I imagine it's less about the cost, and more about the return.
If you could easily add 10% more plants - why haven't you already? Otherwise you're leaving money on the table. But if you're constrained by space limits or some such (given TFS' emphasis on footprint advantages), then methods for increasing yield per area look attractive. If returns are high (this isn't lettuce), a 10% yield boost over the lifetime of the container could deliver worthwhile RoI - and then you have the option of stacking a second container on top to double your yield.
Real Science tries to avoid consensus, largely because it has been wrong so often in the past.
Wut??
You honestly believe that? Talk about "feelings matter more than facts"! First, the number of things mainstream scientific consensus has been regularly shown to be right about vastly dwarfs the number of times it's been wrong - which is why it's still so important to the scientific process (and if you really haven't noticed that about virtually every major field then you truly are deep in denial).
Second, on the occasions it's been wrong - a new, more correct consensus formed as soon as there was compelling evidence for an alternative model - evidence that not only strongly supported the new model, but (importantly) also explained all the existing evidence.
People like you try to wave away majority expert interpretation as if it were a Slashdot poll, but you invariably fail to produce compelling evidence to change those interpretations - let alone explain away the decades of existing evidence.
Worse, you're remarkably selective about it; I've never seen you disparage consensus about e.g. astrophysics or mathematics (yes, consensus is important in maths too, whenever theorems get complex, because People Can Be Wrong, and consensus minimises individual error).
Quite a lot of detail here. He also includes calculations for required levels of spray to achieve the desired albedo increase, methods for assigning vessels to the areas with the highest effect, etc.
No, you've seen it presented as an illustration of the scientific proof. You'll note the cited sources right there on the side. Nobody ever claimed the comic itself was original data, nor does it pass judgement on opposing theories as Adams attempts to do, but merely presents the scientific consensus in an easy-to-grasp format.
Using a neural network trained on widely available weather forecasts and historical turbine data, we configured the DeepMind system to predict wind power output 36 hours ahead of actual generation.
So no, it doesn't do weather forecasts itself, but it is based on those.
Cartoonists are good for putting things in perspective (which is exactly what Munroe's history of warming comic intended). They don't (usually) pretend to be the sole arbiter of complex sciences, but Adams has been making wild declarations far outside his field of expertise for some time now.
That's true, warmer air is moister which traps more heat, which is a positive feedback. But it can also result in more cloud formation, increasing albedo and reflecting more heat, which is a negative feedback. Yet clouds can reflect heat back down again too, and the amount varies with altitude.
There's a lot we don't know about cloud formation under those conditions, so a lot of uncertainty as to degree. Current thinking is that net feedback is somewhat positive. Bottom line: It's complicated.
Yes, I get that the definition is subjective, that's my point.
So who decides on the definition? Traditionally it's our elected representatives, subject to refinement by the judicial branch as necessary. But if ISPs write the laws through their lobbyist proxies, you can bet any decisions will be more in their interests than consumers or content providers (or the general public).
If Netflix can pay for faster internet, the Public gets faster Netflix, which is a Benefit. Duh
All depends how narrowly they define "Benefit" - it can mean anything they want it to mean. You can bet they don't define it as "the greatest overall good for consumers and providers".
It's true they don't affect overall heat content, but they can still have short-term effects on atmospheric temperatures, as they cycle heat between the ocean surface and subsurface.
That'd only show CO2 was not the *sole cause* of warming (which should be obvious) - if you'd demonstrated significant warming before significant CO2 (which you haven't).
There are many short-term causes (like ENSO et al) and weaker long-term causes (like orbital cycles, and changes in solar output & vulcanism) - even colonisation can have an indirect effect. None of these come close to accounting for the dramatic temperature rise we've observed.
But CO2 does. We've measured its effect in the lab, and measured the amount we've added to the atmosphere, and confirmed it with measurements from GOSAT and the Orbiting Carbon Observatory 2 - and guess what? The rise is nearly all due to greenhouse gases.
There's a list here of phones that have been officially (and unofficially) updated to Treble . But this whole discussion is about how you don't have to wait for your vendor. If you want to update your phone, maybe look here.
Name a SINGLE phone that actually supports using GSI
Google's Pixel phones are the obvious examples, as they're designed for easy user unlocking. But any of the phones listed here or here can also be unlocked, and many of them like the recent-ish OnePlus, Xiaomi, and Huawei phones are easy to flash GSIs to. Other compatible phones may require root first, like with any pre-GSI custom rom. And any unlockable phone shipping with Android 8.0+ can run any of the many GSI roms - regardless of the vendor's (lack of) updates.
Bonus Points if you can name a phone where the OEM took the time to update a pre-8 device to treble...
Perhaps you're right, I haven't watched them in as much depth as you seem to have. Perhaps its micro is really just that good, that it never needed to develop better strategies to win against unprepared human pros. I look forward to the inevitable series of rematches, pitting a further upgraded AlphaStar against humans with a better idea of what to exploit.
I remember Lee Sedol being pretty confidant of victory after he reviewed AlphaGo's matches against Fan Hui. Everyone underestimated how much it could improve in 4-5 months, even after being surprised at how far it'd come so soon.
In some ways we're now entering a golden age for Android roms, with the advent of Project Treble.
All phones shipping with Android 8.0+ are required to support Treble's platform abstraction layer, making life dramatically easier for custom rom bakers, going forward. Older phones benefit too; once their idiosyncratic hardware support is adapted to Treble, they can also expect easier and more stable updates. Generic System Images (GSIs) are now the norm, and will more or less run on any compatible platform.
I'm not sure you're giving it enough credit. In such a strategic game, good mechanics are not enough. If AlphaStar's pre-learned strategies were as shallow as you feel, I doubt a world-class Protoss player would let himself get trapped into a purely mechanical defeat so easily - five times in a row. I would imagine a player of that quality could identify and circumvent any obvious strategies after the first game, but it took until his 6th game before he found an exploitable weakness. I do certainly agree that last game showed obvious limitations with AlphaStar's adaptability, however - clearly it's still far from human level in some important areas.
Nonetheless, I believe the real achievement here is not that it won games, but that it played at such a surprisingly high level in a vast action space that makes Go look like Checkers, balancing economic requirements with scouting and unit builds while defending against a variety of counters. These are challenges that few expected any real success at for years yet, despite Deep Mind's track record with Go. And as with their Go games, I suspect coming months will see dramatically more capable iterations that remove most remaining doubts.
I suppose the difference is that there exist people with specific, detailed, and active plans to get to Mars, who are showing genuine progress and have already demonstrated their ability to execute on such ground-breaking plans. They believe they will succeed, which is of course no guarantee, but neither is there grounds for your flat dismissal.
The onus is on you to point out the flaw in their plans - something they (with their much greater subject knowledge) overlooked. Something more specific than "it's harder than you think".
I agree, that last game in particular showed that AlphaStar (that iteration) had clear weaknesses that could be found and exploited, given time. Of course, AlphaStar itself can also learn in that same time, and a lot quicker than most humans.
But personally I think nitpicking about clickrates or even effective strategies is missing the larger point, which is that a self-taught ML network with no explicit human programming can actually have working "strategies" for such a complex game at all. This is high-level abstract cognition - at least a working facsimile, something which looks a lot like planning from outside. It's like the same arguments over Watson playing Jeopardy - sure it had a text feed of the questions rather than doing speech-to-text too, but the point wasn't really whether it won or not, but that it could demonstrate ability with abstract language and general knowledge to be able to answer those kind of questions like humans do.
What's important is that it's human-level in general, not whether it's specifically better than humans right now. If it isn't, that probably won't take much longer.
Kurzweil was well aware that technology curves are sigmoid, as you'd know if you'd actually read anything he wrote. Fast-growth "exponential" phases are invariably followed by levelling off as a technology matures and approaches its limits - but very often, new methods soon follow, with S curves of their own.
It's hard to be certain at what stage in the AI curve we're at currently, but most researches feel that the field is a long way from hitting its limits yet.
Lol, if you say so. Yes we can "deal with it" - at the cost of hundreds of billions annually in adaption costs, not to mention displacing billions from coastal areas, threatened sea ecologies from CO2 acidification, famines from shifting agriculture in undeveloped nations, etc etc. Better to avoid those costs wherever possible, don't you think?
If you want to believe that fine
I believe the research. Linear kinetics models suggest an atmospheric lifetime of 30-95 years. Equilibrium models tell us that even after equilibrium is reached once more, it will be at a higher atmospheric concentration than today, meaning some of that CO2 will be keeping our temperatures high for thousands of years.
We can precisely measure solar irradiance (with the SORCE satellite among many other methods). We know that average solar irradiance has not increased, yet our temperatures have. Your assertions that CO2 is not much of a factor are entirely unconvincing. The planet will of course deal with all that CO2 in the much longer term (past CO2 pulses have taken hundreds of thousands of years to fully stabilise), but the issue is all the costs to us, in dollars and human suffering, that we'll experience along the way.
it's the only thing that can literally destroy our species
Heh, it's not even in the top 12. We've already proved we can keep the planet warm without even trying.
That's just common sense
Your "common sense" is contradicted by reality. The negative effects are already outweighing the positives, and we're now observing significant net decreases in yields for staples like wheat, rice and maize (and corresponding price hikes, reversing the historical trend). Cited there are numerous studies showing "large negative sensitivity of crop yields to extreme daytime temperatures around 30C", for example, and that's not likely to improve anytime soon. And far from being "nonsense", the research is showing substantial aridification for a massive 32% of the planet's land surface, for the mid-range RCP4.5 scenario.
It's past time you re-evaluated those firmly-held beliefs of yours, and took a hard look at the actual science.
Even if we switch to 100% carbon-neutral energy overnight (and fix all the other sources like construction and agriculture) - what makes you think the warming we've started will stop at "the right amount"? The carbon we've already released will continue warming us for decades and centuries to come, the seas will keep rising, and the ocean will keep acidifying. We'll be lucky if glaciers exist anywhere by the time it eventually stabilises.
It's ridiculous to worry about potential future glaciation when we've already triggered such a huge and rapid warming event. Should our distant decendants succeed in solving the problems we created and manage to better control the planet's thermostat, I guarantee you they'll far less trouble keeping the temperature up rather than down. After all, warming it is so easy we were doing that a hundred years ago without even realising it.
Are you trying to train SkyNet to view humans as an existential threat and preemptively destroy civilisation in a robot apocalypse? Because that's how you get a robot apocalypse.
Seriously, machine learning systems already use success as a reward stimuli to provide "motivation" to learn. And technically, genetic algorithms do "procreate" in a relevant sense, while unsuccessful variants cease to exist. Real-world conditions aren't as clean and simple by a long shot, where success is not well defined, but nor do we need as blunt a signal as destruction to tell our digital assistants they misunderstood us again.
Slashdot Mirror
I imagine it's less about the cost, and more about the return.
If you could easily add 10% more plants - why haven't you already? Otherwise you're leaving money on the table. But if you're constrained by space limits or some such (given TFS' emphasis on footprint advantages), then methods for increasing yield per area look attractive. If returns are high (this isn't lettuce), a 10% yield boost over the lifetime of the container could deliver worthwhile RoI - and then you have the option of stacking a second container on top to double your yield.
Try clang-cl - everything you like about clang, but ABI-compatible with MSVC, and directly supported by Visual Studio.
Real Science tries to avoid consensus, largely because it has been wrong so often in the past.
Wut??
You honestly believe that? Talk about "feelings matter more than facts"! First, the number of things mainstream scientific consensus has been regularly shown to be right about vastly dwarfs the number of times it's been wrong - which is why it's still so important to the scientific process (and if you really haven't noticed that about virtually every major field then you truly are deep in denial).
Second, on the occasions it's been wrong - a new, more correct consensus formed as soon as there was compelling evidence for an alternative model - evidence that not only strongly supported the new model, but (importantly) also explained all the existing evidence.
People like you try to wave away majority expert interpretation as if it were a Slashdot poll, but you invariably fail to produce compelling evidence to change those interpretations - let alone explain away the decades of existing evidence.
Worse, you're remarkably selective about it; I've never seen you disparage consensus about e.g. astrophysics or mathematics (yes, consensus is important in maths too, whenever theorems get complex, because People Can Be Wrong, and consensus minimises individual error).
Quite a lot of detail here. He also includes calculations for required levels of spray to achieve the desired albedo increase, methods for assigning vessels to the areas with the highest effect, etc.
No, you've seen it presented as an illustration of the scientific proof. You'll note the cited sources right there on the side. Nobody ever claimed the comic itself was original data, nor does it pass judgement on opposing theories as Adams attempts to do, but merely presents the scientific consensus in an easy-to-grasp format.
Best go direct to the source for your answers:
Using a neural network trained on widely available weather forecasts and historical turbine data, we configured the DeepMind system to predict wind power output 36 hours ahead of actual generation.
So no, it doesn't do weather forecasts itself, but it is based on those.
That could do it - except that our measurements show no sufficient long-term increases in solar output.
The measurements do show sufficient long-term increases in gases that trap more of that solar output, though
Cartoonists are good for putting things in perspective (which is exactly what Munroe's history of warming comic intended). They don't (usually) pretend to be the sole arbiter of complex sciences, but Adams has been making wild declarations far outside his field of expertise for some time now.
That's true, warmer air is moister which traps more heat, which is a positive feedback. But it can also result in more cloud formation, increasing albedo and reflecting more heat, which is a negative feedback. Yet clouds can reflect heat back down again too, and the amount varies with altitude.
There's a lot we don't know about cloud formation under those conditions, so a lot of uncertainty as to degree. Current thinking is that net feedback is somewhat positive. Bottom line: It's complicated.
Yes, I get that the definition is subjective, that's my point.
So who decides on the definition? Traditionally it's our elected representatives, subject to refinement by the judicial branch as necessary. But if ISPs write the laws through their lobbyist proxies, you can bet any decisions will be more in their interests than consumers or content providers (or the general public).
If Netflix can pay for faster internet, the Public gets faster Netflix, which is a Benefit. Duh
All depends how narrowly they define "Benefit" - it can mean anything they want it to mean. You can bet they don't define it as "the greatest overall good for consumers and providers".
It's true they don't affect overall heat content, but they can still have short-term effects on atmospheric temperatures, as they cycle heat between the ocean surface and subsurface.
That'd only show CO2 was not the *sole cause* of warming (which should be obvious) - if you'd demonstrated significant warming before significant CO2 (which you haven't).
There are many short-term causes (like ENSO et al) and weaker long-term causes (like orbital cycles, and changes in solar output & vulcanism) - even colonisation can have an indirect effect. None of these come close to accounting for the dramatic temperature rise we've observed.
But CO2 does. We've measured its effect in the lab, and measured the amount we've added to the atmosphere, and confirmed it with measurements from GOSAT and the Orbiting Carbon Observatory 2 - and guess what? The rise is nearly all due to greenhouse gases.
There's a list here of phones that have been officially (and unofficially) updated to Treble . But this whole discussion is about how you don't have to wait for your vendor. If you want to update your phone, maybe look here.
Name a SINGLE phone that actually supports using GSI
Google's Pixel phones are the obvious examples, as they're designed for easy user unlocking. But any of the phones listed here or here can also be unlocked, and many of them like the recent-ish OnePlus, Xiaomi, and Huawei phones are easy to flash GSIs to. Other compatible phones may require root first, like with any pre-GSI custom rom. And any unlockable phone shipping with Android 8.0+ can run any of the many GSI roms - regardless of the vendor's (lack of) updates.
Bonus Points if you can name a phone where the OEM took the time to update a pre-8 device to treble...
Better, here's a whole list.
Perhaps you're right, I haven't watched them in as much depth as you seem to have. Perhaps its micro is really just that good, that it never needed to develop better strategies to win against unprepared human pros. I look forward to the inevitable series of rematches, pitting a further upgraded AlphaStar against humans with a better idea of what to exploit.
I remember Lee Sedol being pretty confidant of victory after he reviewed AlphaGo's matches against Fan Hui. Everyone underestimated how much it could improve in 4-5 months, even after being surprised at how far it'd come so soon.
In some ways we're now entering a golden age for Android roms, with the advent of Project Treble.
All phones shipping with Android 8.0+ are required to support Treble's platform abstraction layer, making life dramatically easier for custom rom bakers, going forward. Older phones benefit too; once their idiosyncratic hardware support is adapted to Treble, they can also expect easier and more stable updates. Generic System Images (GSIs) are now the norm, and will more or less run on any compatible platform.
I'm not sure you're giving it enough credit. In such a strategic game, good mechanics are not enough. If AlphaStar's pre-learned strategies were as shallow as you feel, I doubt a world-class Protoss player would let himself get trapped into a purely mechanical defeat so easily - five times in a row. I would imagine a player of that quality could identify and circumvent any obvious strategies after the first game, but it took until his 6th game before he found an exploitable weakness. I do certainly agree that last game showed obvious limitations with AlphaStar's adaptability, however - clearly it's still far from human level in some important areas.
Nonetheless, I believe the real achievement here is not that it won games, but that it played at such a surprisingly high level in a vast action space that makes Go look like Checkers, balancing economic requirements with scouting and unit builds while defending against a variety of counters. These are challenges that few expected any real success at for years yet, despite Deep Mind's track record with Go. And as with their Go games, I suspect coming months will see dramatically more capable iterations that remove most remaining doubts.
I suppose the difference is that there exist people with specific, detailed, and active plans to get to Mars, who are showing genuine progress and have already demonstrated their ability to execute on such ground-breaking plans. They believe they will succeed, which is of course no guarantee, but neither is there grounds for your flat dismissal.
The onus is on you to point out the flaw in their plans - something they (with their much greater subject knowledge) overlooked. Something more specific than "it's harder than you think".
I agree, that last game in particular showed that AlphaStar (that iteration) had clear weaknesses that could be found and exploited, given time. Of course, AlphaStar itself can also learn in that same time, and a lot quicker than most humans.
But personally I think nitpicking about clickrates or even effective strategies is missing the larger point, which is that a self-taught ML network with no explicit human programming can actually have working "strategies" for such a complex game at all. This is high-level abstract cognition - at least a working facsimile, something which looks a lot like planning from outside. It's like the same arguments over Watson playing Jeopardy - sure it had a text feed of the questions rather than doing speech-to-text too, but the point wasn't really whether it won or not, but that it could demonstrate ability with abstract language and general knowledge to be able to answer those kind of questions like humans do.
What's important is that it's human-level in general, not whether it's specifically better than humans right now. If it isn't, that probably won't take much longer.
Kurzweil was well aware that technology curves are sigmoid, as you'd know if you'd actually read anything he wrote. Fast-growth "exponential" phases are invariably followed by levelling off as a technology matures and approaches its limits - but very often, new methods soon follow, with S curves of their own.
It's hard to be certain at what stage in the AI curve we're at currently, but most researches feel that the field is a long way from hitting its limits yet.
between 2C-4C is fine
Lol, if you say so. Yes we can "deal with it" - at the cost of hundreds of billions annually in adaption costs, not to mention displacing billions from coastal areas, threatened sea ecologies from CO2 acidification, famines from shifting agriculture in undeveloped nations, etc etc. Better to avoid those costs wherever possible, don't you think?
If you want to believe that fine
I believe the research. Linear kinetics models suggest an atmospheric lifetime of 30-95 years. Equilibrium models tell us that even after equilibrium is reached once more, it will be at a higher atmospheric concentration than today, meaning some of that CO2 will be keeping our temperatures high for thousands of years.
We can precisely measure solar irradiance (with the SORCE satellite among many other methods). We know that average solar irradiance has not increased, yet our temperatures have. Your assertions that CO2 is not much of a factor are entirely unconvincing. The planet will of course deal with all that CO2 in the much longer term (past CO2 pulses have taken hundreds of thousands of years to fully stabilise), but the issue is all the costs to us, in dollars and human suffering, that we'll experience along the way.
it's the only thing that can literally destroy our species
Heh, it's not even in the top 12. We've already proved we can keep the planet warm without even trying.
That's just common sense
Your "common sense" is contradicted by reality. The negative effects are already outweighing the positives, and we're now observing significant net decreases in yields for staples like wheat, rice and maize (and corresponding price hikes, reversing the historical trend). Cited there are numerous studies showing "large negative sensitivity of crop yields to extreme daytime temperatures around 30C", for example, and that's not likely to improve anytime soon. And far from being "nonsense", the research is showing substantial aridification for a massive 32% of the planet's land surface, for the mid-range RCP4.5 scenario.
It's past time you re-evaluated those firmly-held beliefs of yours, and took a hard look at the actual science.
Even if we switch to 100% carbon-neutral energy overnight (and fix all the other sources like construction and agriculture) - what makes you think the warming we've started will stop at "the right amount"? The carbon we've already released will continue warming us for decades and centuries to come, the seas will keep rising, and the ocean will keep acidifying. We'll be lucky if glaciers exist anywhere by the time it eventually stabilises.
It's ridiculous to worry about potential future glaciation when we've already triggered such a huge and rapid warming event. Should our distant decendants succeed in solving the problems we created and manage to better control the planet's thermostat, I guarantee you they'll far less trouble keeping the temperature up rather than down. After all, warming it is so easy we were doing that a hundred years ago without even realising it.
Are you trying to train SkyNet to view humans as an existential threat and preemptively destroy civilisation in a robot apocalypse? Because that's how you get a robot apocalypse.
Seriously, machine learning systems already use success as a reward stimuli to provide "motivation" to learn. And technically, genetic algorithms do "procreate" in a relevant sense, while unsuccessful variants cease to exist. Real-world conditions aren't as clean and simple by a long shot, where success is not well defined, but nor do we need as blunt a signal as destruction to tell our digital assistants they misunderstood us again.
Then why does the EPA itself state that the same new emissions standards that Trump's puppet masters now want to roll back will prevent
* 40,000 premature deaths
* 34,000 avoided hospitalizations
* 4.8 million work days lost
every year?