Ack -- I can't
stand the noise any longer! Those
who dispute the physical viability of cooling via quantum tunneling, please see
Hishinuma
et al, Applied Physics Letters, 23 April, 2001.
The CoolChips folks claim (see the slide show) to have
beaten the Stanford group by several years, but CoolChips also relies upon
details the Hishinuma paper for justification.
As for the various concerns voiced about large electrostatic
forces between the cathode and anode, there are no such forces because there is
no charge imbalance. When an
electron tunnels across the gap it is replaced. Its a circuit.
The physical principle behind the device is quite sound. The materials science and device
engineering are, of course, another matter entirely. I wish them luck, but I am not yet any more likely to invest than
the rest of you.
For the record, as the author of the essay in question, let me point out a few things.
1. Many slashdot readers commented that having the genome isn't the whole game, and this is correct. As I suggest in the essay, real progress in building things out of biology will happen only with the advent of design tools. These design tools will evolve from predictive models, just like in every other field of engineering. The models are being built now, and are probably 5 to 10 years from being useful as design tools. (Humans are, or course, already manipulating genetic systems without models to aid in understanding the process.)
2. The models will be aimed at predicting the behavior of biological systems at the most appropriate resolution. In general, the goal is to build models that quantitatively describe protein-protein interactions and genetic networks, though there are efforts underway to plug metabolism in early on. One difficulty inherent in the process is that the ability measure and model varies significantly depending on the system studied. There are some systems where a single molecule can be identified in a cell, and others where millions of cells are required to get a signal. This directly affects what sorts of models are built and how appropriate they are for the system studied.
For example, there are many models now built on a framework of deterministically solved differential equations. These models are only useful to predict the behavior of deterministic systems. However, many real biological systems involve only a small number of molecules and therefore must be treated stochastically. Yet the experimental tools to provide data at this level of resolution are only just being developed. There is significant progress on both modeling and experimental fronts ongoing.
See, for example: opnsrcbio.molsci.org/alpha/
3. The cost of biological technologies are falling exponentially while the productivity they enable is increasing exponentially. These are independent in the following way: reagents and instruments are becoming cheaper while simultaneously (and independently) becoming more capable through automation and parallelization. The parts for a "state of the art" parallel DNA synthesizer cost only ~$10,000, and both this monetary sum and the assembly cost are well within that expended by computer and car hobbyists. This machine will let you make enough DNA to assemble smallpox from scratch in about a week. Of course, I am not advocating this course of action, but it is worth thinking about. It seems relatively obvious that given the opportunity to tinker, humans will.
More information along these lines is available at: www.molsci.org/~rcarlson/exp_biol_tech.html
4. A general comment about Open Source Biology: Biology is, of course, already very close to open source when you think about how methods are shared. The whole issue of whether the open source software model is useful or appropriate for IP in biology is open. The real question you need to ask yourself is: What happens if there is a Biosoft? Could be good, could be bad -- I don't have the answer. But I suggest we begin considering the question since there are already commercial efforts underway to create proprietary sets of molecules for manipulating biological systems.
5. As far as people's thoughts about whether it will work or not -- it already does. Many pharmaceuticals and industrial enzymes are produced in vats growing bacteria (BACTERIA, not yeast for you home fermentation enthusiasts...). Laundry enzymes are a multi-billion dollar market. Like your snowy slopes? Snowmax is an ice-nucleating protein produced recombinately and used in snowmaking machines world-wide. If memory serves, it ranks first among recombinant proteins by produced weight.
Many labs are inducing plants to produce "biopharmaceuticals". Moreover, there are many projects underway, both commercial and academic, to grow tissues in vats. Artificial bladders have been done and will be tested in humans soon, livers are in progress, and one of the heart projects, headed up by people at MIT, is making very interesting progress. This is difficult work, but not pie in the sky by any means.
Finally, given the audience here, you can help decide how fast this goes and whether it takes place in the open or not. Help figure out the models, help write the code, help write the Bio-GPL (actually the Lesser licence is our favorite model at the moment). Or not.
Slashdot Mirror
See previous slashdot discussions of "Open Source Biology And Its Impact on Industry" (slashdot discussion) from 2001, and for how fast costs are falling and skills are spreading "The Pace and Proliferation of Biological Technologies" (slashdot discussion) from 2003.
I am, by the way, posting this from the Synthetic Biology 1.0 Meeting at MIT. Open source is under discussion, as are the risks.
- Rob
For the record, as the author of the essay in question, let me point out a few things.
1. Many slashdot readers commented that having the genome isn't the whole game, and this is correct. As I suggest in the essay, real progress in building things out of biology will happen only with the advent of design tools. These design tools will evolve from predictive models, just like in every other field of engineering. The models are being built now, and are probably 5 to 10 years from being useful as design tools. (Humans are, or course, already manipulating genetic systems without models to aid in understanding the process.)
2. The models will be aimed at predicting the behavior of biological systems at the most appropriate resolution. In general, the goal is to build models that quantitatively describe protein-protein interactions and genetic networks, though there are efforts underway to plug metabolism in early on. One difficulty inherent in the process is that the ability measure and model varies significantly depending on the system studied. There are some systems where a single molecule can be identified in a cell, and others where millions of cells are required to get a signal. This directly affects what sorts of models are built and how appropriate they are for the system studied.
For example, there are many models now built on a framework of deterministically solved differential equations. These models are only useful to predict the behavior of deterministic systems. However, many real biological systems involve only a small number of molecules and therefore must be treated stochastically. Yet the experimental tools to provide data at this level of resolution are only just being developed. There is significant progress on both modeling and experimental fronts ongoing.
See, for example: opnsrcbio.molsci.org/alpha/
3. The cost of biological technologies are falling exponentially while the productivity they enable is increasing exponentially. These are independent in the following way: reagents and instruments are becoming cheaper while simultaneously (and independently) becoming more capable through automation and parallelization. The parts for a "state of the art" parallel DNA synthesizer cost only ~$10,000, and both this monetary sum and the assembly cost are well within that expended by computer and car hobbyists. This machine will let you make enough DNA to assemble smallpox from scratch in about a week. Of course, I am not advocating this course of action, but it is worth thinking about. It seems relatively obvious that given the opportunity to tinker, humans will.
More information along these lines is available at: www.molsci.org/~rcarlson/exp_biol_tech.html
4. A general comment about Open Source Biology: Biology is, of course, already very close to open source when you think about how methods are shared. The whole issue of whether the open source software model is useful or appropriate for IP in biology is open. The real question you need to ask yourself is: What happens if there is a Biosoft? Could be good, could be bad -- I don't have the answer. But I suggest we begin considering the question since there are already commercial efforts underway to create proprietary sets of molecules for manipulating biological systems.
5. As far as people's thoughts about whether it will work or not -- it already does. Many pharmaceuticals and industrial enzymes are produced in vats growing bacteria (BACTERIA, not yeast for you home fermentation enthusiasts...). Laundry enzymes are a multi-billion dollar market. Like your snowy slopes? Snowmax is an ice-nucleating protein produced recombinately and used in snowmaking machines world-wide. If memory serves, it ranks first among recombinant proteins by produced weight.
Many labs are inducing plants to produce "biopharmaceuticals". Moreover, there are many projects underway, both commercial and academic, to grow tissues in vats. Artificial bladders have been done and will be tested in humans soon, livers are in progress, and one of the heart projects, headed up by people at MIT, is making very interesting progress. This is difficult work, but not pie in the sky by any means.
Finally, given the audience here, you can help decide how fast this goes and whether it takes place in the open or not. Help figure out the models, help write the code, help write the Bio-GPL (actually the Lesser licence is our favorite model at the moment). Or not.
Your choice.
- Rob
Is it just me or does this article read like it was written by a sixth grader?
Yep, it's just you.