annual per-capita GDP increase. This can be determined from the fact that I predict that a factor of 1 million increase from 2010 to 2030 coincides*** with “only” an increase from 3% per year to 4.5% per year.
***NOTE: The word “coincides” is much more appropriate than “causes.”. As I’ve noted before, the increases in per-capita GDP growth rate that I project for the 21st century include increases expected from ALL technology gains, e.g., increased computer speed, increased hard drive (or holographic) memory, increased wired/wireless data transfer rates, etc. etc.
That was NOT the analogy I was making. The analogy I was making was between the Manhattan and Apollo projects versus space elevators and hydrogen-boron fusion.
The Manhattan and Apollo projects had a large number of human brains dedicated to bringing them to fruition. At present, space elevators and hydrogen-boron fusion have relatively few human brain equivalents dedicated to them. But circa 2049, if a computer with the calculating capacity of all the biological human brains on earth can be purchased for $1000, then essentially trillions of human brain equivalents can be brought to bear on the problems related to development of space elevators and hydrogen-boron fusion. (That’s in the somewhat-unlikely event that space elevators and hydrogen-boron fusion aren’t *already* commercially available in 2049.)
As Ray Kurzweil notes, the specifics of Moore’s Law is not a particularly good way of tracking progress in computers. A better way is the number of calculations per second that can be performed by $1000 worth of computer.
As Ray Kurzweil points out, the number of calculations per second that can be performed by $1000 worth of computer is increasing not merely exponentially (which would be a doubling every X number of years) but “double exponentially” (i.e., the doubling period is actually getting shorter and shorter).
If you go to the curve titled, “Exponential Growth of Computing” (under the heading, “The Exponential Growth of Computation Revisited,” you’ll see that Kurzweil predicts not just exponential growth, which would be a straight line on the exponential graph, but “double exponential” growth, in which the doubling time shrinks (rather than remaining the same, as would occur with “only” exponential growth).
Ray Kurzweil on the “Law of Accelerating Returns.”
Willis and Ian have asked why the rate of world per-capita GDP increase hasn’t already gone up, due to computer usage. I think it may already have, but I’ll have to address that later (if I have time).
One thing I’ve never done myself is to take the data I retrieved from the Internet on the number of personal computers built each year, and combine it with Ray Kurzweil’s data and predictions for the number of calculations per second per $1000.
I did that during lunch today, and posted the results on my own blog. Specifically, I calculated an “Annual addition of computer power, in Human Brain Equivalents (HBEs),” using the data on production I got from the Internet, Kurzweil’s data and predictions for calculations per second per $1000, and Kurzweil’s estimate of the calculating speed of a human brain (2 x 10^16 calculations per second):
Annual addition of computer power…or why economic growth will be spectacular
The results are in Figure 1 on my blog post, but can be summarized as follows:
1) Only ONE human brain equivalent (HBE) added in 1995,
2) One million HBEs added in 2015,
3) One billion HBEs added in 2025,
4) One TRILLION HBEs added in 2033, and…
5) One SEPTILLION (i.e. 1 x 10^24, or 1,000,000,000,000,000,000,000,000) HBEs added in 2057.
One TRILLION human brain equivalents added as early as 2033! It’s pretty ridiculous for anyone to try to figure out what will happen when that amount of calculating power is added every year. But it should be pretty easy to hit 5% per year per-capita GDP growth. After all, we’ve been able to get above 3% per year the last couple of years…and we’re currently only adding about 1000 human brain equivalents per year.
But will it run on Linux?