None of those have the same impact as driving a car in the real world, nobody cares if a chess playing model makes a mistake 
(and you’ll be surprised by how easy some of those models are easy to confuse).
And we’re talking about being able to handle all kinds of situations that weren’t seen before.
Level 5 full autonomy 2020, if you believe Elon.
https://www.google.com/amp/s/www.forbes.com/sites/johnkoetsier/2020/07/09/elon-musk-tesla-will-have-level-5-self-driving-cars-this-year/amp/
I find it funny, that we discuss about AI and self-driving and at the same time to prove I am not a robot surfing the internet, I have to choose which of 9 photos don’t contain a photo of a hydrant or a stop sign. Let’s hope the robot drivers are more intelligent than the internet robots.
Be careful about AI… AI not exist it’s just advertising right now.
For AI we need to develop a new programming language that doesn’t need to be compiled and can be write itself and directly understand by the processor something like assembler.
AI now = computer executing an algorithm created by humans with big databases and a lot of calculation power for process a lot of data quickly.
Not view last progress for Tesla but high speed way is cool because all are at the same speed but city more complicated because need to detect what happen and they have to anticipate what can happen. It’s not a chess game where all rules are fixed. Everyone can do almost everything on the road. I think everyone have seem some unbelievable thing like someone who changing lanes side on the highway. Someone going the wrong way on the road.
An algorithm should know all possible situations and that’s very complex task. What do a tesla if a car wrong way come at 120kmh and tesla is a 120kmh too on the same lane ?
Just break, try to changing way even if a car is on other lane to avoid a head-on collision or just accelerate because the vehicle with the lowest speed have the highest damaged and highest chance of dying ? What human can program an algorithm for that ?
When you drive your subconscious mind can make these kinds of decisions sometime emergency break isn’t a solution 
EDIT : like in a plane the pilots are often responsible for crashes it’s the same for car. People not follow some rules and create crash. 100% autonomous vehicle is really safe because all follow the same rules but free driver and autonomous vehicle isn’t really good 
The thing is, AI will not have to be perfect to be beneficial. Imagine that a human driver causes 6 accidents every million miles. Now AI reduces this number to 1, but also adds 1 weird accident that a human would not make. You still have 3x less accidents, more lives saved, less property destroyed. Not to mention other economic advantages, such as: no need for garages, no need to own a car, now you can work/study/relax instead of driving.
That’s how you’re training these neural networks :). They take some edge cases and crowd source them via CAPTCHA. At least that’s what I heard.
the weird phase will be in between first adopter and global AI rollout
Most weird situation are caused by humans. When all car “follow” the rule and drive around by a computer, I think potentially the situation is less dangerous than a mix of lots of unpredictable humans and some AI driven car
You want it, you got it. Here is a fresh wrap up of the current state of Tesla’s self driving tech. I’m happy to see that it confirms what I already wrote above. They are going from covering 99% to 99.9% to 99.99% of situations, which is just very time-consuming.
I wait to view that in real situation and not “planned” situation by an algorithm.
A head-on collision is very interesting to view what happen really and what choose the computer for you. Save your life, other life or just emergency brake.
Musk said that Tesla’s autopilot is optimized for avoiding an accident. The trolley problem is very theoretical and probably happens very rarely. How realistic is it to assume that the autopilot will find itself in a situation, where the only two choices are: kill two pedestrians or kill the driver? And how would a human react in such a case?
I’ve given this a lot of thought recently and decided to leave the rule only for what it was conceived, that is the all country equal weight ETF portfolio.
Rebalancing is dangerous for a stock portoflio, as you are basically trimming winning companies, to reinvest into the losing ones. At this point, I decided to let all the winners run, and systematically get rid of the laggards.
By the way, guys. Thank you very much for discouraging me from buying TSLA stock. On the day of publishing of this topic, the price was $800, a now it’s $1544 
You’re welcome.
Glad to see you didn’t do such an unwise investment.
Happy to give you more of the same “you and I don’t know more than the market”-advice whenever you need to hear it. 
Just imagine you listened to the advice of buying NIO instead. It was $3.4 when you opened the topic. It is $15 today.
Or Wirecard lol.
Always easy to call the winner in hindsight.
It was also below 400 in March, and it will be either below 1000 or above 2000 in 2021 - so what can one do about it rather than just see it as a bet (which it is), and act in line with it? 
In case of Tesla, it has been long criticized by “experts” and it was a favourite pick for shorting. It oscillated around $300 for a few years and finally it broke out. The market is no longer in denial. I have a feeling that I wrote this topic in the last moment where I would still consider buying. Now it would just feel too much like chasing the bull.
Well, I followed news about Tesla, not about Nio, so I had some foundations to base my belief in Tesla on. But it’s nice to have that perspective… it’s crazy how much a stock can go up in just 2 months.
Evaluating investment decisions on the outcome is a really bad way to evaluate investment decisions.
You can’t rally make this kind of decisions based on online advice. You need your own convictions to be able to open a position and then sustain volatility and drawdowns. Additionally, if you go into single stocks, it’s very risky to only own one company. You should probably own at least 15-20 names to make a diversified portfolio.
BTW, Tesla now at $1600 premarket
Another perspective:
I once (around 2018) held XXX Tesla Shares (average buy price of 250$). But after 12 months of Elon being “troll” CEO with “420-twitter” speeches and other fail situations in the media I had enough. Its hard to hold a stock which you hear from the CEO every week another buzz story. I sold when it hit 350$ and called it a day and never looked back. Tesla, Solarcity and even Space X and boring are great companies but they all need 1 person to succeed and that is Elon Musk - what happens when he gets hit by a car or the company says we dont want you to be CEO anymore?
In conclusion:
It was okay to hold Tesla because under the aspect of “high risk high reward”. But looking back the position I had in comparison to my overall portfolio was simply too big. This was my mistake. When I would do it again I rather hold a smaller position with my “fun money portfolio” and set “cash out” prices. For example: Sell 20% at 500$, sell another 20% at 800$, sell 20% at 1400$ etc.
Would I buy Tesla now at 1500$?
No. When buying single stocks: buy when they are cheap or in accumulation phase and not when they spike.
And eventually end up with 0% of your biggest winner? Upside is theoretically unlimited if the company can keep on growing and expand into other addressible markets
And yet, you’re more likely to win by buying winners, than looking for “undervalued” companies, also known as loosers.
Just look up MSCI World Momentum. It’s a consistent outperformer because winners tend to keep on winning.
GM, Enron Corp, Lehmann Brothers, Kodak and Yahoo were once considered winners and then something else happened.
