Tesla unveils new Semi with a 500 mile range, Roadster that can hit 250 miles per hour

123457»

Comments

  • Reply 121 of 140

    Soli said:
    PS: I don't think it was mentioned, but I assume that Tesla Semi has an onboard air compressor for the trailer's air brakes. The only other option is to only use Tesla trailers, which I'm not even sure would be legal.


    He said the tractor has side flaps that match to the trailer being pulled whether it's an old or new trailer.

    But I don't think he mentioned whether the trailer would have regen motors itself. 



  • Reply 122 of 140
    tzeshan said:
    This is the main reason Musk is introducing two new models to recapture the heavily government subsidized tax credit and adding on to national debt.
    It'd be dumb if he was doing that. The main credit is about to run out for Tesla, he'd be better focusing on 1 car.

    Musk has said several times he'd like all incentives to be removed for gasoline, diesel, and electric - and let competition go for it. You guys in the US get great subsidies on your gasoline at various points in the chain.
    cgWerks
  • Reply 123 of 140
    cgWerks said:
    k2kw said:
    I can't wait for self driving cars to take away the tedium and tension of my rush hour every day.    And of course make it safer.
    What makes you think it's going to be safer... and safer than what? If you're drunk or texting, then of course it's going to be safer. 
    To me it's going to be safer than those who dart in and out of traffic on the local highways, driving so close behind other cars to try to push them to go faster. And the old people who insist on having a car. And instead of giving provisional licenses to people who were caught dangerous or drink driving who have to get to work, they should lose their license. 

    Unfortunately it's the hoons darting in and out of traffic who make a sport of it, and enjoy it, who are both dangerous and the last ones to want AI drivers.  So my car with AI will need to be ready to cope with them.

    Perhaps if as a society we suspend licenses more liberally we can force repeat bad drivers to switch to autonomous systems. If you replace the worst drivers with AI then an AI with double the driving competence will have a disproportionate benefit. 
  • Reply 124 of 140
    timmillea said:
    Testorone-driven marketing of this kind looks very last century in retrospect.
    Heaven forbid half the population of the planet be catered to.
  • Reply 125 of 140
    lukei said:
    flydog said:
    How is this article relevant to Apple products?
    Because Elon Musk will Be Apple CEO after they buy his businesses?
    Never happen. Musk previously said his plan was to step down from Tesla as CEO once their Model 3 was shipping in good numbers and that SpaceX and Mars colonization would be his focus for the rest of his life.
    He said that originally. But more recently he said he's stay with Tesla for as long as they'd have him.

    However, he's shifting Tesla too to move the world to renewable energies. It's making solar panels and cells. Incredibly strong glass. Massive factories that can process from raw materials to final products. Artificial intelligence and automated production systems. Tunnel boring machines. 

    All of these things are geared to both work on Earth, and I believe also will be good for Mars. SpaceX will focus on getting people and cargo to Mars, while Tesla is involved in other aspects. 

    Anachr0n said:
    Aw, c’mon. There nothing like the feeling of a fine leather buggy whip in your hands. And the instant acceleration of a team of horses. ;-)
    I know how to measure horsepower if you have 2 horses. But how do you measure torque?

  • Reply 126 of 140
    cgWerkscgWerks Posts: 2,952member
    gregalexander said:
    To me it's going to be safer than those who dart in and out of traffic on the local highways, driving so close behind other cars to try to push them to go faster. And the old people who insist on having a car. And instead of giving provisional licenses to people who were caught dangerous or drink driving who have to get to work, they should lose their license. 

    Unfortunately it's the hoons darting in and out of traffic who make a sport of it, and enjoy it, who are both dangerous and the last ones to want AI drivers.  So my car with AI will need to be ready to cope with them.

    Perhaps if as a society we suspend licenses more liberally we can force repeat bad drivers to switch to autonomous systems. If you replace the worst drivers with AI then an AI with double the driving competence will have a disproportionate benefit. 
    Oh, ***SO*** much could be done to make the roads a safer place which has nothing to do with AI. Driving education and exams are a joke. People keep their licenses after doing things that should come with huge penalties. Basic common sense seems to have gone out the window (texting, drinking, etc.).

    But, AI vehicles don't have any driving competence... they are just following an elaborate program. They'll be as good as the DB data-set and programming competence of their designers (and given the state of the rest of technology and/or AI-attempts, that isn't necessarily saying much). It has been way over-hyped and over-sold, and unfortunately our stupid politicians are starting to buy into it.
    tallest skil
  • Reply 127 of 140
    nhtnht Posts: 4,522member
    cgWerks said:
    gregalexander said:
    To me it's going to be safer than those who dart in and out of traffic on the local highways, driving so close behind other cars to try to push them to go faster. And the old people who insist on having a car. And instead of giving provisional licenses to people who were caught dangerous or drink driving who have to get to work, they should lose their license. 

    Unfortunately it's the hoons darting in and out of traffic who make a sport of it, and enjoy it, who are both dangerous and the last ones to want AI drivers.  So my car with AI will need to be ready to cope with them.

    Perhaps if as a society we suspend licenses more liberally we can force repeat bad drivers to switch to autonomous systems. If you replace the worst drivers with AI then an AI with double the driving competence will have a disproportionate benefit. 
    Oh, ***SO*** much could be done to make the roads a safer place which has nothing to do with AI. Driving education and exams are a joke. People keep their licenses after doing things that should come with huge penalties. Basic common sense seems to have gone out the window (texting, drinking, etc.).

    But, AI vehicles don't have any driving competence... they are just following an elaborate program. They'll be as good as the DB data-set and programming competence of their designers (and given the state of the rest of technology and/or AI-attempts, that isn't necessarily saying much). It has been way over-hyped and over-sold, and unfortunately our stupid politicians are starting to buy into it.
    You have no idea of what you are talking about. While some systems rely on programmed behavior others are built using convolutional neural networks or other types of learning systems and not “databases” and human written rules.  The primary lines of research are now in the areas of deep learning.

    In the nineties I worked on a forward chaining expert system to autonomously control spacecraft like what you are thinking of with programmed rules but the state of the art twenty years later has significantly changed.

    For decades speech recognition was “5 years from widespread use” and widely poo-poo’d by folks like you with a little knowledge until today where we have Siri, Cortana, etc on everyone’s phones. 

    For decades autonomous driving was “5 years from widespread use” and today I believe we really are that close.  We could do almost all of this shit 20 years ago but the processing power wasn’t there for in vehicle systems, the sensors too large or coarse and the learning techniques still too unrefined for mainstream commercial adoption.

    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.  There will still be humans in the loop somewhere in the convoy for safety but that use will be in the more benign highway environment vs the more complex city environment.

    Most people think they are good drivers. Most of them are wrong.  Even the good ones can suffer from driver fatigue and inattention.  Autonomous driving systems will be able to beat the average safety outcome for highway driving and we’ll start to see deployments in trucking where driver fatigue and driver shortages are problems.
    SoliSpamSandwich
  • Reply 128 of 140

    I wonder if it also has regenerative braking, and whether someone could also build trailers with regenerative braking, 18 wheels all working together would have to generate a fair bit of energy back into the batteries and perhaps extend range a little more. I thought also it might have been more of a hybrid as well with a smaller petrol or diesel engine to assist in charging the batteries to extend the range which might have made it more attractive for longer hauls.

    I can see these being deployed more in Europe, where distances are smaller and populations more dense, and trucks travel more frequently through built up suburban areas. Charging would be done on turnaround and there wouldn't be any more grumble of big diesels downshifting and engine braking through residential corridors to annoy people.

  • Reply 129 of 140
    SoliSoli Posts: 10,035member
    markg71 said:

    I wonder if it also has regenerative braking, and whether someone could also build trailers with regenerative braking, 18 wheels all working together would have to generate a fair bit of energy back into the batteries and perhaps extend range a little more. I thought also it might have been more of a hybrid as well with a smaller petrol or diesel engine to assist in charging the batteries to extend the range which might have made it more attractive for longer hauls.

    I can see these being deployed more in Europe, where distances are smaller and populations more dense, and trucks travel more frequently through built up suburban areas. Charging would be done on turnaround and there wouldn't be any more grumble of big diesels downshifting and engine braking through residential corridors to annoy people.

    1) It does.

    2) Tesla has announced no trailers, but it wouldn’t be hard to do. They could even put solar panels on the top to extend the range even more, but these things will require dedicated trailers, which means only a company like Walmart that can have dedicated routes where these can get used in an efficient manor instead of trailers sitting in a yard. This would also mean that the distance warrants the extra cost.

    3) Because current brakes are air brakes the Tesla will have to have an electric air compressor for trailers. That was never mentioned, but it has to exist for this to work.
  • Reply 130 of 140
    nht said:
    For decades speech recognition was “5 years from widespread use”…
    For decades autonomous driving was “5 years from widespread use”…
    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    cgWerks
  • Reply 131 of 140
    cgWerkscgWerks Posts: 2,952member
    nht said:
    You have no idea of what you are talking about. While some systems rely on programmed behavior others are built using convolutional neural networks or other types of learning systems and not “databases” and human written rules.  The primary lines of research are now in the areas of deep learning.
    ...
    Most people think they are good drivers. Most of them are wrong.  Even the good ones can suffer from driver fatigue and inattention.  Autonomous driving systems will be able to beat the average safety outcome for highway driving and we’ll start to see deployments in trucking where driver fatigue and driver shortages are problems.
    It's all programmed behavior. It's more a matter of the capacity to index new information and then be trained on what it means. But 'learning' is anthropomorphizing what is happening. Will it improve due to situational data collection? Yes, as that's what they are doing, it isn't actually thinking and learning. The key thing is that it is artificial and always will be. While I'm no expert on AI, I've talked to people in the field in the past and have done some study (i.e.: written graduate level papers) on *attempts* to explain consciousness, etc.

    It's fascinating stuff, but I think the average person in the public is being a bit hoodwinked about what is actually going on and eventual capabilities. (As seems Musk, as he talks about it more in terms of a Sci-Fi novel or maybe he watched the Terminator a few too many times.)

    Maybe read this:
    https://www.technologyreview.com/s/608911/is-ai-riding-a-one-trick-pony/
    then this:
    https://www.axios.com/ai-pioneer-advocates-starting-over-2485537027.html

    "And most of the big leaps that came about as it developed didn’t involve some new insight about neuroscience; they were technical improvements, reached by years of mathematics and engineering."
    "A real intelligence doesn’t break when you slightly change the requirements of the problem it’s trying to solve."

    re: good drivers - Yes, and for most of our other capabilities, especially in Western cultures. But, over the years, I've done a number of things to actually improve my skills, not just *think* I might be better than the average person. I've done some amateur SCCA racing and autocross, as well as driver training. I also seem to have more of a 'feel' for the dynamics of what is going on (maybe due to the racing) than it seems others I drive with do. I'm also a pretty serious driver in how I pay attention, and 'read' traffic situations and behavior of other drivers.

    While I don't know exactly where that puts me on the spectrum, I'm pretty sure it's above average... and I think I have the evidence to back that up.

    tallest skil said:
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    Too many scientists and futurists haven't had enough philosophy, anthropology, or critical thinking training, it seems. Technological advancement is one thing... impossible/highly-improbable leaps are quite another. And, then, as you've pointed out, wrong predictions or over-estimation and such as well. Technology can't just *do anything* as long as we give it enough time.
    tallest skil
  • Reply 132 of 140
    nhtnht Posts: 4,522member
    nht said:
    For decades speech recognition was “5 years from widespread use”…
    For decades autonomous driving was “5 years from widespread use”…
    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    Except that you can be right for decades and then be wrong.  Past performance is not indicative of future results.  Meaning that simply being a curmudgeon isn't predictive of when certain technologies will mature.

    Autonomous driving systems are at that tipping point.  That's my opinion as a curmudgeon with more than a passing interest in these systems.

    Hydrogen fueled cars are more than 5 years away because the infrastructure doesn't exist and there isn't a Tesla equivalent.  

    Electric vehicles ARE in widespread use. 159,000 EVs were sold in the US during 2016.  When the Nissan Leaf and Volt came out as production vehicles EVs were finally available to the average consumer.
    Soli
  • Reply 133 of 140
    SoliSoli Posts: 10,035member
    nht said:
    nht said:
    For decades speech recognition was “5 years from widespread use”…
    For decades autonomous driving was “5 years from widespread use”…
    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    Except that you can be right for decades and then be wrong.  Past performance is not indicative of future results.  Meaning that simply being a curmudgeon isn't predictive of when certain technologies will mature.

    Autonomous driving systems are at that tipping point.  That's my opinion as a curmudgeon with more than a passing interest in these systems.

    Hydrogen fueled cars are more than 5 years away because the infrastructure doesn't exist and there isn't a Tesla equivalent.  

    Electric vehicles ARE in widespread use. 159,000 EVs were sold in the US during 2016.  When the Nissan Leaf and Volt came out as production vehicles EVs were finally available to the average consumer.
    With the 1 million worldwide milestone happening in 2015 and then the 2 million milestone the next year. China first started outselling the US in 2016.

    edited November 2017
  • Reply 134 of 140
    tmaytmay Posts: 6,309member
    nht said:
    nht said:
    For decades speech recognition was “5 years from widespread use”…
    For decades autonomous driving was “5 years from widespread use”…
    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    Except that you can be right for decades and then be wrong.  Past performance is not indicative of future results.  Meaning that simply being a curmudgeon isn't predictive of when certain technologies will mature.

    Autonomous driving systems are at that tipping point.  That's my opinion as a curmudgeon with more than a passing interest in these systems.

    Hydrogen fueled cars are more than 5 years away because the infrastructure doesn't exist and there isn't a Tesla equivalent.  

    Electric vehicles ARE in widespread use. 159,000 EVs were sold in the US during 2016.  When the Nissan Leaf and Volt came out as production vehicles EVs were finally available to the average consumer.
    Electric Vehicles are not, by any metric, currently in widespread use. If anything, EV's are looking to crack the one percentile mark of total yearly sales, and, I doubt that will even be in 2018. Autonomous driving systems are not at a tipping point for anything but controlled access highways and roads, and most of that would be interstate highway in the U.S, or equivalent worldwide.

    To crack that single percentile, manufacturers would need to deliver close to a million EV's in a single year, in a market expected to sell some 93 M vehicles.
    edited November 2017 tallest skilcgWerks
  • Reply 135 of 140
    nhtnht Posts: 4,522member
    tmay said:
    nht said:
    nht said:
    For decades speech recognition was “5 years from widespread use”…
    For decades autonomous driving was “5 years from widespread use”…
    We’re going to see autonomous vehicles in fleet use (trucking) in that 5 year timeframe.
    Thing is, for decades we’ve also heard “hydrogen fueled cars are 5 years from widespread use.” “Electric vehicles are 5 years from widespread use,” “fusion power plants are 5 years from widespread use,” and “revolutionary battery technology is 5 years from widespread use”, so you’ll have to forgive people for not expecting anything from utterly game-changing technology.
    Except that you can be right for decades and then be wrong.  Past performance is not indicative of future results.  Meaning that simply being a curmudgeon isn't predictive of when certain technologies will mature.

    Autonomous driving systems are at that tipping point.  That's my opinion as a curmudgeon with more than a passing interest in these systems.

    Hydrogen fueled cars are more than 5 years away because the infrastructure doesn't exist and there isn't a Tesla equivalent.  

    Electric vehicles ARE in widespread use. 159,000 EVs were sold in the US during 2016.  When the Nissan Leaf and Volt came out as production vehicles EVs were finally available to the average consumer.
    Electric Vehicles are not, by any metric, currently in widespread use. If anything, EV's are looking to crack the one percentile mark of total yearly sales, and, I doubt that will even be in 2018. Autonomous driving systems are not at a tipping point for anything but controlled access highways and roads, and most of that would be interstate highway in the U.S, or equivalent worldwide.

    To crack that single percentile, manufacturers would need to deliver close to a million EV's in a single year, in a market expected to sell some 93 M vehicles.
    This is like claiming that Mazda cars, with 150,000 new car sales in the US in 2016, are not in widespread use.

    That’s amusing if dumb.

    Not to mention:

    ”US electric car sales had another record month in December and also again reached a record market share in the country. Fully electric and plug-in hybrid electric car sales reached 19,133 in December (and 144,455 sales for 2016 as a whole), which put EV market share among all passenger cars at a record 1.13% (and 0.82% for 2016 as a whole)*.”

    https://cleantechnica.com/2017/01/06/7-models-81-electric-car-sales-us-us-electric-car-sales-report/
    edited November 2017
  • Reply 136 of 140
    nht said:
    Except that you can be right for decades and then be wrong.
    And wrong for decades and then be right.
    Electric vehicles ARE in widespread use.
    You and I differ on the definition of that word, but then again I’m still waiting on the aforementioned revolutionary battery tech to make my definition a reality.
  • Reply 137 of 140
    nhtnht Posts: 4,522member
    nht said:
    Except that you can be right for decades and then be wrong.
    And wrong for decades and then be right.
    Electric vehicles ARE in widespread use.
    You and I differ on the definition of that word, but then again I’m still waiting on the aforementioned revolutionary battery tech to make my definition a reality.
    NiMH, Lithium Ion, Lithium Polymer are revolutionary battery techs unless you think smartphones, quads, EVs, etc are not revolutionary.

    The energy densities are much higher than the batteries used widely before that in consumer items.  Takes decades for tech to mature from military/space use to commercial.  Autonomous vehicle technology is simply another example.
    edited November 2017 Soli
  • Reply 138 of 140
    nhtnht Posts: 4,522member
    cgWerks said:
    nht said:
    You have no idea of what you are talking about. While some systems rely on programmed behavior others are built using convolutional neural networks or other types of learning systems and not “databases” and human written rules.  The primary lines of research are now in the areas of deep learning.
    ...
    Most people think they are good drivers. Most of them are wrong.  Even the good ones can suffer from driver fatigue and inattention.  Autonomous driving systems will be able to beat the average safety outcome for highway driving and we’ll start to see deployments in trucking where driver fatigue and driver shortages are problems.
    It's all programmed behavior. It's more a matter of the capacity to index new information and then be trained on what it means. But 'learning' is anthropomorphizing what is happening. Will it improve due to situational data collection? Yes, as that's what they are doing, it isn't actually thinking and learning. The key thing is that it is artificial and always will be. While I'm no expert on AI, I've talked to people in the field in the past and have done some study (i.e.: written graduate level papers) on *attempts* to explain consciousness, etc.

    It's fascinating stuff, but I think the average person in the public is being a bit hoodwinked about what is actually going on and eventual capabilities. (As seems Musk, as he talks about it more in terms of a Sci-Fi novel or maybe he watched the Terminator a few too many times.)

    Maybe read this:
    https://www.technologyreview.com/s/608911/is-ai-riding-a-one-trick-pony/
    then this:
    https://www.axios.com/ai-pioneer-advocates-starting-over-2485537027.html

    "And most of the big leaps that came about as it developed didn’t involve some new insight about neuroscience; they were technical improvements, reached by years of mathematics and engineering."
    "A real intelligence doesn’t break when you slightly change the requirements of the problem it’s trying to solve."
    First, it's not "programmed behavior" in the traditional sense.  It's called "learning" because you train the system.  Different stimulus will result in different end behaviors using the exact same code.  The end goal of autonomous systems is NOT machine consciousness.  The end goal of autonomous systems is to be able to carry out a complex task with limited human input and supervision.  To equate autonomous systems with machine consciousness is clear evidence of not understanding what you are poo-pooing as impossible.

    The technology review author fundamentally doesn't understand how technology advancement works (or more likely is just click baiting) and this is clearly evident in its opening line:

    "Just about every AI advance you’ve heard of depends on a breakthrough that’s three decades old."

    Yes, that's how it works. Because:

    "But it took another 26 years before increasing computational power made good on the discovery."

    The foundation of every new cool thing today was laid decades ago by someone with a brilliant moment of inspiration.  Then it takes a while before you can actually do anything with it.

    The other fundamental problem is that the author writes as if any single technique (aka deep learning neural networks) must single handedly solve every AI problem or it's a dead end or one trick pony.  Nope.  Almost all the stuff we've done in AI is still used today and add to our capabilities.  The hot items in AI during the early 00s (like artificial neural networks (ANNs), genetic programming (GP), evolutionary polynomial regression (EPR), support vector machines (SVM), etc)  haven't gone away.  They've just matured and are simply another tool in the toolbox.  In 10 years other techniques will be hot and deep learning will just be another useful tool in the toolbox.  Same for all the neuromorphic chips that the cool kids are currently playing with.

    The other thing the author doesn't understand is that we necessarily have to wait another 30 years for the next big advance in AI (which is separate from autonomous systems) and we have to look at current dissertations resetting the clock to today for the next big advancement in AI.  As opposed to things that have been percolating since the 90s.  There is no guarantee of breakthroughs on a certain cadence but every decade or so something hits a tipping point (reduction in cost, increase in power, another piece of a puzzle found) where you see a period of rapid advance and the creation of another tool in the toolbox.

    Hinton's quoting of Planck is correct, but the next advance isn't necessarily a grad student who is deeply suspicious of everything Hinton has said but a grad student that was deeply suspicious of what Koza did with genetic programming in the 90s (or other evolutionary algorithm technique popular at the time).  EA is more or less out of style but still useful for folks solving problems in certain areas.  Or swarming is another good example of an eclipsed AI technique that is useful.  There's no need to throw everything away and start over.  Someone, somewhere is already working on the next thing.

    And the comment that we are doing engineering without the science like building bridge without the physics is somewhat true but ignores the fact that we were able to build lots of very functional bridges long before we understood the underlying math.  And AI techniques are simply tools among many used in autonomous vehicles. I can use several different techniques to address different challenges.
    Soli
  • Reply 139 of 140
    brucemcbrucemc Posts: 1,541member
    For all those who want to believe that the Model 3 is only a delay of a month or two and then everything is going to be purring along smoothly.  Tesla has never met their production targets on an annual basis.  They have never "over delivered".  Now they plan on going from producing less than 100K vehicles/year to over 500K/year overnight (their production goals - not my view).  Sorry, don't believe it.  Tesla doesn't have the track record to back it up.

    This isn't about hating Tesla.  I wish Tesla all the best.  But their sky-high valuation as a company is not backed up by any proof of execution, nor of any reasonable expectations of outsized profit (even if they *could* meet their production targets AND make a small profit per car, then hey - they are a small slightly profitable car manufacturer).  Unless you think that Tesla will "corner the market" on autonomous driving technology, then their company valuation is ludicrous.
  • Reply 140 of 140
    cgWerkscgWerks Posts: 2,952member
    nht said:
    NiMH, Lithium Ion, Lithium Polymer are revolutionary battery techs unless you think smartphones, quads, EVs, etc are not revolutionary.
    Relatively big advances, but not what is needed to really hurdle the problems. The most promising tech I saw was around aluminum batteries... but I'm not sure where it is at or how long. They were talking 1/2 charge times and more than 2x energy density. Plus, they didn't use any hard to get or dangerous materials, if memory serves.

    nht said:
    The end goal of autonomous systems is NOT machine consciousness.  The end goal of autonomous systems is to be able to carry out a complex task with limited human input and supervision.  To equate autonomous systems with machine consciousness is clear evidence of not understanding what you are poo-pooing as impossible.
    Thanks for the response, and you certainly know more about the various technologies. But, the problem for me lies in the above. You're right that what they are shooting for (despite Elon's stated fears and such) isn't really machine consciousness (which, IMO, is impossible). I think the general public thinks it is... and I think that's what we'd need to have a truly excellent automated vehicle. What it seems we're shooting for, is something that is kind of good-enough once it has collected enough data and been 'trained' well enough.

    In other words, a statistical tipping point where it's better, on average, than human drivers. I suppose we'll get there with enough sensor improvement, data collection, and training... and when anomaly situations show up, that's where the deaths will occur (possibly instead of the texting person who hopefully is now using AI). That's not good enough for me. I'm not just an average or statistic. I'm not willing to trade my life because AI got confused by a snow-storm or a semi it couldn't 'see' crossing the road because it does a better job than a drunk or someone texting.

    My solution would be to do far more to stop drunks and texting... increase driving training and responsibility... AND implement some of that awesome sensor and AI tech to ***supplement*** human drivers abilities and awareness. THAT I think would lower fatality rates far more than starting to depend on AI for fully autonomous vehicles. The problem is that people are already sold on this idea of just having vehicles shuttle us around, and companies are heavily invested in making that happen.

    brucemc said:
    This isn't about hating Tesla.  I wish Tesla all the best.  But their sky-high valuation as a company is not backed up by any proof of execution, nor of any reasonable expectations of outsized profit ...
    I somewhat agree. I think this is a money raising and marketing move (possibly because there won't be good news?). I too, hope it works, but I'm not 100% sold.
Sign In or Register to comment.