Comments on: Thinking quantitatively about technological progress

By: jabelar

jabelar — Tue, 19 Jul 2011 20:12:58 +0000

I think you can measure replacement and obsolescence. Simply measuring sales isn’t quite right because the product definition changes — a smart phone next year is quite different than one this year.

I think the way you’d quantify it is you’d look at the sales of a very specific technology product declining and see if it is being replaced by a similar volume of a next generation device. Presumably replacement is due to something better (in any dimension of functionality, performance, cost, etc.) and therefore constitutes progress.

Looking at replacement would help clarify what is actually saturating a market (i.e. the example of railroads in your article) versus things declining due to obsolescence.

By: Jake Witmer

Jake Witmer — Mon, 18 Jul 2011 19:08:33 +0000

What I think might help: plugging in as much data as possible to a Numenta / Hawkins-type brain architecture, and tell it to separate like/similar patterns (while keeping track of what is known about each pattern). I’d expect to see a large jump in mining and things attached to physical limits as well, come from better computation (much of mining is prospecting, which is bound by information laws, not so much physical capacity, once one has the economic resources), even if the jump wasn’t as large as the one regarding software capacity. Also, you’d need to drill down, and determine when a killer app was written for mining companies, etc…

Go strictly by what is able to be directly physically measured. Ask companies for raw data in units produced, and plug their data in. That’s the best way to get reliable predictions.

By: RBynum

RBynum — Sun, 17 Jul 2011 21:47:32 +0000

You may be interested in the book “What Technology Wants” by Kevin Kelly. In particular he discusses the drivers of technology and some of the effforts of others to project the path of technological progress.

I thnik it its interesting to note the rapid decrease in the cost of DNA sequencing, lab on a chip diagnostic devices, progress in electronic records keeping and patient monitioring. What could this technological path lead to in the near future?

By: RBynum

RBynum — Sun, 17 Jul 2011 03:36:28 +0000

Many advances result from a cross polination beween multiple lines of techonology advances. This would be simlar to genetic selection and mutations. The fitness selection process is through the economic/cultural forces. So what we desire to have and what we can afford to produce and sell at a profit determines the direction of techological progress.

Consider the current desire for affordable green energy and advances in solar cell efficency, battery storage improvements, and electric smart grids. These drivers are moving techology forward in one direction. While biofuels are being driven by advances in genetic studies of cellular organisms, synthetic life, and bioreactor designs are driving energy technology in another direction.

To predict technological progress you need multiple future experience/ cost curves for various tecnology lines and some polling of desirable future features. When the experience/cost lines meet the desirable future features at a reasonable time in the future (say 5 years from now) you will have a movement towards that technology point and a convergence of the technolgy lines needed to acheive that point.

By: jaknight89

jaknight89 — Wed, 13 Jul 2011 00:50:16 +0000

Diminishing marginal utility is that factor that causes the technological trend to reach a limit point and fall. I think we need a reconstruction of multiple facets of our culture to create economic incentive for improved technological advancement.

By: gillammi

gillammi — Mon, 11 Jul 2011 16:23:57 +0000

Quantifying the Frequency of Disruptive Technologies is something I have yet to see done. For example, railroads in every city may saturate, but the cost of movement continues to decrease and the breadth of places to travel increases from innovation from other transport mediums – ships, cars, buses, trucks, planes. The replacement of hard drives with flash memory is an example from Clayton Christiansen. These punctuated disruptions must occur to create the step-wise stack of sigmoidal curves. The question is whether this step-wise stack is predictable. Can we predict the rough moment when disruption is expected?

Another principle to quantify is the impact of Product Upgrades. For example, though the world may saturate itself with the iPhone 4 – customers may upgrade to the iPhone 5 whether needed or not. What is the quantitative role in sales of replacements & upgrades?

By: Khannea Suntzu

Khannea Suntzu — Mon, 11 Jul 2011 13:33:43 +0000

Now that’s interesting – you are trying to model progress as a diffuse, holistic, abstract (algorithmic) mechanisms. If you can do that you can also adjust for the same diffuse, holistic, abstract (algorithmic) detractors to progress. There are quite a few, and I am sure you, me, and your side kick could both come up with a list. Peak Oil, Climatic disturbance and Disparity are a few that would be modelacious.

Even better, if you use the world quantitative you miught also be able to model a deep field, where you can depict outcomes (ranging from utopian to desirable to dystopian and hellish) in a graph. I.e. depict policies and choices in a nice visualization. Kinda Asimovian, right?

People respond out of bounds to visualizations and pretty pictures. If you are able to spur on/create a visualization tool with colors and timelines and signposts and critical transition moments, you can rub that same thing in the face of our beloved policy makers.

Might even change their more flawed policies a little.