interesting interview by rajiv malhotra with a technology VC. listen and weep, both india and the US: having lost manufacturing, the US will lose its R&D as well. david teece said this long ago in a seminal paper from UC Berkeley in like 1980 or so.
PV/EV, semiconductor manufacturing technology, and pervasive AI/ML.
I can so relate to what Mukesh Chatter says at 25:30:
"I will compare this to one more example because I come from that world of computer engineering. When you design chips and circuit boards there's a team that designs the integrated circuits. There's a completely independent team that verifies independently that the chip design is what it is supposed to do. And those two teams, the verification team does not work for the design team. They independently verify and only come together at the top. So there's no incentive for compromises at any level. The same mechanism must exist if we want to succeed in India, which is when the budget allocation is done somebody else should be verifying that it is being implemented right. Innovation isn't enough. Innovation must be implemented and if it is implemented then it becomes Incredible India".
The idea has universal application in governance, in corporations, in defense, in strategy, everywhere. For example, an independent strategy-verification team would have foreseen the reaction to CAA or to Farm laws by India-Breaking/Andolanjeevi crowd and armed the government with counter measures.
If some day, we develop the courage to dismantle Indian Administrative Services, it should be based on this very idea. If you have the appetite for a seminal work on decision-making, be it in governance or at corporate level, , Rajeev, it should be built on this simple idea.
"Again, please don't ignore the key problem, which is the problem of successful countries self destroying themselves in multiple dimensions. The science/technology stagnation of US is caused by the non-technical people actively out-breeding the scientists and engineers. This is a key trend visible in almost all developed countries."
the US will figure out what to do, they are smart guys. their main problem has been short-term profit chasing by CEOs and management consultants: so they exported all manufacturing to china. this they will reverse. the far more important question for us is, "what is india doing to upgrade its technology and R&D capability?"
In semiconductors, we can start at 65 nanometer, which is old enough for the US to part with the technology (assuming willingness on their part), but capable enough for many applications. Most chips in industrial, military and even automobiles do not need latest and greatest FinFet technology. Mobile phones , PCs and server chips are a different matter. But all this will be worthless unless we sustain R&D spending, not for a year or two but forever.
in general, yes, earlier generation tech is enough to get on the semiconductor bandwagon, tho i'm not sure who owns the tech: is it people like applied materials in the US and canon in japan, or is it now taiwan semiconductor and samsung and intel? it should be possible to get going with older tech (i thought it was like 20nm) and built up capability esp for military applications and then slowly move up the learning curve.
- 65nm is 15-year-old tech. 65 can be optically shrunk to 55nm. Anything more advanced, I doubt anyone will be willing to share with us.
- China's SMIC was making rapid strides, even mastering FinFET technology, till Trump put them on blacklist.
- I don't think any one company holds the whole tech; ASML, a Dutch company, is the sole provider of advanced lithography technology for 10nm and below.
ok. i thought canon was doing well on photolithography. but there's no question that we have to have fabs (even if primitive) here for military and security purposes. there are new RISC-V designs (eg shakti from iitm) that would benefit from fab manufacturing know-how. there is also the possibility of GaN, GaAs and other exotic materials to do R&D on.
I can so relate to what Mukesh Chatter says at 25:30:
"I will compare this to one more example because I come from that world of computer engineering. When you design chips and circuit boards there's a team that designs the integrated circuits. There's a completely independent team that verifies independently that the chip design is what it is supposed to do. And those two teams, the verification team does not work for the design team. They independently verify and only come together at the top. So there's no incentive for compromises at any level. The same mechanism must exist if we want to succeed in India, which is when the budget allocation is done somebody else should be verifying that it is being implemented right. Innovation isn't enough. Innovation must be implemented and if it is implemented then it becomes Incredible India".
The idea has universal application in governance, in corporations, in defense, in strategy, everywhere. For example, an independent strategy-verification team would have foreseen the reaction to CAA or to Farm laws by India-Breaking/Andolanjeevi crowd and armed the government with counter measures.
If some day, we develop the courage to dismantle Indian Administrative Services, it should be based on this very idea. If you have the appetite for a seminal work on decision-making, be it in governance or at corporate level, , Rajeev, it should be built on this simple idea.
Comment by Vinay Krishnan on Youtube makes sense:
"Again, please don't ignore the key problem, which is the problem of successful countries self destroying themselves in multiple dimensions. The science/technology stagnation of US is caused by the non-technical people actively out-breeding the scientists and engineers. This is a key trend visible in almost all developed countries."
the US will figure out what to do, they are smart guys. their main problem has been short-term profit chasing by CEOs and management consultants: so they exported all manufacturing to china. this they will reverse. the far more important question for us is, "what is india doing to upgrade its technology and R&D capability?"
In semiconductors, we can start at 65 nanometer, which is old enough for the US to part with the technology (assuming willingness on their part), but capable enough for many applications. Most chips in industrial, military and even automobiles do not need latest and greatest FinFet technology. Mobile phones , PCs and server chips are a different matter. But all this will be worthless unless we sustain R&D spending, not for a year or two but forever.
in general, yes, earlier generation tech is enough to get on the semiconductor bandwagon, tho i'm not sure who owns the tech: is it people like applied materials in the US and canon in japan, or is it now taiwan semiconductor and samsung and intel? it should be possible to get going with older tech (i thought it was like 20nm) and built up capability esp for military applications and then slowly move up the learning curve.
- 65nm is 15-year-old tech. 65 can be optically shrunk to 55nm. Anything more advanced, I doubt anyone will be willing to share with us.
- China's SMIC was making rapid strides, even mastering FinFET technology, till Trump put them on blacklist.
- I don't think any one company holds the whole tech; ASML, a Dutch company, is the sole provider of advanced lithography technology for 10nm and below.
ok. i thought canon was doing well on photolithography. but there's no question that we have to have fabs (even if primitive) here for military and security purposes. there are new RISC-V designs (eg shakti from iitm) that would benefit from fab manufacturing know-how. there is also the possibility of GaN, GaAs and other exotic materials to do R&D on.