8 ms·

That's interesting perspective. > And after seeing many, many breakthroughs where the 'new' bit was doable in small scales at great expense in a laboratory, bu

by tomp 8y ago

That's interesting perspective. > And after seeing many, many breakthroughs where the 'new' bit was doable in small scales at great expense in a laboratory, but impossible to do at scale in an automated process, Can you elaborate on that, maybe give a few examples? I always thought that inventions were mainly judged by their merit (efficiency), I guess I was blind to the "scaling" (i.e. "actual industry") variable, and just assumed that there must be some way to automate and streamline almost everything, as long as the demand is there (as Elon Musk has recently been showing, with batteries/electric cars and rockets).

petra 8y ago
Chemistry(and biology) is one such field: the chemical/biological process behaves very differently whether it's don't in small quantities of the lab or in the huge volumes required for a commercial effort. So think of that large($100mm) factory as a research process , that can often fail. Example: commercializing biotech often fails because it's not possible to control contamination and genetic evolution of the bacteria you grow in huge processes .
ChuckMcM 8y ago
I have found that many times when people are "talking past" each other it is because there is a belief in there somewhere that one holds one way and the other holds a different way. In this case you have illuminated it nicely, which (slightly edited) is this: I ... assumed that there must be some way to automate and streamline almost everything, as long as the demand is there. I have experienced many things over my life where this assumption was invalid. The most common reason I have seen that this assumption breaks down is that "demand" is expressly tied to "price." And more specifically when the price goes up, demand goes down (in the economic sense, the number of people willing to pay that price for a widget) and at some price the demand goes to zero. Thus "not possible" which is not literally true, it is only actually true. Do a search on "Battery breakthrough" (Google Scholar will give you better results but the web search works too) and pick any one where the improvement is >= 50% improvement in any axis (cost, charging, capacity, weight, what ever). This works best for articles that are 5 or more years old (you have to allow some time for them to come up with a way to do it), and then try to find that breakthrough used in any battery today. Sometimes you will find a 'why we couldn't do it' article but mostly it will just not be there. Even though batteries that were 50% better on any axis would seem to have a market right? And then read through to find the 'hard bit' in the break through and go back to Google Scholar and see if you can find papers related to that bit. You will find "the rest of the story" as it is sometimes said where some thing prevents wider adoption than laboratory curiosity. Here is a simpler technology, silicon anodes (announced in 2010) reviewed in 2018 [1], maybe hitting the market if they can go to production [2]. This is a simpler process and as it has matured its impact has gone down to 10 - 15%. [1] https://sci-hub.tw/10.1016/j.nanoen.2016.07.023 https://sci-hub.tw/10.1016/j.nanoen.2016.07.023 [2] [BMW] plans to incorporate Sila Nano batteries into vehicles by 2023 to boost energy capacity by 10 to 15 percent. --https://venturebeat.com/2018/08/16/sila-nanotechnologies-raises-70-million-for-next-generation-battery-development/ https://venturebeat.com/2018/08/16/sila-nanotechnologies-rai...
- csours 8y ago
  > The most common reason I have seen that this assumption breaks down is that "demand" is expressly tied to "price." I knew this, but reading it again in this context makes many business decisions very clear.
  - logicallee 8y ago
    >I knew this, but reading it again in this context makes many business decisions very clear. Yeah. You can get this very clear in your mind by considering a few examples. Here's one. How much demand is there for legal, genuine, clean $2 bills that are legal currency and have no meaningful restrictions (won't get you in trouble, nothing special about it)? Barring anything I can think of, just your personal demand for such bills should be at least a billion of them... at the price point of $1.99 each. (For a profit of 1 billion * $0.01 = $10 million USD. There are some logistics involved, according to Google a U.S. dollar bill weighs about a gram, so a billion grams weigh in at around 1102 tons. So that is approximately $9074 per ton. But for $10 million you'll figure the logistics out.) And now compare your personal demand for them at $2.00. That's probably about 0 in number. Why would you want even one of them, they're less useful than the $2 in other currency you bought it for. So just your personal demand for a legal $2 bill is either a billion of them or 0 of them, depending on whether we're talking about the price point of $2.00 or $1.99. Demand is at a given price point.
- tomp 8y ago
  Right, but that the demand/supply curve isn't really my point. (Maybe we're still "talking past" each other.) Nobody wants to buy 2GB of memory at a "640 kB ought to be enough for anybody" price point; that's fairly obvious. My point is, human civilisation has proven time and again, especially when it comes to micro/nano-electrical/chemical technology (batteries, transistors, screens, chips, ...), that it is possible to optimize, improve, reduce price and make more efficient, production (while making the product better) beyond what anyone thought possible (well except Gordon Moore). Why would batteries be different? If I find some way to make a battery better, 50% more efficient and/or 50% lighter, why can't I just dump $1-10bn into research/production and then dominate the market (likely worth 10-100x more)? Is there a fundamental technological reason (for some technologies/inventions at least), or is it just a matter of economics/management (too risky)? I'm guessing this comment reads fairly unclear (my thoughts aren't particularly clear on this matter either), but hopefully it gets some of my point across. Edit: petra gives a good example below - scaling of biological processes often fails; which makes sense because biology - we don't understand how life works, and it changes/evolves/moves by itself! But it seems that non-biological processes should be manageable, at least at this scale (i.e. not quantum/high-energy processes).
another-cuppa 8y ago
> and just assumed that there must be some way to automate and streamline almost everything, as long as the demand is there Scaling a solution to the travelling salesman problem would solve many other problems. The demand is definitely there. But many people believe it cannot be done.
- tomp 8y ago
  That's a fundamentally different problem. "Scaling" it requires figuring out how to solve a different concrete problem every time. Scaling e.g. battery production requires figuring out how to solve the same real-world problem very efficiently.
  - another-cuppa 8y ago
    No, you could (possibly) solve TSP once by providing an algorithm that scales well just like inventing an industrial process. My point was that it looks very much like some problems are fundamentally difficult and have no scalable solutions.

by tomp 8y ago

petra 8y ago

Chemistry(and biology) is one such field: the chemical/biological process behaves very differently whether it's don't in small quantities of the lab or in the huge volumes required for a commercial effort. So think of that large($100mm) factory as a research process , that can often fail. Example: commercializing biotech often fails because it's not possible to control contamination and genetic evolution of the bacteria you grow in huge processes .

ChuckMcM 8y ago

I have found that many times when people are "talking past" each other it is because there is a belief in there somewhere that one holds one way and the other holds a different way. In this case you have illuminated it nicely, which (slightly edited) is this: I ... assumed that there must be some way to automate and streamline almost everything, as long as the demand is there. I have experienced many things over my life where this assumption was invalid. The most common reason I have seen that this assumption breaks down is that "demand" is expressly tied to "price." And more specifically when the price goes up, demand goes down (in the economic sense, the number of people willing to pay that price for a widget) and at some price the demand goes to zero. Thus "not possible" which is not literally true, it is only actually true. Do a search on "Battery breakthrough" (Google Scholar will give you better results but the web search works too) and pick any one where the improvement is >= 50% improvement in any axis (cost, charging, capacity, weight, what ever). This works best for articles that are 5 or more years old (you have to allow some time for them to come up with a way to do it), and then try to find that breakthrough used in any battery today. Sometimes you will find a 'why we couldn't do it' article but mostly it will just not be there. Even though batteries that were 50% better on any axis would seem to have a market right? And then read through to find the 'hard bit' in the break through and go back to Google Scholar and see if you can find papers related to that bit. You will find "the rest of the story" as it is sometimes said where some thing prevents wider adoption than laboratory curiosity. Here is a simpler technology, silicon anodes (announced in 2010) reviewed in 2018 [1], maybe hitting the market if they can go to production [2]. This is a simpler process and as it has matured its impact has gone down to 10 - 15%. [1] https://sci-hub.tw/10.1016/j.nanoen.2016.07.023 https://sci-hub.tw/10.1016/j.nanoen.2016.07.023 [2] [BMW] plans to incorporate Sila Nano batteries into vehicles by 2023 to boost energy capacity by 10 to 15 percent. --https://venturebeat.com/2018/08/16/sila-nanotechnologies-raises-70-million-for-next-generation-battery-development/ https://venturebeat.com/2018/08/16/sila-nanotechnologies-rai...

csours 8y ago
> The most common reason I have seen that this assumption breaks down is that "demand" is expressly tied to "price." I knew this, but reading it again in this context makes many business decisions very clear.
- logicallee 8y ago
  >I knew this, but reading it again in this context makes many business decisions very clear. Yeah. You can get this very clear in your mind by considering a few examples. Here's one. How much demand is there for legal, genuine, clean $2 bills that are legal currency and have no meaningful restrictions (won't get you in trouble, nothing special about it)? Barring anything I can think of, just your personal demand for such bills should be at least a billion of them... at the price point of $1.99 each. (For a profit of 1 billion * $0.01 = $10 million USD. There are some logistics involved, according to Google a U.S. dollar bill weighs about a gram, so a billion grams weigh in at around 1102 tons. So that is approximately $9074 per ton. But for $10 million you'll figure the logistics out.) And now compare your personal demand for them at $2.00. That's probably about 0 in number. Why would you want even one of them, they're less useful than the $2 in other currency you bought it for. So just your personal demand for a legal $2 bill is either a billion of them or 0 of them, depending on whether we're talking about the price point of $2.00 or $1.99. Demand is at a given price point.
tomp 8y ago
Right, but that the demand/supply curve isn't really my point. (Maybe we're still "talking past" each other.) Nobody wants to buy 2GB of memory at a "640 kB ought to be enough for anybody" price point; that's fairly obvious. My point is, human civilisation has proven time and again, especially when it comes to micro/nano-electrical/chemical technology (batteries, transistors, screens, chips, ...), that it is possible to optimize, improve, reduce price and make more efficient, production (while making the product better) beyond what anyone thought possible (well except Gordon Moore). Why would batteries be different? If I find some way to make a battery better, 50% more efficient and/or 50% lighter, why can't I just dump $1-10bn into research/production and then dominate the market (likely worth 10-100x more)? Is there a fundamental technological reason (for some technologies/inventions at least), or is it just a matter of economics/management (too risky)? I'm guessing this comment reads fairly unclear (my thoughts aren't particularly clear on this matter either), but hopefully it gets some of my point across. Edit: petra gives a good example below - scaling of biological processes often fails; which makes sense because biology - we don't understand how life works, and it changes/evolves/moves by itself! But it seems that non-biological processes should be manageable, at least at this scale (i.e. not quantum/high-energy processes).

another-cuppa 8y ago

> and just assumed that there must be some way to automate and streamline almost everything, as long as the demand is there Scaling a solution to the travelling salesman problem would solve many other problems. The demand is definitely there. But many people believe it cannot be done.

tomp 8y ago
That's a fundamentally different problem. "Scaling" it requires figuring out how to solve a different concrete problem every time. Scaling e.g. battery production requires figuring out how to solve the same real-world problem very efficiently.
- another-cuppa 8y ago
  No, you could (possibly) solve TSP once by providing an algorithm that scales well just like inventing an industrial process. My point was that it looks very much like some problems are fundamentally difficult and have no scalable solutions.