Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
The dirty little secret is that aircraft manufacturing currently is a crappy, difficult business with a small total addressable market.
And that's bc nothing is automated in Aircraft Assy, lol. I worked as a consultant for 6 months at Eclipse Aviation, animating the procedures for Visual Manufacturing before they when BK. Every freaking Hole/Cleco/Rheam/Inspection/Rivet is by hand, including the measurements. Rework is nuts involving Engineering... for one aircraft, can they be any more manual? They seriously needs to automate it with smaller robotics (to get in there tight). It could be done and with a heck of a lot more reliably. Scares me to fly after that whole experience. Stuff left inside planes even though they try to eliminate it at every step.
 
Considering we are talking about Waymo FSD vs Tesla I'll throw in my 2 cents.

I'm not an expert on neural nets or any of this. I do work in tech so I know just enough to realize how little I do know. We live in a real physical world and we can take examples from the longest running NN program ever, life.

Life on Earth has continually adapted and improved for 6 billion or so years. Life has tested all sorts of ways to navigate our physical environment, but one system is dominant. Vision. The surface of our planet is bathed in radiation constantly and life has decided to mostly use the visible spectrum to get around. Yes there are exceptions for animals that live in dark areas such as echo-location among Cetaceans and bats, and electrical impulses like you see in Sharks etc. but those are only used because visible light is not available. Furthermore, human life/cities/streets are all designed to be navigated through vision.

Given that, I don't know why anyone thinks LIDAR is the way forward.

Dave Lee put together a nice video on this. I think most of us that truly grew up middle class or lower have to deal with this mindset if we want to truly gain wealth.

In general, biological diversity is a great place to turn for lessons about what works and what doesn't. Evolution doesn't leave many stones unturned.

Ecology and evolution offer a lot of insight into how change happens in the short term and long term. My formal training in these fields has been surprisingly handy when it comes to investing (especially in Tesla).
 
Different rates in FIRST year as the manufacturers include the fee in the car cost

15-18 euros a month PLUS €0.29 to 0.33 per kWh from year TWO

IONITY network dramatically increases EV fast charging costs

"Discounts on the new fees Europe will be available to members of Connected Mobility Service Providers networks, such as Audi e-tron Charging Service, Mercedes me Charge, BMW ChargeNow, Porsche Charging Service and Volkswagen WeCharge.

Mercedes me Charge, for example, has already announced that it will charge €0.29/kWh at IONITY chargers, with no annual subscription fee for the first year. Audi e-tron Charging Service will cost €0.31/kWh plus a monthly subscription of €17.95, while Porsche Charging Service will cost €0.33/kWh plus a basic annual fee of €179. BMW ChargeNow and Volkswagen WeCharge have not yet announced their tariffs.

As a comparison, Tesla charges an average of €0.24/kWh in France, €0.33/kWh in Germany, €0.28/kWh in Belgium, €0.25/kWh in the Netherlands and €0.30 in Italy."

While relevant, this is old news - the article you linked is dated Jan 21.
 
  • Like
Reactions: UkNorthampton
Yes, like Minnesota, Colorado, Wisconsin, Maryland, Missouri, Indiana, Tennessee, Massachusetts, Arizona, Washington, Virginia, New Jersey, Michigan, North Carolina, Georgia, Ohio, Illinois, Pennsylvania, New York, Florida, taxes and California.

Actually, you guys have more small states (by population) than I expected!

upload_2020-7-9_9-21-41.jpeg
 
In general, biological diversity is a great place to turn for lessons about what works and what doesn't. Evolution doesn't leave many stones unturned.

Ecology and evolution offer a lot of insight into how change happens in the short term and long term. My formal training in these fields has been surprisingly handy when it comes to investing (especially in Tesla).


One of my favorite reads:

The Nature of Technology: What It Is and How It Evolves

Leading scientific theorist W. Brian Arthur puts forth the first complete theory of the origins and evolution of technology, in a major work that achieves for the invention of new technologies what Darwin’s theory achieved for the emergence of new species.


The evolution of technology is natural selection.
 
The dirty little secret is that aircraft manufacturing currently is a crappy, difficult business with a small total addressable market.

I'm thinking battery-powered, heavily automated personal aircraft of the type you can't make unless you're the world's leading battery pack manufacturer, or batteries somehow get so light and cheap it's obvious. It's a market that doesn't exist yet. But once the battery tech gets there, it has enormous potential. We're almost there, IMHO.
 
Yes exactly, that seems to be the current model, but I think they will end up going further. Essentially the code is:

if(NeuralNetDetectsATrafficCone())
{
Do some hard coded C++ avoidance stuff here;
}

And I think thats one level too low. I think the number of edge cases and weirdness is just too massive for any normal software engineering to handle. I reckon that code will end up looking like this:

if(NeuralNetSeesSomethingWeird())
{
neural nets decides what to do here;
}

I get the impression thats what Karpathy wants, but there is likely pushback because its scary as hell. The problem is, the first approach STILL requries humans to anticipate every poissible 'issue' and wriote code to respond. You are *bound* to miss things. Thus the 75%/25% NN/C++ split may be a 'local maxima' where that approach gets you to 99% autonomous but never 100%. They may *have* to switch to pure NN systems to get to 100%. We know a pure NN can do it, because human drivers do it.

It doesnt worry me as an investor though, because the value of Tesla having autopilot at 99% FSD is still HUGE, so they can ride out that cash for however long it takes them to go 99-100%.

Disclaimer: Teslas coding team know way, way more about this than me, obviously.

I dunno. In my 2 weeks owning a brand-new 2020 Model S, I would have thought the code was:

if (NeuralNetSeesSomethingAlongsideTheRoad())
{
show an orange cone on screen;
}

Cuz that's sure what it seems like is going on every time I drive anywhere :)
 
I dunno. In my 2 weeks owning a brand-new 2020 Model S, I would have thought the code was:

if (NeuralNetSeesSomethingAlongsideTheRoad())
{
show an orange cone on screen;
}

Cuz that's sure what it seems like is going on every time I drive anywhere :)
Sometimes it moves over to the next lane to get away from the cones (the display actually says that).
 
  • Informative
Reactions: SOULPEDL
I dunno. In my 2 weeks owning a brand-new 2020 Model S, I would have thought the code was:

if (NeuralNetSeesSomethingAlongsideTheRoad())
{
show an orange cone on screen;
}

Cuz that's sure what it seems like is going on every time I drive anywhere :)
Or trash can.

But each one cost $100M.
There is so much Tesla/SpaceX Synergy to make it happen, should let it go to waste. ...
Once you have L5 FSD on the ground...how much more difficult is it to add a Y axis to that data?
 
But I've also read that you can get trailer sway and I don't know what else where the trailer starts wagging the car, and thus you want the human handling the reaction to that (rather than autopilot). Then again, the car already makes minute adjustments to damp that out that I don't feel or worry about (I might suck at towing trailers with anything other a Tesla, as I don't really understand issues that others towing trailers need to deal with; I just drive).

In the manual: When trailer sway is detected, Model X’s electronic stability control system attempts to apply the appropriate amount of braking to minimize trailer sway.

Bjørn Nyland experienced this when driving a caravan in winter:

 
EV startups feel squeeze from Tesla

Tesla Inc.’s new Shanghai plant has churned out popular Model 3 sedans for the past six months, catapulting the company atop the electric-vehicle sales chart and piling the pressure on cash-strapped local rivals. There was another casualty last week.

Byton is at least the third sizable EV upstart to throw in the towel since Elon Musk started his made-in-China offensive, after Bordrin Motors and Jiangsu Saleen Automotive Technology Co. wound their operations down earlier this year.

They fell victim to plummeting demand amid the trade war and coronavirus pandemic, and as the government scaled back the subsidies that turned China into the world’s biggest EV market with hundreds of producers.

Yet Tesla, in just half a year, grabbed a hefty slice of that shrinking pie -- and its portion keeps getting bigger. The market leader’s sales now approach a quarter of the total tally for EVs, the China Passenger Car Association said Wednesday, as wealthier buyers are drawn to Tesla’s brand cachet. That’s making life difficult for the slew of local contenders and risks exposing the multibillion-dollar Chinese EV push as a bubble.
 
Once you have L5 FSD on the ground...how much more difficult is it to add a Y axis to that data?

Seriously, is there anything more complex than Level 5 autonomy in car traffic? Once you have solved that (vision in 3D space) doesn‘t automowers, autovacuums, drones, fleets of manufacuring robots, trains and aircraft carriers look light childs toy in comparison? What else could this autopilot AI be useful for.... hmmm.