HopJokey
Member
Back to price action, there's some movement in TSLA pps in the pre market right now.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
Artful Dodger
"Neko no me"
During Investor's Autonomy Day Karparthy stated that the software continuously verifies the consistency between the sensor input and the orders for the actuators. So in principle the car should be able to react reasonably even during an unfolding accident.
I speculate that the psychological stress from experiencing a near-miss while driving can be amplified when the AP/FSD does something unexpected to avoid or diminish the accident.
I think this stress could be reduced if the FSD computer would use a synthetic voice to concisely announce its actions as they unfold.
Let's say it detects and starts reacting to an impending collision. While doing so it could then say e.g.
"Collision Evasion" - or if during maneuvering it realizes that it is not able to fully avoid the collision, it could complete the already spoken "Collision" with "from behind" - or something along those lines.
And I fully agree that to reduce the stress of driving, Tesla should make it known how the car can react - e.g. with a little textual slide show, played while upgrading the firmware, or maybe during charging.
I'm talking about mainly software. Mobileye has hundreds. Tesla has 0. Even for hardware, that's not even autonomous cars related at all and as Tesla said themselves, neither of those will ever be approved (probably because of prior art or being too broad).
In acouple hours i will post the NN models in the Q4 2017 Eyeq4 versus Tesla's Feb 2019 AP.
I meant in terms of accident avoidance of the intended swerve. It probably already prevents swerving the same way it prevents passing if there is traffic in the other lane.
Please post it in the appropriate forum.
Total speculation: The cap raise was today..maybe oversubscription or realization that a big supporter bought up a ton of shares?
having 100s of patents is not enough. it's rather easy to get a patent. they also have to be defensible once challenged. and it's much harder to defend a patent successfully, especially in a space like machine learning
bhtooefr
Active Member
This could explain some of it too: Tesla CEO Elon Musk spent $25 million to buy more stock, boosting stake to nearly 20%
Maybe it becoming official that Elon purchased $25 million worth of shares got some early movement?
Anecdotal story from Denmark. Called Tesla to book a test ride in a model 3 around Copenhagen.
They have 3 cars doing doing test drives at a rate of 1 pr hour pr car.
First available time slot was on the 29th of May.
They have 3 cars doing doing test drives at a rate of 1 pr hour pr car.
First available time slot was on the 29th of May.
My understanding is (correct me if I'm wrong), that the new shares issued for the capital raise are delivered today at $243 price to the buyers. So what is the expectation for those ?
Are the new buyers going to be long holders ? Probably not all.
If some of them are bought with the goal of short term profit, at what point can we expect them to be dumped on the market ?
Immediately today as the SP is more than 5% above their cost ?
Wait until they make a 10% gain ~$267 and then start selling them at that point as limit order ?
I would guess the last option and so I expect a new resistance at 267.
Are the new buyers going to be long holders ? Probably not all. If some of them are bought with the goal of short term profit, at what point can we expect them to be dumped on the market ? Immediately today as the SP is more than 5% above their cost ? Wait until they make a 10% gain ~$267 and then start selling them at that point as limit order ?I would guess the last option and so I expect a new resistance at 267.
...or maybe bc bmw posted a loss on its car unit for the first time since 2009
Tesla, TSLA & the Investment World: the 2019 Investors' Roundtable
(i guess most of you blocked me, oh well)
Tesla, TSLA & the Investment World: the 2019 Investors' Roundtable
(i guess most of you blocked me, oh well)
Neural network models are not "developed" but "trained", with Large amount of labelled data, and trial and error in setting hyper parameters or tweak of architecture, whose results are again verified by large amount of labelled data. Development of NN is completely nonsense when you don't have access to huge amount of labeled data.
In multiple presentations and blogs Dr. Karparth addressed the hard problem of large scale data labelling and how to automate most of it. He reiterated some of those approaches in autonomous day and how Tesla's large, connected fleet plays a vital role in not only collecting but also filtering labelling of the data.
Unless you provide some evidence to how mobile eye is better at data collection and labelling, you sounds like certain troll on Twitter who likes to say "I know much more about this than you".
I ran across this Ars article about Electrify America's effort to "simplify" electric vehicle charging. Given the benchmark of simplicity set by Tesla, any article with a headline like this one should comprise a single sentence saying something to the effect of, "Just open an account, link your credit card, and plug in at any Electrify America charger." Suffice it to say, it's more complicated than that. Go figure.
HG Wells
Martian Embassy
Headline: Shortsville Times
Elon buys more stock on margin. Brings Tesla closer to DOOM.
We now only need to get the stock price below $28 to create a margin call and then BANKWUPCY !
Elon buys more stock on margin. Brings Tesla closer to DOOM.
We now only need to get the stock price below $28 to create a margin call and then BANKWUPCY !
That would be the kindergarten variety.
Real developers follow a procedure by which documentation, release information and unit tests(*) are all updated along with the actual code.
(*) Some believe that the unit tests should be written by a separate group of coders who only rely on the documentation (since this helps to test also the documentation), in which case the non-test coder gets to write more detailed documentation (instead of just writing "see unit test code for examples").
Similar threads
- Locked
- Locked
- Locked
