Planet-9 Porsche Forum banner
  • NOTICE - Before adding photos to posts on Planet-9, please review: Posting Photos on Planet-9

Status
Not open for further replies.
246K views 2K replies 142 participants last post by  chows4us  
Hot take from PEC LA: it's pretty damn awesome but I wouldn't sell the 981 to get into one. 718S is not a monster engine like I thought it would be - definitely moves but doesn't kick your stomach or peel your eyelids back. The turbo whistle could be heard in the driver's seat, but not the passenger seat. Startup exhaust sounds great, tooling around in 1st gear is notchy-sounding for sure, at full throttle the sound is good.

Image


Image



Sent from my iPhone using Tapatalk
 
Save
I heard the new president doesn't care how much gas we waste. So maybe we can get a real 981.3 instead of that abortion of a 718.


Sent from my iPhone using Tapatalk
 
Save
If you want sound get an M4. But hug those turns get the 718
The M4 does not sound good from outside of the car. The 2018 S4 sounds a lot better. Mercedes V8 biturbo is the king of the roost.

If my Boxster got stolen or wrecked, I'd get a new 718. I wouldn't replace a perfectly good 981 with a 718. I also wouldn't buy a used 981 if my 981 disappeared. The sound is a regression but the rest is an evolutionary improvement worth having.

Sent from my iPhone using Tapatalk
 
  • Like
Reactions: PcarBob
Save
Here's another aspect of autonomous driving vehicles that most people don't think about; self driving trucks. I lead technology for a large logistics company, and the number one constraint trucking companies face is a driver shortage. There are thousands of open truck driving positions right now and the issue will be much worse over the next 10 years as the veterans retire. Our average driver made $68,000 last year, but nobody is raising their kids to drive trucks anymore, it's a hard life and kids these days just aren't cut out for the work and sacrifice it entails. MB out of Germany is leading the charge here. Current state is one driver in a truck up front with 3 or 4 driverless units following behind him. Raising driver pay to say $100,000 a year would drastically increase the supply chain costs across everything that trucks carry, and trucks carry everything....everything.
I 100% think that current technology can support this use case, for long hauls only. Take a container from exurban Las Vegas to West Texas, 100 miles from Fort Worth. Then switch to a local driver who can do the last miles. The city streets stuff is just too dangerous for current gen robots.


Sent from my iPhone using Tapatalk
 
  • Like
Reactions: Rickyb
Save
Let's assume it costs more than the F35 JSF to build a self driving car program that someone with a 70 IQ can use without risking the death of himself of those around him. I'm going to ballpark it at $50B. You've got 10 companies that have put $1B into their programs, which are silos. None of them are going to get close. The company that brings this to market doesn't exist yet, or isn't in this business yet. I think GE could do it. I think Ford could do it. I think Boeing could do it. I don't think a unicorn startup can pull it off before everyone gets bored and quits. It's a century pursuit.

The most dangerous phase is when the car says it's mostly driving, you trust it, and then it doesn't react to changing conditions like a tractor trailer turning in front of it. It assumes this is light playing a trick on its eyes, and you end up decapitated because you were on your phone. True story.

There are different levels of scrutiny for transportation. We have no tolerance for airplane accidents even if planes are generally safer than driving. United can't murder 200 families on the way home for Thanksgiving and say "oh well you peasants, you probably would have died driving statistically anyway -- eventually." It's probably why we don't have pilotless passenger aircraft. It's because one nerd who isn't a pilot can't codify everything Captain Sully knows into 5000 lines of C.
 
  • Like
Reactions: Grand-Cayman
Save
"using software written by a bunch of snowflakes hyped up on caffeine that float from job to job..."

<grin>
+1000. Otto was started by 2 google engineers in January 2016 and by May 2016 it was sold to Uber for $700 million, about $10 million per engineer. Most of these guys will be pursuing passion projects semi-retired before they even hit 15 years in the workforce. Not exactly the kind of follow through needed for a project this complex.


Sent from my iPhone using Tapatalk
 
  • Like
Reactions: Rickyb
Save
Vision 2020. Not vision "I need to delegate care for my own life to an unfinished computer program so I can tweet".


Sent from my iPhone using Tapatalk
 
  • Like
Reactions: gcurnew
Save
I think you guys are all smoking a pretty wild strain down in Santa Clara. The hubris is staggering. As if you're just one all nighter code bender away from pushing a binary that will:

1/ not make the same mistakes as humans
2/ not make mistakes safe driving humans don't make
3/ won't maim or kill
4/ will follow the ethical profile of the owner
5/ will obey US laws

I've seen every type of self driving prototype tooling around SF with 3 engineers in the car with laptops. I'm still saying 25 years before it can go 5 miles on SF surface streets at 5:00PM with parking lanes as travel lanes and get me home in the same 27 minutes it takes a risk calculating human who can deal with insane ambiguity.

Here are a few snaps. Also saw a ton of the XC90 before they got banned for blowing red lights they couldn't see (but any sober human would know to stop when other cars are stopped).

Image


Image


Image





Sent from my iPhone using Tapatalk
 
  • Like
Reactions: gcurnew
Save
Has anyone actually driven a new Cayenne Turbo? The instructor at PEC told me it was his favorite car they have. It does everything, he exclaimed. Plenty of profit margin in that.

Self driving cars are a marketing stunt in 2017. I think they're 25 years away because we don't have mobile edge artificial intelligence. We don't even have it in data centers because the chips, low level instruction sets, and the software platforms are not developed much less mature. Right now AI is just running on graphics cards for childrens games, as a stop-gap. It's not going to run on Intel x86 microprocessors. Takes decades to build out fabs for this.

Until then, the self driving programs will grow to billions of lines of code as they try to account for every edge case and avoid every accidental death they run into. It's unmaintainable, which is why you need a learning computer to truly self drive. With a local distributed database with the "knowledge".

25 years.




Sent from my iPhone using Tapatalk
 
  • Like
Reactions: erik_plus8
Save
My Boxster is fun to drive but the quality of the standard interior is quite mediocre for a 65k car. The paint is also poor.
Porsche and quality, perhaps yeah although some may dispute that. Porsche and value? is a joke.
Really? I can't think of a more reliable, well built sports car. Ever hear McLaren owners complain about basic things like the headlights don't work and parts fall off? Porsche gets all of these details perfect. Objectively, the quality measures show. No squeaks, no rattles, etc. as in any BMW.


Sent from my iPhone using Tapatalk
 
  • Like
Reactions: Rickyb
Save
$90,000 autocross weapon lol. I guess the problem with a 718 is you can't hear it with a helmet on. Like a 918 spyder in pure electric mode.


Sent from my iPhone using Tapatalk
 
Save
Having driven the 718S on Porsche's track, it's nothing like a Subaru in engagement or aura. It's the classic argument of a snobby detractor who thinks old is better than new. Not better, just different. But if you are short on talent the 718 will help you keep up with me around a circuit :)


Sent from my iPhone using Tapatalk
 
Save
Do you know how much dirt 5 yards? Have you ever seen it dumped in a driveway?

The kind that needs it "occasionally" across a lifetime. The kind that needs peat moss, or potting soil or any one of a myriad of yard work bulk products when it happens, not at a specific time. That kind that needs to haul stuff to the dump, or the kind that needs to haul large items from stores.

If you don't have a large yard, or small one either, then you will never understand. Life "happens" and at any time, I will need to go get stuff. Your approach seems to be one that send you to the poor house quickly.

I'm going to guess you don't have a spouse telling you to go NOW to go get something.
I've shoveled 5 yards of dirt and mulch more times than I care to count, as a teenager doing some freelance landscaping. Good motivation to go to and finish "hard college".

Amazon delivers this stuff same day (5 hours) so I don't have the problems you describe. Instacart brings things in 2 hours. Uber brings things in 22 minutes. There are dozens of other apps that make items appear or go away. How else could I own two completely impractical cars?
 
Save
Sure, I'm going to pay $23.59 for a $1.69 bag of topsoil. Good luck with those prices. Timberline 1 cu. ft. Top Soil-50051562 - The Home Depot
For city folks, it's whether or not I'd pay $1500 a month to own and park an Audi Allroad so I can drive the u-pick Amazon warehouse. Or the grocery store. Or any big box store for that matter.

What kind of fool buys top soil by the bag when a dump truck from the nursery can deliver you 5 yards for a $60 delivery fee?
 
Save
Save
Having seen a lot of 991.2 on the road now as a pedestrian in SF, I have to say my base 2.7L Boxster has a lot more presence than the turbo 3.0L. That doesn't leave the 718 a leg to stand on.


Sent from my iPad using Tapatalk
 
Save
That's your metric for software and complexity? HA HA! I hadn't realized you were posting your messages from sometime in the late 80's. Welcome to the future, time-traveler!

Seriously, you almost certainly have more than 5K lines of C in your 981 engine management system. Modern SW development moved past C long, long ago. And far more fundamentally, the idea that anyone is even trying to "codify everything Captain Sully knows" is a very dated view of how complex decision-making systems operate. Read up on machine learning.

Anyway, modern passenger planes are highly automated already, and as others have pointed out, air travel is much less dangerous than car travel. So the benefits of closing the remaining gaps in automation are much less compelling than for cars.



Why on earth would you assume the most expensive military weapon in history is a reasonable proxy for the cost of self driving car development? Because private industry is known for outspending the military on vehicles? It just makes no sense.

/Brett
I think you are proving my point. We don't have edge mobile artificial intelligence that we can stuff into a vehicle. Nvidia is taking us there but it's several generations of chips away from being great. I'm not convinced GPUs are the best architecture, but they are a workable one as they pivot towards their big vision. I could imagine them developing and fabing new chipsets specifically for this purpose, from scratch.

So in the meantime, the size of the self driving program resembles traditional software. The program grows in size as you bug patch every time the car killed a human or made a mistake. Because it's not actually smart or sentient or able to hold the context the human brain has. It's just evaluating polygons and deciding whether to continue on the route or avoid/stop.

I don't think this problem is going to be solved by 150 engineers at Uber who all want to quit but can't because they're wearing golden handcuffs. I think it's going to take tens of thousands of engineers on a large expensive program with a lot of complexity to get the 200 car models currently for sale able to drive themselves and meaningfully reduce fatalities.
 
Save
End of thread, end of topic. No other arguments are necessary, correct, or, in fact, possible.
Robocop doesn't need to be good, he just needs to shoot fewer unarmed civilians than regular cops. Everyone will celebrate the corporation for producing such great code. I'm picturing a parade for the robot that only shot and killed one resident.

What strain are you smoking up in Washington?
 
Save
There are pretty good stats on what causes the US road deaths. Saying the robot will save all lives is about as foolish as saying PSM will end all loss of vehicle control.

The most common causes of fatal accdients have all been addressed by driver assistance systems - intersection collision warning, lane departure warning, auto braking, etc. I ordered PAS on the Boxster and it is a total lemon, and I've lost trust in current generation systems. I realize Tesla is leading but they also killed a customer with no remorse, which is concerning because their meat grinder is pursuing a higher purpose than your life.

The other causes of fatal accidents are mistakes the computer still make, like running red lights and stop signs.

No computer is going to be able to do a city drive until T1000 is literally sitting in the driver's seat, except a lot smarter and hopefully less homocidal.

3rd Street in between Folsom and Howard in SF, with the self driving Volvo XC90 which I had been watching the engineers manually drive for months until Travis said it was GO TIME BOYS and unleashed his billion dollar project for livery rides:

https://m.youtube.com/watch?v=_CdJ4oae8f4

https://www.youtube.com/watch?v=_CdJ4oae8f4


Sent from my iPhone using Tapatalk
 
Save
I'm not about to forecast when automated cars will be available, but when the technology is perfected, accidents should virtually be eliminated. There are approximately 27,000 accidents every day when cars are driven by humans. These accidents provide good income for doctors, lawyers, body shops, and towing companies. So as I posted earlier, economic upheavals will abound.
If you don't recognize that humans add value to the world and when they die, GDP falls then you are one cold bastard! You're talking about losing salvage from the economy causing a dislocation? Quite the opposite friend. When some poor sap making $40k a year ends up with an $80k medical bill and a $20k auto repair bill, nobody wins. In fact, you can see our country's inability to afford insurance playing out on TV right now!


Sent from my iPhone using Tapatalk
 
Save
Status
Not open for further replies.
You have insufficient privileges to reply here.