About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« Harvard to the Rescue | Main | Cyclodextrin's Day in the Sun »

October 17, 2011

The Singularity, Postponed

Email This Entry

Posted by Derek

I've had some problems over the years with the Singularity-Is-Near line of thought, and some problems with the "If we can build a new generations of microchips in five years, we ought to be able to cure cancer in ten" idea. Here's an article by Paul Allen in Technology Review that takes aim at both of these simultaneously:

The complexity of the brain is simply awesome. Every structure has been precisely shaped by millions of years of evolution to do a particular thing, whatever it might be. It is not like a computer, with billions of identical transistors in regular memory arrays that are controlled by a CPU with a few different elements. In the brain every individual structure and neural circuit has been individually refined by evolution and environmental factors. The closer we look at the brain, the greater the degree of neural variation we find. Understanding the neural structure of the human brain is getting harder as we learn more. Put another way, the more we learn, the more we realize there is to know, and the more we have to go back and revise our earlier understandings. We believe that one day this steady increase in complexity will end—the brain is, after all, a finite set of neurons and operates according to physical principles. But for the foreseeable future, it is the complexity brake and arrival of powerful new theories, rather than the Law of Accelerating Returns, that will govern the pace of scientific progress required to achieve the singularity.

Very true. Imagine a fiendishly complex chip diagram, but with not a single component of it standardized. It's one bespoke piece of hardware after another, billions of them, and the wiring between them was put together the same idiosyncratic way. And it's altering while you study it - in fact, it may be altering because you're studying it. Glorious stuff, and understanding it is going to give us extraordinary powers. But that's not happening soon, or on anyone's schedule.

Comments (19) + TrackBacks (0) | Category: General Scientific News | The Central Nervous System


1. Chemjobber on October 17, 2011 10:23 AM writes...

Yes. This explains a lot, including the craptastic performance of the Portland Trailblazers.

Permalink to Comment

2. lazybratsche on October 17, 2011 10:56 AM writes...

Each component bespoke, and each seemingly designed by a different grotesquely baroque madman. And, in spite of itself, everything comes together in an astonishingly robust manner.

Permalink to Comment

3. Bernard Munos on October 17, 2011 11:02 AM writes...

I agree with your perspective, Derek. In a world bounded by constraints, growth curves tend to be not exponential, but S-shaped. The first part, until the inflection point, looks exponential, and this has spawned much of the singularity thinking. Unfortunately, because of limitations from quantum physics, heat dissipation issues, or whatever else, the growth curves will eventually level off. When it comes to modeling, brute force has seldom been a good substitute to an understanding of the system being modeled.

Permalink to Comment

4. Morten g on October 17, 2011 11:39 AM writes...

But that means that Ray K won't get to live forever! Oh no!

Permalink to Comment

5. DCRogers on October 17, 2011 11:45 AM writes...

A species of animal contains individuals, all different, created by evolution -- and yet though study, we actually have amazingly deep and useful understanding of the animal kingdom.

And it's hard to give hard-wired evolution full credit for the diversity within the brain, as much of the complexity is emergent from interaction with the environment. Given human DNA contains ~3B base pairs, that's about 1.5GB of information -- even it were all to code for brain function, that's large, but hardly insurmountable.

His conclusion (it will take a very long time before we understand this complex thing) may or may not be true, but it's not supported by arguments of evolutionary complexity.

Permalink to Comment

6. johnnyboy on October 17, 2011 12:59 PM writes...

Your quote deals with the brain, but pretty much the same thing could be said for the cell. And investigation of molecule function at the cellular level is a couple of orders of magnitude more complex and difficult than investigation of neurons, which at least you can image to some extent.

Permalink to Comment

7. luysii on October 17, 2011 1:18 PM writes...

Very much agree with #6 johnnyboy. For some examples of how even a wiring diagram of the brain (such as we have for any computer chip) won't be enough see
It also contains some references to the current literature if you're still game after reading it.

Permalink to Comment

8. matt on October 17, 2011 1:32 PM writes...

@DCRogers #5: 1.5GB of information in DNA base pairs? You haven't been paying attention for several years now. Metagenomics vastly increases the information content of DNA: various different points on or bonded to DNA (or bonded to something bonded to DNA) can be methylated, aminated, phosphorylated, etc. And that's just construction: once constructed, the interactions explode in complexity and feed back. It's more like 2 to the 1.5GB.

Permalink to Comment

9. Anon anon anon on October 17, 2011 1:52 PM writes...

Just to follow up on @7 matt:

"Parts list" is a better model for the information in DNA than "computer program". Furthermore, if you consider post-translational modifications such as alternative splicing, you see that the parts can be cut and shimmed and glued and welded...

Permalink to Comment

10. Olivier Galibert on October 17, 2011 3:50 PM writes...

Planes don't flap wings.

You don't need to, nor want to, imitate the brain to build an AI (whatever an AI is). And in any case, we build tools that are useful for tasks, and it's not really clear what tool/task can end up in a singularity. Giving human motivations to computer programs which do not have the associated meatware makes little sense.


Permalink to Comment

11. monoceros4 on October 17, 2011 10:54 PM writes...

God, it's always grotesque to watch computer programmers try to think, isn't it?

Permalink to Comment

12. Jose on October 17, 2011 11:24 PM writes...

"In a world bounded by constraints, growth curves tend to be not exponential, but S-shaped"

I go even beyond that and say non-biological growth curves can be anything you could possibly imagine, and probably some you (plural) can't.

Permalink to Comment

13. dreamer on October 18, 2011 6:47 AM writes...

I'm happy that such concepts get some attention, even though they seem very improbable in the near future. It takes a lot of black swan eggs to cook up the singularity and you have to find the swans in the first place, but once we build the first intelligence capable of designing even more advanced advanced intelligences than it and so on, then just maybe the next plateau
in our growth curve will be significantly higher... After all, right now there is a very large gap between what is physically possible and what is practical and intelligence is the best tool to close it.

Permalink to Comment

14. Jim on October 18, 2011 9:11 AM writes...

@#6 johnnyboy- if neurons are cells, how can cells be more complex than neurons? Not only do the type of complex cellular processes that you're hinting at (such as spatially controlled release of calcium from the endoplasmitc reticulum) affect neuronal functioning, you also have to deal with glial cells that modulate neuronal functioning in order to begin to understand neuronal processing.

I don't remember the who said it, but my favorite quote along these lines is "If the brain were so simple that we could understand it, then we'd be too simple to do so."

Permalink to Comment

15. luysii on October 18, 2011 10:27 AM writes...

#14 -- I think what johnnyboy is referring to is the rather simplistic description of neurons (due to McCulloch and Pitts in the 40's) as integrate and fire devices (with essentially no internal structure). It certainly had its uses.

We've moved on and now one can regard each synapse on a neuron (held to average 10,000 per neuron) as a separate computational device. For an interesting review of why most synapses on neurons are found on dendritic spines see Neuron vol. 71 pp. 772 - 781 '11

Permalink to Comment

16. RKN on October 18, 2011 11:55 AM writes...

The most immediate and relevant information for understanding phenotype is protein - its abundance, modification, structure (folds), and interactions - none of which is found in the information of DNA.

And alternative splicing is a post-transcriptional process, not a post-translational process.

Permalink to Comment

17. johnnyboy on October 18, 2011 4:20 PM writes...

@ #14: A neuron is not a brain. The quote was discussing the complexity of brain organization (ie. considering mainly the interconnections between the neurons that compose it). I compared this to the complexity of all the molecules that make up a single cell, whatever that cell is. My own impression is that the functioning of one cell is actually more complex than the functioning of a brain.

Permalink to Comment

18. Matt Bamberger on October 18, 2011 8:25 PM writes...

The brain is complex, but not as complex as all that. There are billions of neurons, yes, but they aren't all bespoke. There are maybe hundreds of different types, but not billions.

Permalink to Comment

19. Vince on October 24, 2011 7:33 AM writes...

Silly recycled argument. I'm quite disappointed that this comes from the likes of a Paul Allen or Derek Lowe.

Just in addressing the above quoted portion, once again someone (Allen) is confusing the map for the territory. The bulk of the information in any neural system (or physical system for that matter) is contained at the lowest levels of description where the universe is computing itself. Fortunately, the 'exciting' things can be summed up in a more abstract, vastly lower resolution map. We don't need to be concerned with quarks when dealing with a question of GPCR's or worry about quantum effects influencing even the firing of a single voltage gated sodium channel as eloquently explained by Max Tegmark.

The more we learn, the more we can artfully find shortcuts that nature, using a more random walk on a very rugged fitness landscape, couldn't possibly 'see' to efficiently scale our map to the desired resolution, irrespective of the ultimate terrain that might exist below. The flaw is in your assumption of the size and accuracy of the configuration space we NEED to describe to be able to affect an outcome: I do believe you are mistaken, it's much more compact and much more sparse.

And as an aside, we DO know the ultimate entropic limit of the terrain thanks to the Bekenstein-Hawkins bound. We also have no current complete quantum-gravity theory that can actually describe the majority of the computation going on in the world: that happens well, well, well under the systems biology level being discussed here. Yet, we get on just fine in neurophysiology using the Hodgkin-Huxley equations -- that are descriptive equations with no direct mapping to a physical description, as per HH themselves. We also know Derek has been successful at designing molecules without having a complete quantum-gravity theory that allows him to 'understand' what's actually happening in his reactions. Yet, as one of my neural-engineering professors once told me: things just work. Systems are robust. Symbol manipulation and Pattern matching work,

Invariably, every few months we get another person who is quite intelligent and knowledgeable in their respective field, write a piece on why their field is so deep and unknowable. Why it's their field which will buck the overarching statistical trends. But, the beat still goes on: quicker and quicker.

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry