6 ms·

Yes, using obsolete data from when tech was bleeding edge in 1999 to imply things about cost going forward is misleading. On the other hand using improvement o

Yes, using obsolete data from when tech was bleeding edge in 1999 to imply things about cost going forward is misleading. On the other hand using improvement over the same timeline, drawing a line on a graph and saying "look how X Y is gonna be in Z years" is the same exact type of stupid but pointed in a different direction. In 1991 lithium was highly immature technology and would take about a decade to make it into fragile electronics. It took another decade to make it into power tools. Now it's viable in high end commuter vehicles. If it was easy to predict the future a decade out with any reliability we wouldn't be having this discussion.

ip26 6y ago
drawing a line on a graph and saying "look how X Y is gonna be in Z years" is the same exact type of stupid And yet we do it every ~18 months with semiconductors. Imperfect to be sure, but if you don’t skate to where the puck is headed, you will miss opportunity.
- dmoy 6y ago
  > And yet we do it every ~18 months with semiconductors. So, this is about to stop. Very soon. Quantum tunneling, yield rates, etc are all starting to be cost prohibitive. <10nm stuff was already delayed. There's plans for 3nm. It's unclear yet whether 2nm will work. Even if hypothetically it does, there's a very real limit - an atom is only 10x smaller than 2nm.
  - ben_w 6y ago
    On the one hand, I remember people saying the same about nodes much larger than those we currently use. On the other, yes: continue current trends and single atom transistors become standard in 10-32 years.
    - usrusr 6y ago
      I think it boils down to we will hit a wall, but we don't know exactly when we'll hit it. (and how hard, chances are that higher-hanging-fruit refinements will make the transition to stagnancy so gradual that we may not notice at all) A jump from "all past predictions failed" to "and so will all future predictions" seems rather bold to me. In the end it's like a somewhat upended variation of the "x decades to practical fusion" thing where we all hope that the old joke that x might be a natural constant is eventually proven wrong.
      - ghaff 6y ago
        Well, the main direction of refinements at this point seem to be around composable/heterogeneous computing where we basically have a lot of hardware optimized for specific workloads and throw the complexity at the software people. i.e. now deal with GPUs, DPUs, FPGAs, xCPUs, etc. instead of (largely) just a standardized set of CPU instructions.
  - ip26 6y ago
    Sure. But there’s been questions about the viability of the next node for at least twenty years. If you played it safe and stuck with the current node, you’d have been wrong & at a process disadvantage 14 times out of 14. No improvement trend goes on forever. But why is this the moment lithium ion hits the wall? It’s like trying to call the end of a bull market.
  - mlindner 6y ago
    > So, this is about to stop. Very soon. Quantum tunneling, yield rates, etc are all starting to be cost prohibitive. People have been saying that for over 10 years though. "Very soon" keeps being punted off by another 5 years every 5 years.
    - vkou 6y ago
      It has stopped, though. Single-threaded performance, either in absolute value, or per dollar has not been following Moore's law for more than a decade. And outside of data centers, single-thread performance is still king. Performance per watt has improved, but that's not a metric the typical end user cares much about.
      - jononor 6y ago
        Performance per watt is key for all portable devices, ie the laptops and phones that many do almost all of their end-user computing on.
      - mlindner 6y ago
        Single-thread performance is not king though. Software nowadays is extensively multi-threaded. If you're still writing single threaded code for a desktop program people are going to look at you like you have antenna growing out of your head. And performance per watt absolutely is important. People care a lot about battery life of their laptops and phones and that's strongly dependent on performance per watt.
  - cbozeman 6y ago
    Not according to Jim Keller. He believes we still have plenty of room, and no offense to you, but I'll take his word before yours.
    - ksec 6y ago
      And Jim Keller is not a Fab / Processing guy. And if you actually listen to what he said, he is predicting plenty of room to improve, not it will improve every 24 months, which arguably has already stopped happening since ~2017
  - thaumasiotes 6y ago
    > Even if hypothetically it does, there's a very real limit - an atom is only 10x smaller than 2nm. That's certainly a real limit, but what is it a limit on? It is not obvious from first principles that it takes n atoms to implement n transistors.
  - fpoling 6y ago
    2nm is just a commercial name. The real features are much bigger, see, for example, https://en.m.wikipedia.org/wiki/10_nm_process https://en.m.wikipedia.org/wiki/10_nm_process. So we are quite far from 10 atoms per transistor. The biggest problem is lithography. ASML managed unexpectedly for many in the industry including Intel to solve technological problems with extreme ultraviolet sources, but shrinking transistors significantly further requires soft x-rays and the perspectives of that now for mass lithography are much more uncertain than for EUV ten years ago.
  - imtringued 6y ago
    We'll get to 256 core server chips and that will be the end of the line. You are forgetting that the advertised version number denoted by nm is about what size a theoretical planar transistor would need to have to be equivalent to the advertised process. Since that number is purely theoretical, we can construct a theoretical scenario in which its theoretical nature would become absolutely obvious. Take a 7nm process node transistor and stack it 100 times. Such a process would be called a 0.07nm process.
- sir_bearington 6y ago
  Except we don't. Where's my single-core 100GHz processor? Improvements still happen, but not always in the same way. If you implemented an application in 2003 assuming we'd have such a processor you'd be very disappointed. Counting on exponential improvements to continue is risky bet.
  - rictic 6y ago
    Scaling may take us to some strange places, but it's worth noting that an Apple M1 chip has >100x the transistors of a 2004-era Pentium 4, and achieves ~500x the FLOPs at a fraction of the power draw. Yes this isn't single-threaded performance, but I think we should keep in mind that exponential improvement in price/performance over many decades is possible, if never certain. Sources - https://www.alternatewars.com/BBOW/Computing/Computing_Power.htm https://www.alternatewars.com/BBOW/Computing/Computing_Power... - https://en.wikipedia.org/wiki/Apple_M1 https://en.wikipedia.org/wiki/Apple_M1
clomond 6y ago
Except when you consider where each technology fits within its own S-curve of adoption (X axis over time, Y axis is % of the technology adopted by the market). When factoring in the shape of the exponential decreases in costs, and that penetration of most of these technologies is at or before the inflection point (between 5%-15% market penetration), it is more likely that the cost declines will ACCELERATE moving forward rather than slow down. Why has it felt that laptops and PCs haven’t progressed as much in the 2010s as in the 1990s or 2000s? Because in 1995, there was not a computer on every desk in every home. But now not only is the market saturated with laptops and PCs, people are walking around with mini internet connected “super computers” everywhere they go.
- ben_w 6y ago
  > Except when you consider where each technology fits within its own S-curve of adoption (X axis over time, Y axis is % of the technology adopted by the market). Unfortunately, even a very small amount of noise in the data makes is basically impossible to know where you are in an S-curve. Much safer to make predictions based on the far more limited good news that PV+battery is already cheaper than coal for electricity or ICE for cars. Hmm… question for anyone who knows: with current tech, how much would it cost to develop a significant PV-powered electrolysis-and-Sabatier-process plant in any of the big coastal deserts, for exporting methane?
  - clomond 6y ago
    > Unfortunately, even a very small amount of noise in the data makes is basically impossible to know where you are in an S-curve. While true, my point is that when combined with the fact that we are pre-inflection point, and the economics now stand on their own (renewables, Electric Vehcile TCO and various Energy Storage applications being already cheapest, competitive or very close too) it is not unreasonable when mapping out the 5-15 year future to bet on an acceleration of cost declines over a deceleration. Particularly because the actual driver of unit cost declines (Wrights Law/Moore's Law) is the doubling / magnitude of units manufactured and put through the system, for which with all the factories being ramped up and planned - point to the positive in my view on it. Regarding your PV-powered and electrolysis-Sabatier (electrofuel) methane, I think there are two important considerations. In order for methane (or other e-fuels like hydrogen or longer chain hydrocarbons) to be made economically, the capital cost of the equipment needs to be utilized as close to 100% of the time as possible. We already know that PV excess will be centered around the daytime peak (5-7 hours per day) meaning that there would also need to be plenty of excess wind to balance this out to get anywhere close to 100% utilization of the excess energy. Until the electricity grids get sufficiently saturated with renewables broadly, most e-fuel applications will continue to not be competitive, particularly as things like energy storage applications (possibly run off an e-fuel) are likely to be economical prior there being an opportunity for the export of excess e-fuels. That's more at the a end of the S-Curve as far as I can tell.
nardi 6y ago
These are not the same kind of stupid. One makes the assumption that costs will always be the same, and the other makes the assumption that cost decreases are linear, or predictable. The former is much stupider.
- ekianjo 6y ago
  Nope, it's the same thing: excess of confidence in predicting the future.
  - LMYahooTFY 6y ago
    The commenter was point out the nuance between the two, it's obviously about confidence in an assertion. You just re-reduced it to what was already obvious?
  - Qwertious 6y ago
    Investing in government bonds vs investing in penny stocks: both are the same thing, as they are both an excess of confidence in predicting the future.
- wpietri 6y ago
  They are exactly the same kind of error in that both assume stability over time. One assume prices are stable; the other assumes the rate of change is stable.
  - Brakenshire 6y ago
    The rate of change has been stable over the last 15 years, do you bet it continues for a 16th year or that it stops this year?
    - zdragnar 6y ago
      They'll never be free. They'll never be cheaper than the raw materials that go into them, or the copper to wire them together and into the car. At some point, there's going to be a price floor that the research, materials supply and competition simply won't break through. Guessing when that is going to happen is more luck than anything. I do not see it continuing to get exponentially cheaper for long, though. LiFePho has removed most of the precious metals out of the equation, and the demand for electric cars will continue to compete against the growth in demand for battery storage for renewables. For an analogy, lumber prices have shot through the roof over the past few years where I live due to construction booms. Nothing about the technology has changed, and supply hasn't fluctuated greatly. These same pressures are going to be pushing against lithium batteries getting exponentially cheaper over the next few years. I don't doubt that they will find room to bring prices down, but there is a floor out there somewhere close by.
blake1 6y ago
It actually is pretty easy to predict the future, in a limited sense, a decade out: the cost of new technologies generally follows an exponentially declining “learning” curve. This has been extremely well studied in the case of the aircraft and semiconductor industries, and is the subject of hundreds/thousands of articles. Of course, some conditions must be met—hence the articles—like plentiful inputs, lack of monopoly power, and government (dis)incentives, but we’d see evidence of these constraining battery tech by now. The input variable is the cumulative number of units, and of course we can’t be exact about the trajectory of that number, but we can infer from X MWh manufactured -> $/MWh.
clairity 6y ago
that's a learning or experience curve[0]. related is the technology adoption curve, characterized for instance by the bass diffusion model[1]. [0]: https://en.wikipedia.org/wiki/Experience_curve_effects https://en.wikipedia.org/wiki/Experience_curve_effects [1]: https://en.wikipedia.org/wiki/Bass_diffusion_model https://en.wikipedia.org/wiki/Bass_diffusion_model

ip26 6y ago

drawing a line on a graph and saying "look how X Y is gonna be in Z years" is the same exact type of stupid And yet we do it every ~18 months with semiconductors. Imperfect to be sure, but if you don’t skate to where the puck is headed, you will miss opportunity.

dmoy 6y ago
> And yet we do it every ~18 months with semiconductors. So, this is about to stop. Very soon. Quantum tunneling, yield rates, etc are all starting to be cost prohibitive. <10nm stuff was already delayed. There's plans for 3nm. It's unclear yet whether 2nm will work. Even if hypothetically it does, there's a very real limit - an atom is only 10x smaller than 2nm.
- ben_w 6y ago
  On the one hand, I remember people saying the same about nodes much larger than those we currently use. On the other, yes: continue current trends and single atom transistors become standard in 10-32 years.
  - usrusr 6y ago
    I think it boils down to we will hit a wall, but we don't know exactly when we'll hit it. (and how hard, chances are that higher-hanging-fruit refinements will make the transition to stagnancy so gradual that we may not notice at all) A jump from "all past predictions failed" to "and so will all future predictions" seems rather bold to me. In the end it's like a somewhat upended variation of the "x decades to practical fusion" thing where we all hope that the old joke that x might be a natural constant is eventually proven wrong.
    - ghaff 6y ago
      Well, the main direction of refinements at this point seem to be around composable/heterogeneous computing where we basically have a lot of hardware optimized for specific workloads and throw the complexity at the software people. i.e. now deal with GPUs, DPUs, FPGAs, xCPUs, etc. instead of (largely) just a standardized set of CPU instructions.
- ip26 6y ago
  Sure. But there’s been questions about the viability of the next node for at least twenty years. If you played it safe and stuck with the current node, you’d have been wrong & at a process disadvantage 14 times out of 14. No improvement trend goes on forever. But why is this the moment lithium ion hits the wall? It’s like trying to call the end of a bull market.
- mlindner 6y ago
  > So, this is about to stop. Very soon. Quantum tunneling, yield rates, etc are all starting to be cost prohibitive. People have been saying that for over 10 years though. "Very soon" keeps being punted off by another 5 years every 5 years.
  - vkou 6y ago
    It has stopped, though. Single-threaded performance, either in absolute value, or per dollar has not been following Moore's law for more than a decade. And outside of data centers, single-thread performance is still king. Performance per watt has improved, but that's not a metric the typical end user cares much about.
    - jononor 6y ago
      Performance per watt is key for all portable devices, ie the laptops and phones that many do almost all of their end-user computing on.
    - mlindner 6y ago
      Single-thread performance is not king though. Software nowadays is extensively multi-threaded. If you're still writing single threaded code for a desktop program people are going to look at you like you have antenna growing out of your head. And performance per watt absolutely is important. People care a lot about battery life of their laptops and phones and that's strongly dependent on performance per watt.
- cbozeman 6y ago
  Not according to Jim Keller. He believes we still have plenty of room, and no offense to you, but I'll take his word before yours.
  - ksec 6y ago
    And Jim Keller is not a Fab / Processing guy. And if you actually listen to what he said, he is predicting plenty of room to improve, not it will improve every 24 months, which arguably has already stopped happening since ~2017
- thaumasiotes 6y ago
  > Even if hypothetically it does, there's a very real limit - an atom is only 10x smaller than 2nm. That's certainly a real limit, but what is it a limit on? It is not obvious from first principles that it takes n atoms to implement n transistors.
- fpoling 6y ago
  2nm is just a commercial name. The real features are much bigger, see, for example, https://en.m.wikipedia.org/wiki/10_nm_process https://en.m.wikipedia.org/wiki/10_nm_process. So we are quite far from 10 atoms per transistor. The biggest problem is lithography. ASML managed unexpectedly for many in the industry including Intel to solve technological problems with extreme ultraviolet sources, but shrinking transistors significantly further requires soft x-rays and the perspectives of that now for mass lithography are much more uncertain than for EUV ten years ago.
- imtringued 6y ago
  We'll get to 256 core server chips and that will be the end of the line. You are forgetting that the advertised version number denoted by nm is about what size a theoretical planar transistor would need to have to be equivalent to the advertised process. Since that number is purely theoretical, we can construct a theoretical scenario in which its theoretical nature would become absolutely obvious. Take a 7nm process node transistor and stack it 100 times. Such a process would be called a 0.07nm process.
sir_bearington 6y ago
Except we don't. Where's my single-core 100GHz processor? Improvements still happen, but not always in the same way. If you implemented an application in 2003 assuming we'd have such a processor you'd be very disappointed. Counting on exponential improvements to continue is risky bet.
- rictic 6y ago
  Scaling may take us to some strange places, but it's worth noting that an Apple M1 chip has >100x the transistors of a 2004-era Pentium 4, and achieves ~500x the FLOPs at a fraction of the power draw. Yes this isn't single-threaded performance, but I think we should keep in mind that exponential improvement in price/performance over many decades is possible, if never certain. Sources - https://www.alternatewars.com/BBOW/Computing/Computing_Power.htm https://www.alternatewars.com/BBOW/Computing/Computing_Power... - https://en.wikipedia.org/wiki/Apple_M1 https://en.wikipedia.org/wiki/Apple_M1

clomond 6y ago

Except when you consider where each technology fits within its own S-curve of adoption (X axis over time, Y axis is % of the technology adopted by the market). When factoring in the shape of the exponential decreases in costs, and that penetration of most of these technologies is at or before the inflection point (between 5%-15% market penetration), it is more likely that the cost declines will ACCELERATE moving forward rather than slow down. Why has it felt that laptops and PCs haven’t progressed as much in the 2010s as in the 1990s or 2000s? Because in 1995, there was not a computer on every desk in every home. But now not only is the market saturated with laptops and PCs, people are walking around with mini internet connected “super computers” everywhere they go.

ben_w 6y ago
> Except when you consider where each technology fits within its own S-curve of adoption (X axis over time, Y axis is % of the technology adopted by the market). Unfortunately, even a very small amount of noise in the data makes is basically impossible to know where you are in an S-curve. Much safer to make predictions based on the far more limited good news that PV+battery is already cheaper than coal for electricity or ICE for cars. Hmm… question for anyone who knows: with current tech, how much would it cost to develop a significant PV-powered electrolysis-and-Sabatier-process plant in any of the big coastal deserts, for exporting methane?
- clomond 6y ago
  > Unfortunately, even a very small amount of noise in the data makes is basically impossible to know where you are in an S-curve. While true, my point is that when combined with the fact that we are pre-inflection point, and the economics now stand on their own (renewables, Electric Vehcile TCO and various Energy Storage applications being already cheapest, competitive or very close too) it is not unreasonable when mapping out the 5-15 year future to bet on an acceleration of cost declines over a deceleration. Particularly because the actual driver of unit cost declines (Wrights Law/Moore's Law) is the doubling / magnitude of units manufactured and put through the system, for which with all the factories being ramped up and planned - point to the positive in my view on it. Regarding your PV-powered and electrolysis-Sabatier (electrofuel) methane, I think there are two important considerations. In order for methane (or other e-fuels like hydrogen or longer chain hydrocarbons) to be made economically, the capital cost of the equipment needs to be utilized as close to 100% of the time as possible. We already know that PV excess will be centered around the daytime peak (5-7 hours per day) meaning that there would also need to be plenty of excess wind to balance this out to get anywhere close to 100% utilization of the excess energy. Until the electricity grids get sufficiently saturated with renewables broadly, most e-fuel applications will continue to not be competitive, particularly as things like energy storage applications (possibly run off an e-fuel) are likely to be economical prior there being an opportunity for the export of excess e-fuels. That's more at the a end of the S-Curve as far as I can tell.

nardi 6y ago

These are not the same kind of stupid. One makes the assumption that costs will always be the same, and the other makes the assumption that cost decreases are linear, or predictable. The former is much stupider.

ekianjo 6y ago
Nope, it's the same thing: excess of confidence in predicting the future.
- LMYahooTFY 6y ago
  The commenter was point out the nuance between the two, it's obviously about confidence in an assertion. You just re-reduced it to what was already obvious?
- Qwertious 6y ago
  Investing in government bonds vs investing in penny stocks: both are the same thing, as they are both an excess of confidence in predicting the future.
wpietri 6y ago
They are exactly the same kind of error in that both assume stability over time. One assume prices are stable; the other assumes the rate of change is stable.
- Brakenshire 6y ago
  The rate of change has been stable over the last 15 years, do you bet it continues for a 16th year or that it stops this year?
  - zdragnar 6y ago
    They'll never be free. They'll never be cheaper than the raw materials that go into them, or the copper to wire them together and into the car. At some point, there's going to be a price floor that the research, materials supply and competition simply won't break through. Guessing when that is going to happen is more luck than anything. I do not see it continuing to get exponentially cheaper for long, though. LiFePho has removed most of the precious metals out of the equation, and the demand for electric cars will continue to compete against the growth in demand for battery storage for renewables. For an analogy, lumber prices have shot through the roof over the past few years where I live due to construction booms. Nothing about the technology has changed, and supply hasn't fluctuated greatly. These same pressures are going to be pushing against lithium batteries getting exponentially cheaper over the next few years. I don't doubt that they will find room to bring prices down, but there is a floor out there somewhere close by.

blake1 6y ago

It actually is pretty easy to predict the future, in a limited sense, a decade out: the cost of new technologies generally follows an exponentially declining “learning” curve. This has been extremely well studied in the case of the aircraft and semiconductor industries, and is the subject of hundreds/thousands of articles. Of course, some conditions must be met—hence the articles—like plentiful inputs, lack of monopoly power, and government (dis)incentives, but we’d see evidence of these constraining battery tech by now. The input variable is the cumulative number of units, and of course we can’t be exact about the trajectory of that number, but we can infer from X MWh manufactured -> $/MWh.

clairity 6y ago

that's a learning or experience curve[0]. related is the technology adoption curve, characterized for instance by the bass diffusion model[1]. [0]: https://en.wikipedia.org/wiki/Experience_curve_effects https://en.wikipedia.org/wiki/Experience_curve_effects [1]: https://en.wikipedia.org/wiki/Bass_diffusion_model https://en.wikipedia.org/wiki/Bass_diffusion_model