Question about professionals and the Porsche GT

[littlebeast]

you are too cute for me to know what the fuck you are talking about. adore you babe. sincerely. please to assist my feebleness. (short term memory lapse being my strong/short suit hahaha.)

 

[UDRider]

Since you explained I'll explain. Every lap on a track is totally unique, as is every corner. It is NEVER the same. The temp is different, the line is different, the angles, power applied, tire traction, traffic, brake pad wear, engine power, all vary in an infinite number of variables. The interesting thing about animals is that they can perform very complicated calculus and catch other animals, catch balls, run, dodge traffic, and basically exist in a world where these calculations take place constantly.

On the other hand, the very best robots can't do .01% of what humans can. Fuck, we can't even hit missiles with interceptors with any regularity, an area of robotics where more money has been spent than any other. The remainder are fucking DUMB repetitive machines that can do very simple tasks. There are no computers in existence capable of calculating and integrating the amazing and infinite stream of data that flows in one corner of one lap. Nor will there probably ever be. You're explaining how a machine that does not exist and can not be built with current tech will be better than a human. No insult intended, but I have my favorite science fiction writers. You're not one of them.

Not sure about whole dog walking robot or motogp robot, but purely from computation power perspective I think we will get there. If don't blow or drive our selves in to extinction. Look at the increase in computation power over the last 50 years, hell over last 10 years. Although it won't be on silicon and transistors, we are reaching the limits of what can be done there quickly. We are now on 14nm process, don't think it can be shrunk much further beyond another gen or two.

 

[mike23w]

Since you explained I'll explain. Every lap on a track is totally unique, as is every corner. It is NEVER the same. The temp is different, the line is different, the angles, power applied, tire traction, traffic, brake pad wear, engine power, all vary in an infinite number of variables. The interesting thing about animals is that they can perform very complicated calculus and catch other animals, catch balls, run, dodge traffic, and basically exist in a world where these calculations take place constantly.

On the other hand, the very best robots can't do .01% of what humans can. Fuck, we can't even hit missiles with interceptors with any regularity, an area of robotics where more money has been spent than any other. The remainder are fucking DUMB repetitive machines that can do very simple tasks. There are no computers in existence capable of calculating and integrating the amazing and infinite stream of data that flows in one corner of one lap. Nor will there probably ever be. You're explaining how a machine that does not exist and can not be built with current tech will be better than a human. No insult intended, but I have my favorite science fiction writers. You're not one of them.

decades ago people said you couldn't build a machine that could beat the world's best chess player for the same reasons. infinite combinations and complexity. then deep blue beat kasparov.

what you describe is exactly what computers are good at. taking in huge amounts of numeric input(temp, line, angles, power, tire traction, brake pad wear, engine power), doing fast analysis and making a decision. computers are much better than humans if the rules are well defined. for a single computerized bike (not a race) on a track, which is what i was describing, i'm pretty sure it could be done with today's technology. a race situation (multiple bikes) is an order of magnitude harder because you have to react to unpredictable people. i was just talking about how fast a single bike on a track could go (why? because i think it'd be neat to know the fastest possible lap time).

google cars drive automatically already. it controls steering, gas, brakes, clutch (somewhat similar to a bike). auto-pilot flies a plane. modern day rockets fly themselves. on a known track with known conditions, i'm sure you could get a motorcycle riding around it (i assume it's easier than auto piloting a plane which exists). would it be fast? not initially, but those are algorithms that can be improved (when to control the inputs based on various track/bike conditions), and eventually with enough time/money i expect it would beat the times of a professional racer - just like deep blue eventually beat kasparov. would it ever be done? doubt it, but i still think it's an interesting idea.

There are no computers in existence capable of calculating and integrating the amazing and infinite stream of data that flows in one corner of one lap.

this is where we differ. i think the amount of information is finite, large but finite, and the variables can be reduced to a manageable set such that a computer (or possible a data center of computers close by) can crunch the information.

 

[MrIncredible]

I think the difference is that it is much easier for car drivers to hit Tilt, the point which you are past your skill limit, because of the safety of the cage.

Any two-wheeled vehicle you could be in a bad crash at any speed. The penalty for error is likely to be an injury.

I do not know what speed the Porsche was going, but assuming that street is 35mph speed limit, did Paul Walker think 70mph was a fatal speed in his car?

Now does any rider think 35mph is a speed which no injuries could happen? Or 70? Is only when you are doing 170 is the risk too high?

As for the performance capabilities of a liter bike or the ability to utilize it 100%, that is a long discussion as well. But for a machine with good speed on the straights and pedestrian handling, yes, I have utilized my motorcycle to its capacity.

You're assuming all cages are safe. My falcon is a death trap- still going to supercharge it.

 

[RRR70]

thread took a turn i don't understand with regard to OP's original question. which BTW - i don't understand either. except of course, for the professionals who have offered their opinion. (though i'm not sure i can tell who they are.)

so let me speculate - if i were a professional, i would be bored to tears to drive an (even above) average POS, and therefore, driving a porsche GT to the grocery store would seem quite normal to me. except for the lack of ample space to stash my cereal, milk and cat food lol.

Here is an OP's stance on a lot of things

 

[aminalmutha]

you are too cute for me to know what the fuck you are talking about. adore you babe. sincerely. please to assist my feebleness. (short term memory lapse being my strong/short suit hahaha.)

Read a few posts back. VW had a major stake in Porsche for a long time before they became a wholy-owned subsidiary.

 

[afm199]

Not sure about whole dog walking robot or motogp robot, but purely from computation power perspective I think we will get there. If don't blow or drive our selves in to extinction. Look at the increase in computation power over the last 50 years, hell over last 10 years. Although it won't be on silicon and transistors, we are reaching the limits of what can be done there quickly. We are now on 14nm process, don't think it can be shrunk much further beyond another gen or two.

Basically what your premise would entail would be a construct capable of computation AND decision making at the level of the .15 quadrillion synapses of the human brain. That's what seems to be missed here. Pardon me if I don't expect to see this in my lifetime and doubtfully in yours.

The top racers ( or even the slow trackday guys) are making huge numbers of decisions and computations simultaneously while riding on the track. Frankly the idea that a robot is going to be built to mimic this tells me someone is not considering the sheer complexity of the process. In order to program a robot to mimic human behavior, you have to be able to program the human behavior. The incredibly complexity of that is the problem.

I have no doubt that TODAY a robot could be built to run in a straight line faster than a human. That's one set of fairly simple movements. Racing introduces a level of difficulty several degrees of magnitude higher.

 

[aminalmutha]

You may be able to overcome the computational speed and data volume issues, but then you have the coding. Try coding a robot to mimic a human brain. Massive undertaking, to say the least.

 

[afm199]

You may be able to overcome the computational speed and data volume issues, but then you have the coding. Try coding a robot to mimic a human brain. Massive undertaking, to say the least.

I suspect that it might take a project similar to the Manhattan project to reach this level of complexity. Of course it's "possible" but the magnitude is the problem. Mimicking a racer is basically mimicking a human being making decisions. That's called AI. The robot has to be able to "look" at a corner, integrate speed, temp, tire heat, # of laps on tires, calculate traction, brake turn and accelerate accordingly. Those things are done really well by top human racers.

 

[Reli]

I suspect that it might take a project similar to the Manhattan project to reach this level of complexity. Of course it's "possible" but the magnitude is the problem. Mimicking a racer is basically mimicking a human being making decisions. That's called AI. The robot has to be able to "look" at a corner, integrate speed, temp, tire heat, # of laps on tires, calculate traction, brake turn and accelerate accordingly. Those things are done really well by top human racers.

All those items on your list could be more accurately measured, and more quickly responded to, by a robot.

You seem to be thinking these are emotional or subjective decisions where being a human has an advantage. I don't think so. Maybe when you're dicing it out with a competitor and deciding how to respond to him, but even then, motorcycle racing is not the Tour de France, where psy-ops are much more important.

 

[Yakoo752]

you are too cute for me to know what the fuck you are talking about. adore you babe. sincerely. please to assist my feebleness. (short term memory lapse being my strong/short suit hahaha.)

Sorry if I added to the confusion being a smart ass.

The Carrera GT was designed as a LeMans LMP1 race car. Think of it like a Desmosidicci (sp?), like a race vehicle you could buy from a dealer.

A lil history here: Porsche and it's sister company VW Auto Group (VAG) now currently the parent company, wanted to prove/race an Audi Diesel, known as the R10/TDI in LMP-1 class. Porsche at the same time wanted a SUV for the North American market. VW/Audi shook hands w/ Porsche and they pulled the Carrera GT from the race circuit, to give the TDI powered R10 a leg up and Porsche gained the SUV they wanted, known as the Cayenne. Porsche put the GT in limited production, since they were already tooled for it.

The GT is a race car that is legal on the street. As awesome as it may be, it's sucks to drive as a street car. I can't imagine being stuck in bumper to bumper traffic with the clutch it has. The brakes are crap when it's not at race speed and at proper temperature. Don't get me wrong, it's an awesome, well designed vehicle; I really wish I could have flogged it around Sears Point for a few laps- that's what it was designed for.

The Carrera GT was never designed to be an LMP however it was born from used parts from the Porsche LMP2000 (sadly the LMP2000 never raced and was scrapped, specifically for reasons you stated before).

VW/Audi shook hands w/ Porsche and they pulled the LMP2000 from the race circuit, to give the TDI powered R10 a leg up and Porsche gained the SUV they wanted, known as the Cayenne. Porsche created the GT in limited production, since they wanted to recoup the costs on the 5.5L V10 engine and racing transmission they had developed for the LMP2000... and this ended Porsches LMP attempts until the RS Spyder debuted.

Again the C-GT is my dream car ... I actually saw a white one on middlefield yesterday

I should note that the LMP2000 *technically* doesn't exist and never did according to Porsche

 

[Asphaultnaut]

I could buy a cannon but what is the point if I cannot shoot it? I rather have a rifle which I can take to any range.

I can fathom the rationale behind buying a super Porsche, from impressing the dudes at the garage to showing off my penoz substitute.

But I have never driven on a race track, so I am trying to understand power and handling better.

I have both. a smoothbore BP replica that chucks a 1 inch lead ball at a shade under the speed of sound. Only can't use it around here because of NIMBY's buying property next to the range complain about the noise-if high brass offends them, a few scoops of powder would send 'em screaming to their legislators.

But what bugs me is that many BARFers have ridden (or still do) bikes with a higher power/weight ratio with even less safety features and less available traction. Only a few of you 1%ers have traction control available at all, and not too many more have ABS. Many actual riders here have ridden bikes that are absolute deathtraps compared to a Shelby Cobra or Dean's 550. Bad suspensions, horrible drum brakes, peaky torque bands, handling geometry designed by visual appeal...you name it.

When BARFers get on the "unsafe vehicle" bandwagon, especially one that exists right now simply to insulate Hollywood from it's own choices and consequences, it stinks like "wanting to fit in to the popular opinion". Especially when that opinion will be rendered meaningless by action the moment many of us fire up their bike and commute home this evening, or go for a ride this weekend.

I mean FFS, current year Harleys have a higher body count than Porsche's entire "exotic" lineup.

All those items on your list could be more accurately measured, and more quickly responded to, by a robot.

You seem to be thinking these are emotional or subjective decisions where being a human has an advantage. I don't think so. Maybe when you're dicing it out with a competitor and deciding how to respond to him, but even then, motorcycle racing is not the Tour de France, where psy-ops are much more important.

I've been chewing on this one for awhile now, and here's the roadblock:

ONE AI system will choose the perfect line on a given course for given conditions, and will outplan and outmaneuver human drivers (assuming they stay predictable) or out-react (by responding before a chaotic movement can result in a crash by realtime crunching of all vectors and variables).

Now add TWO AIs. Or three. Or the entire field.

ALL are capable of "thinking" thousands of times faster than the physical vehicles can respond. ALL have sensors that are thousands of times more accurate than human senses. ALL "want" the same perfect line and ALL are plotting the best course to get there. Coming to the SAME conclusions. Having control software that can respond a picosecond faster than the competition is meaningless, since the vehicle itself in meatspace will not be able to respond any faster-and the amount of variables will be unchanged so this provides no benefit for predicting accidents. The "better" AI will just be waiting longer for new input from the sensor stack that it already queried a thousand times while the lateral sidewall oscillation of the tires stabilizes to predicted and desired parameters for course correction. Plenty of extra bandwidth to re-count the fuel and air molecules heading into the intake for a 5th time until that valve gets around to overcoming inertia and lifting.

winning becomes only a matter of luck-which vehicle experienced a mechanical or software failure. Everyone else will be gridlocked into all trying for the optimum line. None will make the intuitive leap to "try something" unless that deliberate path has been set in advance-which means pretty much a randomized response that is not optimum. That kind of defeats the purpose of Expert Systems or the philosophy of AI. A machine would need to either compete in a field of imperfect responses, or be able to show intuition, or act outside of the "best" choice.

In other words, deliberately make "mistakes". Which is what human drivers are already doing.

 

[aminalmutha]

All those items on your list could be more accurately measured, and more quickly responded to, by a robot.

You seem to be thinking these are emotional or subjective decisions where being a human has an advantage. I don't think so. Maybe when you're dicing it out with a competitor and deciding how to respond to him, but even then, motorcycle racing is not the Tour de France, where psy-ops are much more important.

Machines are falible, contrary to popular belief. It can take a lot of time and set-up to get even a relatively simple machine to work correctly again and again. Getting one to perform something like winning a MotoGP race would be a massive and very expensive project.

 

[afm199]

All those items on your list could be more accurately measured, and more quickly responded to, by an imaginary device we can't build, don't have the tech for, and that I dreamed up"

You seem to be thinking these are emotional or subjective decisions where being a human has an advantage. I don't think so. Maybe when you're dicing it out with a competitor and deciding how to respond to him, but even then, motorcycle racing is not the Tour de France, where psy-ops are much more important.

Fixed that for you.

I don't see how you equate emotional in this example. Catching a ball is not emotional, it's sub conscious computation and activity. The decisions are not made at the conscious level, but the computation is real. I can stipulate an imaginary robot that mimics a human entirely. That doesn't mean one exists or will ever.Think about how complicated an action it is to see a ball hit, track it visually, estimate where it's going, run there and catch it. The way you have a robot ride on a track is to first program it to emulate human behavior, then look for ways to improve on that. So yes, if you had the money and tech, you could build a robot to follow and exact path that never varied around a track, emulating a human.

Have you ever raced? If you think psyching an opponent is not part of the game, i have news for you.

 

[wazzuFreddo]

You're assuming all cages are safe. My falcon is a death trap- still going to supercharge it.

Sounds like something a reasonable family man shouldn't be involved in, I'm down to help with that.

 

[RRR70]

Sounds like something a reasonable family man shouldn't be involved in, I'm down to help with that.

Sounds like fun. I wanna play.

 

[Reli]

Fixed that for you.

I don't see how you equate emotional in this example. Catching a ball is not emotional, it's sub conscious computation and activity. The decisions are not made at the conscious level, but the computation is real. I can stipulate an imaginary robot that mimics a human entirely. That doesn't mean one exists or will ever.Think about how complicated an action it is to see a ball hit, track it visually, estimate where it's going, run there and catch it.

Have you ever raced? If you think psyching an opponent is not part of the game, i have news for you.

Why do you think a human can predict where a ball is going to land more accurately than a robot? I'd say building a ball-catching robot would be a ton easier than building a MotoGP robot. See the ball in the sky, measure it's arc and altitude, and measure the wind as well. All of those things can be more accurately measured by machines. There is no variation other than a sudden shift in the wind.

And psyching an opponent? Of course it happens, I never said it doesn't, but I don't think it's as big of a factor as it is in cycle racing, where it has a direct impact on the other person's energy reserves, not just their place in the race.

My whole argument is that none of you, not even the best racer, is getting 100% of a bike's potential around a track. So saying "You shouldn't buy a literbike or even a 600, because you aren't using all of its potential" is a stupid statement, because neither is Rossi.

 

[Asphaultnaut]

Fixed that for you.

I don't see how you equate emotional in this example. Catching a ball is not emotional, it's sub conscious computation and activity. The decisions are not made at the conscious level, but the computation is real. I can stipulate an imaginary robot that mimics a human entirely. That doesn't mean one exists or will ever.Think about how complicated an action it is to see a ball hit, track it visually, estimate where it's going, run there and catch it. The way you have a robot ride on a track is to first program it to emulate human behavior, then look for ways to improve on that. So yes, if you had the money and tech, you could build a robot to follow and exact path that never varied around a track, emulating a human.

Have you ever raced? If you think psyching an opponent is not part of the game, i have news for you.

Google's self driving car program is already doing what experts 20 years ago, even within the AI community, claimed was impossible. the DARPA challenge of a couple years ago was also claimed impossible by naysayers, myself included.

Interception of warheads in theater (the field-expedient, repurposed Patriot missile) had a better accuracy than what SDI proponents dreamed of a couple decades earlier-and did it with less time to calculate.

Cryptographic keys that uni profs said would take supercluster systems decades to solve are cracked in hours by software mods to the GPUs on gamer video cards you can score right now at Fry's.

Hell, I can even build GPS guided aerial autopilots for RC planes that are more accurate than any cruise missile fielded in the 80's. And spend less than a trackday's cash or race day's attendance.

You've been around long enough to witness the curve. To see for yourself Moore's Law getting modified again and again. There is absolutely no evidence to suggest that the rate of available computation is slowing down-and remember, you don't *have* to have the computational hardware aboard the vehicle..."cloud" it and do everything by remote!

Psychology is only a factor against humans. But I guarantee you a machine that knows no fear is gonna scare enough riders that they will refuse to compete-citing all sorts of cherry picked reliability "facts" and safety concerns of course...

the only winning move will be....not to play

 

[afm199]

Why do you think a human can predict where a ball is going to land more accurately than a robot? I'd say building a ball-catching robot would be a ton easier than building a MotoGP robot. See the ball in the sky, measure it's arc and altitude, and measure the wind as well. All of those things can be more accurately measured by machines. There is no variation other than a sudden shift in the wind.

And psyching an opponent? Of course it happens, I never said it doesn't, but I don't think it's as big of a factor as it is in cycle racing, where it has a direct impact on the other person's energy reserves, not just their place in the race.

My whole argument is that none of you, not even the best racer, is getting 100% of a bike's potential around a track. So saying "You shouldn't buy a literbike or even a 600, because you aren't using all of its potential" is a stupid statement, because neither is Rossi.

I think a robot could be built to walk dogs. But nobody has. Nobody has built one to catch balls, which is far less complicated. Nobody has built one to do anything outside of a very narrow range of objectives. So your posts remain conjecture. I think the human can run a three minute mile, so it must be true.

PS: I never said no one should but a literbike. I said there are a handful of racers who can ride one close to its potential.

I don't much care what bike people buy. It's none of my business.

 

[Yakoo752]

ball catching robot is done

[youtube]83eGcht7IiI[/youtube]

taking this to the next level wouldn't be impossible

1:40 has some simple juggling