James Douma and I talk about most current Tesla FSD advancements, FSD beta v11 release notes, Elon Musk, Twitter, TSLA and more. Reward section after 1 hour mark on language learning and mental skills development.
James Douma on Twitter,
James Douma on AI playlist,
Tesla FSD Beta v11 release notes:
– Made It Possible For FSD Beta on highway. This merges the vision and planning stack on and off-highway and replaces the legacy highway stack, which is over four years old. The tradition highway stack still counts on several single-camera and single-frame networks, and was setup to handle easy lane-specific maneuvers. FSD Beta's multi-camera video networks and next-gen organizer, that permits more complex representative interactions with less dependence on lanes, give way for adding more intelligent habits, smoother control and better choice making.
– Enhanced Tenancy Network's recall for close by barriers and accuracy in severe weather with 4x increase in transformer spatial resolution, 20% boost in image featurizer capability, enhanced side electronic camera calibration, and 260k more video training clips (real-world and simulation).
– Improved merging behavior by leveraging lane geometry and lane bounds, association with coarse map information and better gap selection algorithms, enabling a smoother and much safer experience.
– Added highway habits to balance out away from blocked lanes and generic barriers like roadway particles while likewise adding a smooth handoff in between in-lane offsetting and lane altering.
– Improved speed-based lane modification decisions to much better avoid decreasing traffic in fast lanes, and interfere less with navigation.
– Reduced level of sensitivity for speed-based lane changes in CHILL mode.
– Improved lane modifications to enable greater jerk maneuvers if required to stay on-route or to move far from lane obstructions.
– Enhanced smoothness at highway lane divides by being less strict about focusing in between lane lines and allowing lower jerk maneuvers, where safe to do so.
– Decreased latency of trajectory optimization by 20% usually, without compromising habits, by leveraging mathematical tricks for more effective calculations.
Social
Twitter
Apple Podcast:
Spotify.
Instagram.
Dave Lee on Parenting,.
KEEP IN MIND: Please do not succumb to scammers who might impersonate me or others in the remark area. I do NOT offer my number or ask people to call me.
♂ Disclaimer: All content on this channel is for discussion and illustrative functions just and must not be construed as professional monetary recommendations or recommendation to purchase or sell any securities. Ought to you require such advice, seek advice from a certified monetary or tax advisor. All views revealed are individual opinion as of date of recording and undergo change without duty to upgrade views. No assurance is given concerning the precision of information on this channel. Neither host or guests can be held responsible for any direct or incidental loss incurred by using any of the details used. Author is long TSLA and other stocks at time of original video publish date. When you purchase through links in this video description, author might make an affiliate commission.
Open an brokerage account at Interactive Brokers:.
Have a look at my archived articles/posts on Tesla and investing:.
#Tesla #TSLA #Stocks.
Wealth Builders Club Secrets Revealed – Click Here to Discover the #1 Investment Resource!
10:00 Thanks James for your calm and rational explanation of how the stock market works. Agree with everything you said.
Me too. I think we should repost it on Twitter and have it played before every Tesla Youtuber does a video.
With regard to stock buyback, I disagree with James that it’s a one-quarter pump. There is a cumulative effect of reducing the float over time. THAT’s the benefit of buyback.
We got FSD Beta last night, can’t wait to try it out. So happy to see it finally rolling out. Expectations are very high.
Make sure you keep your hover hands ready until you get a feel of where it’s great and not so great. Enjoy it!
When karpathy said the autopilot team was on autopilot I knew everything was going to be okay 😉
Whoopie! Housebound, recovering from a successful medical procedure, I look up at my TV and see that Dave Lee and James Douma have a new episode up. I learn so much listening to these guys.
I’d like to see some kind of recovery by February, so that when my son comes to visit, I can demonstrate that “Father Knows Best”. . .
Hope you have a good Thanksgiving and a speedy recovery!
@D R-K Seconded.
I hope they don’t do a buy-back. They can surely put their excess cashflow to better use. If they want to scale to insane levels, they will need lots of cash.
@Swiffah145 That’s interesting. In what way? It seems like the stock price only matters to Tesla if they want to raise money. But they have basically no debt and many billions of free cash flow, even with continuing to build out GT, GB and expand Fremont. I don’t yet understand how stock price has any meaning for Tesla.
Dana: Companies with lots of cash can grow when / as they need, re applying resources. Companies without debt and able to pay their bills CANNOT go bankrupt by definition.
So I very much agree with you. Keeping a strong financial position and flexibility is FAR more important than propping up TSLA a little in the short term, at the expense of burning a LOT of cash they could use for factories, mining, research, unforseen issues (like managing a REALLY bad recession for years, if needed), etc.
It doesn’t really matter over an investment timeframe for long term unleveraged TSLA holders whether the bottom is 170ish or 140ish or even 100ish. It DOES matter to Tesla if for some reason, they can’t strike at the right time to build a given factory or do the right investment for Optimus, etc.
For one thing, despite the claims there is no competition, BYD is a blatant example that competition DOES exist for Tesla, and more will arrive over time (even though lots will die as well).
@Roger Geyer 100%
Tesla in unsurpassed in its ability to invest capital. And future capital needs to achieve Tesla goals are huge.
I see how you made this common misinterpretation of Tesla’s excess cash. Tesla is spending all the money that they can efficiently spend. The cash they would use for a stock buyback won’t impact projects or emergency funds.
Interesting possibility of shadow mode A/B test between FSD and AP.
Here are some timestamps for each release notes
1 – 22:01 – – Enabled FSD Beta on highway. This unifies the vision and planning stack on and off-highway and replaces the legacy highway stack, which is over four years old. The legacy highway stack still relies on several single-camera and single-frame networks, and was setup to handle simple lane-specific maneuvers. FSD Beta’s multi-camera video networks and next-gen planner, that allows for more complex agent interactions with less reliance on lanes, make way for adding more intelligent behaviors, smoother control and better decision making.
2 – 37:32 – Improved Occupancy Network’s recall for close by obstacles and precision in severe weather conditions with 4x increase in transformer spatial resolution, 20% increase in image featurizer capacity, improved side camera calibration, and 260k more video training clips (real-world and simulation).
3 – 40:31 – Improved merging behavior by leveraging lane geometry and lane bounds, association with coarse map information and better gap selection algorithms, allowing for a smoother and safer experience.
4 – 49:42 – Added highway behavior to offset away from blocked lanes and generic obstacles like road debris while also adding a smooth handoff between in-lane offsetting and lane changing.
5 – 53:02 – Reduced sensitivity for speed-based lane changes in CHILL mode.
6 – 53:56 – Improved lane changes to allow higher jerk maneuvers if required to stay on-route or to move away from lane blockages.
7 – 55:05 – Improved smoothness at highway lane splits by being less strict about centering between lane lines and allowing lower jerk maneuvers, where safe to do so.
8 – 58:55 – Reduced latency of trajectory optimization by 20% on average, without sacrificing behavior, by leveraging numerical tricks for more efficient computations.
On the Language discussion. . . My family emigrated from the Netherlands to Canada when I was just short of 7 years old. The first few months of exposure to English, in Grade 2, was excruciating. But then, one day, I realized that I was “thinking in English”. At age 23, when I was married, I was able to write my Dutch grandmother a thank you note for the wedding gift. Sadly, since then, I’ve lost almost all my Dutch. IN school, we were taught French, from Grade 2 to 11, 1 hour per day.
None of it has stuck.
Dewiz: The reality for adults for almost any complex learning is that if it sits disused for a long time, it fades, often to just concepts vs. usability. It’s an unfortunate fact of life. The internet and Google help re convenience in brushing up and fact checking (to ensure veracity of memory) conveniently.
Imagine life before books — it’s a wonder humanity made much progress at all.
Jammer voor je, makker! 😉
On the learning topic, it would be great to have an AI tutor that knows after a few lessons what you are interested in and starts talking to you in your skill level and makes suggestions and proposals to what to learn or talk about. Like the friend you talked about that you had in that korean school. If AI could be your debating buddy that would be so awesome.
Dave, I love the idea of using AI for teaching. Ideal methods of teaching / unique pace of learning per child / individual engagement
Dave it would be awesome to get time stamps for these long videos
Actually, I was even more interested in the second topic than the first one. Learning new languages is a bit of a hobby of mine. And it’s almost a shame we only hear James speak about FSD when he’s clearly a very interesting person who we can learn a lot from in general.
And Dave, if in the future I could in any way help with feedback for your language app project as a multilingual European, you’re very welcome to contact me. It sounds super interesting.
For learning Latinic languages, Michel Thomas is useful because he highlights the commonalities between them which immediately builds vocabulary pretty rapidly.
It seems like one stack with all features, then turning off features you haven’t paid for seems like the long term solution. A release would increase all the updates and by your purchase history they would be included in your driving experience.
the trouble with snow driving is knowing the conditions. Even when snow is around, people drive normally when it’s safe. But humans can recognize quickly when behavior needs to change, like when ice is likely to be present.
Language learning: Sounds like a family movie night with a foreign language film with subtitles would be in order.
Then as an activity ask the kids what scene they would like to re-enact and video them renacting the scene. Let them watch their re-enactment and the orginal scene over and over and give them a chance to critique their own performance and re-do the scene. It might be a scene where a character sings a song.
I also like the idea of using Siri, Alexa or Google in a foreign language.
Your drive is awesome to see James. I like you a lot and you are a role model for many of us! Thanks man.
I am bilingual in Japanese and English, and I’m also a technical translator. What I can personally add is this. Kids’ ability to pick up different sounds they’re not regularly exposed to disappear around age 13, so make sure they’re exposed to the sounds of the target language(s) before 13. I started learning English around 13, so to this day, I cannot always differentiate some English sounds. But what’s most important is to make it FUN and keep it fun. I learned English through movies and songs I loved. Also, make friends with native speakers. Good lock!
I can listen to you guys for hours. 🙂
Thank you Dave and James. 😘