The Onslaught of Tesla News Continues and Tesla Stock Soars (TSLA) ⚡️

Welcome to Electrified, it's your host Dylan Loomis. So first up today, Lex uploaded his interview with Walter Isaacson and we got some interesting takeaways. They talked a lot about the pain and the rough childhood that Elon had growing up. They shared anecdotes like when Elon was beat up pretty badly as a kid he was hospitalized for a week and then when Elon got home his dad, Errol, then berated him for more than an hour taking the side of the bully.

欢迎来到Electrified，我是你们的主持人Dylan Loomis。今天首先要说的是，Lex上传了他与Walter Isaacson的采访，我们得到了一些有趣的见解。他们谈了很多关于埃隆成长过程中所经历的痛苦和艰难童年的话题。他们分享了一些趣闻，比如埃隆小时候曾经被打得很惨，住进了医院一个星期，回到家后他的父亲Errol却在一个多小时里责骂他，站在欺凌者的一边。

Later in life there were some anecdotes shared like around the Tesla verge of bankruptcy times in 2008 when Elon's former wife said that he was so stressed he would vomit at times. And Walter said it's some of Elon's childhood trauma that he's actually channeling and using as extra motivation and drive. Paired with the way Elon is wired which is he's addicted to risk and adventure and intensity which is again why he gets so much stuff done.

后来在生活中有一些轶事被分享，比如在特斯拉在2008年濒临破产的时候，埃隆的前妻说他压力大到有时还会呕吐。沃尔特说，这些都是埃隆童年时期的一些创伤，他实际上在利用它们作为额外的动力和推动力。再加上埃隆的性格使然，他对风险、冒险和激烈的追求上瘾，这也是他能够完成那么多事情的原因。

And one deep note they were talking about how Elon's empathy toward humanity is greater than empathy toward individuals. This gives Elon the ability to be a tough boss and to fire people understanding it's for the greater good of the mission. Walter said quote Elon is driven by mission more than any person I've ever seen and it's not only missions it's like cosmic missions. I see it all the time when I'm walking the factory lines with him or in product development where he'll look at say the heat shield under the Raptor engine of a starship booster. And he'll say why does it have to be this way couldn't we trim it this way or make it or even get rid of this part of it. And he can visualize the material science isn't small anecdotes in my book but at one point he's on the Tesla line and they're trying to get 5,000 cars a week in 2018. It's a life or death situation and he's looking at the machines that are bolting something to the chassis and he insists that Drew Bagley not Drew, that Lars Maravy one of his great lieutenants come and they have to summon him and he says why are there six bolts here and Lars and others explained well for the crash tests or anything else the pressure would be in this way so you have to and they were blah blah blah blah. And he said no if you visualize it you'll see if there's a crash it would the force would go this way and that way and it can be done with four bolts. Now that sounds risky and they go test in the engineer but turns out to be right. I know that seems minor but I could give you 500 of those where in any given day he's visualizing the physics of an engineering or manufacturing problem that sounds pretty mundane but for me if you say what makes him special there's a mission driven thing I give you a lot of reasons but one of the reasons is he cares not just about the design of the product but visualizing the manufacturing and of the product the machine that makes the machine.

有一次，他们谈到了埃隆对人类的同理心大于对个体的同理心。这使得埃隆有能力成为一个严格的老板，并解雇员工时明白这是为了整个使命的利益。沃尔特说：“埃隆对使命的驱动力超过了我所见过的任何人，这不仅仅是使命，而是宇宙使命。”我在和他一起走工厂生产线或者产品开发时经常看到这一点，他会看着像星舰助推器上的Raptor引擎下的热防护盾，然后问为什么一定要这样，我们能不能改一下这个，甚至去掉这个部分。他可以用心理图像的方式来看待材料科学，在我写的书中，这并不是小事，但有一次他在特斯拉生产线上，他们试图在2018年每周生产5000辆汽车，这是一场生死攸关的情况，他看着那些螺栓扣住底盘的机器，坚持让德鲁·贝格利（不是德鲁，是他的一位重要副官拉斯·马拉维）来，并且必须召唤他，他问为什么这里有六个螺栓，拉斯和其他人解释说：因为进行碰撞测试或其他情况下，压力会这样传导，所以必须得这样，然后blah blah blah blah。他说不，如果你想象一下，你会发现在发生碰撞时，压力会这个方向和那个方向传导，只需四个螺栓就可以了。现在听起来有风险，他们就让工程师做了测试，结果证明他是正确的。我知道这似乎微不足道，但我可以给你500个类似的例子，就在任何一天，他都在用心理图像的方式来看待工程或制造问题的物理学。对我来说，如果你问他有什么特别之处，除了他对产品的设计关注之外，他还关注产品的制造过程和制造的机器，这就是让他与众不同的原因之一。

I think you're also right on 50-100 years from now what Musk will be most remembered for besides space travel is real world AI not just optimist the robot but optimist the robot and the self-driving car they're pretty much the same they're using GPU clusters or dojo chips or whatever it may be to process real world data. We all got and you did on your podcast quite excited about large language model you know generative predictive text AI. That's fine especially if you want to chat with your chat bot but the holy grail is artificial general intelligence and the tough part of that is real world AI and that's where optimist the robot or full self-drive or I think far ahead of anybody else.

我认为在未来的50到100年里，除了太空旅行之外，人们最多会记住马斯克的是他对现实世界人工智能的贡献，这不仅仅是对机器人的乐观态度，还包括机器人和自动驾驶汽车，它们本质上是一样的，都使用GPU集群或者道场芯片来处理现实世界的数据。我们都对大型语言模型感到兴奋，你在播客中也是如此，你知道生成性预测文本人工智能。这很好，特别是如果你想与你的聊天机器人进行对话，但真正的目标是人工通用智能，而其中困难的部分是现实世界的人工智能，而在这方面，乐观的机器人或者完全自动驾驶系统远远领先于其他任何人。

Since a few years ago when you could see with Andrei Karpathy the software 2.0 evolution of autopilot it became obvious to me that this is not about the car this is about optimist the robot this is like if we look back 100 years from now the car will be remembered as a cool car nice transportation but the autopilot won't be the thing that controls the car it will be the thing that allows embodied AI systems to understand the world so broadly so that kind of approach and you kind of stumble into it will Tesla be a car company will be an AI company will be a robotics company will be a home robotics company will be an energy company and you kind of slowly discover this as you confidently like push forward with a vision which one will Tesla become well probably all of the above.

自几年前当您与Andrei Karpathy一同看到了自动驾驶2.0软件的发展，我就明白了，这并不仅仅关乎汽车，而是关乎机器人的进步。就好像从现在回望100年前的汽车，我们会觉得它是一辆酷车和便利交通工具，但自动驾驶并不仅仅是控制汽车的一项技术，它更是使实体化的人工智能系统能够更全面地理解世界的关键。所以，你开始思考特斯拉会成为一家汽车公司、AI公司、机器人公司、家庭机器人公司，还是能源公司。随着你坚定地推进着这个愿景，你会逐渐发现特斯拉将成为以上所有类型的公司。

The full conversation will be below over the weekend thanks to the updated user manual for the new model 3 as we can see it looks like there are no airbags for the knee area in the new version. I do want to add this is not a guarantee for the north american version that still may be slightly different in different aspects either way I did just want to pass along this article from the IIHS some of their researchers have found a little benefit from these knee airbags they found not only do they have a negligible effect on injury risk but they may even increase it in some cases.

完整的对话将在本周末发布，感谢更新的新款Model 3的用户手册。从我们可以看到的情况来看，新版本中似乎没有膝盖区域的气囊。我想补充一点，这并不是对北美版的保证，因为在某些方面可能会略有不同。无论如何，我只是想传达一下来自IIHS的这篇文章，他们的一些研究人员发现膝盖气囊对减少受伤风险几乎没有影响，甚至在某些情况下可能会增加受伤风险。

To everyone asking how the hazards will work when the trunk is open on the new model. 3 well here is your answer with the lights integrated to the trunk now they have the lights down here below.

对于所有询问新型号车尾箱打开时灯光会如何工作的人们。好的，这里就是你们的答案，现在这些灯光被整合到了车尾箱的下方。

We have a new questionable rumor from car news china according to a Tesla salesperson Tesla might increase the model y price in china next month they said the price hikes could be between 1100 and 1400 us dollars and that Tesla might even cancel the rear wheel drive variants altogether but here's why it's questionable if you go on to read the rest of the article they're also saying that the model y already has everything the new version of the model 3 added this is not true and if you need one example interior ambient lighting is not on the model y for that reason I wouldn't put too much stock into this report but it's certainly possible.

我们从中国汽车新闻中得知了一个新的令人怀疑的传闻，根据一位特斯拉销售员的说法，特斯拉可能会在下个月提高中国市场的Model Y价格，他们表示价格上涨可能在1100到1400美元之间，而且特斯拉甚至可能会取消后驱版本。但这个传闻存在疑问的原因在于，如果你继续阅读文章的剩余部分，他们还说Model Y已经拥有了Model 3新版本的所有功能，这是不正确的，举个例子，Model Y没有内部环境照明。因此，我不会过多关注这个报道，但这种可能性是存在的。

We got the Tesla production number for august and it was 80,832 essentially in line with july more importantly looking at the production breakdown between model 3 and why model 3 production still relatively on par with where it's been albeit slightly decreased coming in at 21.3 thousand for the month of august we'll see what september looks like but it looks like gigashang high is doing a pretty incredible job of maintaining the model 3 production all the way through this highland transition because we know that there are now parking lots filling up with new model threes for export from gigashang high.

我们得到了8月份的特斯拉产量数据，总共为80,832辆，与7月份基本持平。更重要的是，我们看到了Model 3和为什么Model 3的产能仍然相对稳定，虽然稍微有所下降，但在8月份依然达到了21.3万辆。我们将看看9月份的情况，但看起来gigashang高级制造是在这个高产能转换阶段中维持Model 3的产能的一个了不起的工作，因为我们知道现在停车场里正填满了从gigashang高级制造出口的新Model 3车辆。

We got another excerpt from walter isakson's upcoming biography on elon musk and everybody saying it's ruining the book just remember the books about 700 pages and we're only getting a few pages of a preview in this one we got some more context on the bill gates elon musk situation in 2022 they were at a meeting together and bill gates said hey i'd love to come see you and talk about philanthropy and climate fast forward to march 9th of last year and bill gates landed to visit gigatexas then when elon started to give bill gates atora the factory gates argued batteries would never be able to power large semi-trucks and that solar energy would not be a major part of solving the climate problem bill gate said i showed him elon the numbers it's an area where i clearly knew something that he didn't despite this gates found himself impressed by the factory elon had built and his detailed knowledge of every machine and process most of us know about gates shorting tesla stock and by the time he arrived at the factory he had lost about 1.5 billion dollars when elon found out about the situation he was obviously upset bill gates said i did apologize to him to elon but once he heard i'd shorter the stock he was super mean to me but he's super mean to so many people so you can't take it too personally elon said how can someone say they're so passionate about fighting climate change and then do something that reduced the overall investment in the company doing the most it's pure hypocrisy why make money on the failure of a sustainable energy car company after the visit bill gate sent a text to elon trying to talk about some of his philanthropic ideas for elon elon's response to that text do you still have a half a billion dollar short position against tesla to which gate said yes and then elon said sorry i can't take your philanthropy on climate seriously when you have a massive short position against tesla the company doing the most to solve climate change reportedly bill gates was puzzled about why elon was upset about him shorting tesla which is odd then right after this exchange elon texted walter saying at this point i'm convinced he bill gates is categorically insane i did actually want to like him side on x elon said gates also said the tesla semi was impossible even though it was literally being driven all over the country when i asked what battery watt hour per kilogram and truck watt hour per mile he was using to reach the conclusion that it's impossible gates had no idea but still stuck with his conclusion.

我们得到了沃尔特·艾萨克森即将出版的关于埃隆·马斯克的传记的另一个摘录，大家都说这本书被破坏了，但请记住这本书大约有700页，而我们只能看到几页的预览。在这其中，我们对2022年比尔·盖茨和埃隆·马斯克的情况有了更多的背景认识。他们当时在一个会议上，在会上盖茨说：嘿，我愿意去看你，并谈论慈善和气候问题。快进到去年3月9日，盖茨来到了吉加德州参观，然后埃隆开始给盖茨展示工厂。盖茨争辩说电池永远无法为大型半货车提供动力，并且太阳能不会成为解决气候问题的重要组成部分。盖茨说：我向他展示了数据，这是一方面我明显掌握了比他更多信息的领域。尽管如此，盖茨还是对埃隆建造的工厂以及他对每台机器和过程的详细了解感到印象深刻。我们大多数人都知道盖茨做空特斯拉股票，在他到达工厂时，他已经损失了约15亿美元。当埃隆得知这一情况时，显然很不高兴。盖茨说：我向他道了歉。但是，一旦他听说我做空了股票，他对我很刻薄。不过，他对很多人都很刻薄，所以你不能太过于个人化。埃隆说：一个人怎么能说他非常热衷于对抗气候变化，然后做一些减少对这个做出最大贡献的公司的投资的事情，这纯属虚伪。为什么要在一个可持续能源汽车公司失败上赚钱呢？参观结束后，盖茨给埃隆发了一条短信，试图谈论他的一些慈善想法。埃隆对这条短信的回应是：你还是有对特斯拉空头头寸吧？盖茨回答是肯定的，然后埃隆说抱歉，当你对特斯拉这家公司做出最大努力解决气候变化问题的时候，我无法认真对待你的气候慈善。据报道，比尔·盖茨对埃隆对他做空特斯拉感到困惑，这很奇怪。就在这次交流之后，埃隆给沃尔特发了一条短信说：“现在我确信他比尔·盖茨是彻头彻尾的疯子。我真的想喜欢他。”埃隆还说盖茨说特斯拉的半货车是不可能的，尽管这些半货车实际上正在全国各地行驶。当我问他是用什么电池瓦时每公斤和卡车瓦时每英里来得出这个结论时，他一脸茫然，但仍然坚持他的结论。

Here we have an unlikely source cbs sunday morning to potentially leak part of the design for the upcoming robo taxi so first we have this cardboard design as you can see it appears to have one row of seating potentially for two people with ample legroom easy ingress and egress the ability to get in and out and it looks like right behind the seats there would be some room for suitcases and storage.

在这里，我们有一个不太可能的消息来源CBS Sunday Morning，可能会泄露即将推出的机器人出租车的部分设计。首先，我们有这个纸板设计，如您所见，它似乎有一排座位，潜在容纳两人，舒适的腿部空间，方便的上下车，能够进出，而且看起来在座椅后面有一些行李和储物空间。

I've also seen a lot of questions about why they would have a robo taxi with only two seats don't forget the rest of tesla's fleet the s3 xy will also be able to operate as robo taxis and will work just fine for families or people that want to travel in groups bigger than two so think about the millions of instances just one person or a couple will need a ride and then tesla will be able to make smaller more affordable vehicles use less batteries they'll be lighter and more efficient all of those benefits but you may notice something over here on the left and now you can't see these very well. but it looks like this could be an early prototype of the robo taxi cyber futuristic field design and i think it's this image that we're leading some people to believe it could be a three wheel vehicle i do think that's unlikely although yes it would lead to things like lower rolling resistance but i will say i'd be surprised if tesla were to go the aptara route then up top it looks like this could be a shot of the rear

我也看到了很多关于他们为什么只有两个座位的机器人出租车的问题，不要忘了特斯拉的其他车队S3 XY也可以作为机器人出租车运营，对于想要以大于两个人的团体旅行的家庭或个人来说也完全适用，所以想象一下有数百万次只有一个人或夫妻需要搭车的情况，然后特斯拉将能够制造更小、更实惠的车辆，使用更少的电池，它们将更轻更高效。所有这些好处都是显而易见的，但你可能会注意到左侧的一些东西，你现在看不太清楚。但它看起来像是机器人出租车的早期原型，具有未来主义的设计感，我认为正是这张图让一些人认为它可能是一辆三轮车，尽管是这样，它会导致更低的滚动阻力等问题，但我想说如果特斯拉选择走aptara的路线，我会感到惊讶。然后上面看起来像是后部的一张照片。

I do want to say though if this really was a design for the robo taxi this could have been changed a hundred times by now. We know Tesla has a huge library of drawings, sketches, and designs like this that never ultimately stick. But again, hopefully sometime next year, we get an official unveiling event for the robo taxi and the twenty-five thousand dollar Tesla.

不过，我确实想说的是，如果这真的是专为机器人出租车设计的方案，它可能已经被修改了一百次了。我们知道特斯拉有一个巨大的图纸、草图和类似这样的设计库，最终并不会采用。但是再次希望，明年我们能够看到机器人出租车和价值两万五千美元的特斯拉的官方发布活动。

Replying to k-10 Elan did say he has a copy of the book but Walta recommended that he not read it over the weekend. Alex said a VW RND board member said that VW makes no money with BEVs and is slowing down the introduction of new BEV models. Saying "for us it's important not to overwhelm customers by moving too quickly into e-mobility. That's why our combustion engines make sense." Saying quote "with competitors, you see that the focus on low-cost electric cars means they don't make money. The transformation is a slow process." And he said, "Along the way, the internal combustion engine has a right to exist, which is why we are once again investing heavily in technology with the Tiggy One and the Pasote. This is necessary to bring the masses along on the road to electromobility."

回复k-10，埃兰确实说他有这本书的副本，但瓦尔塔建议他不在周末阅读。亚历克斯说，大众汽车（VW）RND董事会成员表示，VW在纯电动汽车（BEV）上不赚钱，并且正在减缓新型BEV车型的推出。他表示：“对我们来说，重要的是不要通过过快进入电动出行而让顾客不知所措。这就是为什么我们的内燃机是合理的。”他还说：“与竞争对手相比，你会发现低成本电动汽车的关注意味着它们不能赚钱。转型是一个缓慢的过程。”他还表示：“在这个过程中，内燃机有存在的权利，这就是为什么我们再次在Tiggy One和Pasote技术上大量投资的原因。这是推动大众踏上电动出行之路所必需的。”

Replying to Alex's post, Doctor Electron X said, "The goal of most auto manufacturers is to figure out how to wind down their businesses. I've been saying this about GM for a while, others are starting to follow GM's lead. The words they say to the media is to keep stockholders from fleeing all at once." To which Elan said, "If they don't switch to autonomous electric cars pronto, that is absolutely true. It may be a few years, but the electric and autonomous revolution will undoubtedly put some legacy car makers into the EV autonomous graveyard."

在回复Alex的帖子时，电子医生X说：“大多数汽车制造商的目标是如何逐渐减少他们的业务。我一直在说这个关于通用汽车已经有一段时间了，其他人也开始效仿通用汽车的做法。他们对媒体所说的话是为了避免股东们全部一起逃离。”对此，埃兰说：“如果他们不迅速转向自动驾驶电动汽车，那绝对是真的。可能还需要几年的时间，但电动和自动驾驶的革命无疑会导致一些传统汽车制造商进入电动自动驾驶汽车的墓地。”

Recently at the Fully Charged event, Sandy Munro saw a VW logo pointed at it and said that they could be in the EV graveyard. Next to that point, we have this: one of the top executives at VW was dispatched to China to review the competitive landscape and his assessment was grim. He told the VW CEO that Europe's largest car maker was losing the electric vehicle race in its most important market and had no prospect of catching up on its own. Echoing Elan's words, VW risks getting confined to the wane market for combustion engine cars and lacking the volumes to support its bloated structure.

最近在“Fully Charged”活动上，Sandy Munro看到大众（VW）标志指向它，并表示它们可能会成为电动汽车墓地。在此之际，我们还有这样一个观点：大众的其中一位高管被派往中国审查竞争环境，他的评估结果是令人沮丧的。他告诉大众CEO，在其最重要的市场上，欧洲最大汽车制造商正在输掉电动汽车竞赛，并且无法凭借自身力量赶上。与Elan的观点相呼应，大众有被限制在燃烧发动机汽车的衰落市场，并且缺乏支撑臃肿结构的销量的风险。

For the last five years or so, we've heard how VW was going to catch Tesla in EV sales but looking at quarter two 2023, Tesla in the black, 466,1000 BEV sales, VW still sits at 145,000. So as of quarter two this year, Tesla is still more than tripling up VW in terms of full BEV sales globally. The boss of VW China said, "It's a marathon and not everyone in a marathon who comes out fastest will be seen first or at all at the finish line."

在过去的五年左右时间里，我们一直听说大众汽车打算在电动汽车销量上赶超特斯拉，但是看看2023年第二季度的数据，特斯拉实现盈利，电动汽车销量达到了466,100辆，而大众汽车仅为145,000辆。所以截至今年第二季度，特斯拉在全球纯电动汽车销量方面仍然是大众汽车的三倍以上。大众汽车中国的负责人说：“这是一场马拉松比赛，并不是每个跑得最快的人都会在终点线上第一个或者能被看到。”

We're already seeing some of these last-ditch type efforts from VW, like paying 700 million dollars for about 5% of e.paying. And that e.paying deal was preceded by new cooperation agreements with a battery maker and autonomous driving specialist and an infotainment developer. Going back to that first quote, that VW's own executive does not think they can catch up at all on their own. What does that mean? Well, ultimately it adds a lot more complexity for VW and takes away a lot of those profit margins. So, will Sandy be right again with Tesla stock up over 10% today?

我们已经看到大众汽车进行了一些最后一搏的努力，比如花费7亿美元购买大约5%的e.paying股份。而在这个e.paying交易之前，他们还与一家电池制造商、一家自动驾驶专家和一家信息娱乐系统开发商签订了新的合作协议。回到第一句引述，也就是大众汽车自己的高管都认为他们无法单独迎头赶上。这意味着什么呢？嗯，这最终给大众汽车增加了很多复杂性，并减少了很多利润空间。那么，今天特斯拉股价上涨超过10%，桑迪会再次正确吗？

At least part of it could probably be attributed to this from Adam Jones and Morgan Stanley. They put out a new Tesla stock research note highlighting the benefits of Tesla Dojo, and they've upgraded their Tesla price target to 400. This would be about a 60% increase from their prior price target of 250 dollars per share. I believe this is now the highest Tesla stock price target on the street, and this would give Tesla a valuation of about 1.4 trillion.

至少一部分原因可能可以归功于亚当·琼斯和摩根士丹利的这一点。他们发表了一份新的特斯拉股票研究报告，强调了特斯拉Dojo的好处，并将其特斯拉股票价格目标提高到400美元。这将比他们之前的每股250美元的价格目标增长约60%。我相信这现在是华尔街上特斯拉股票价格目标最高的，这将使特斯拉的估值约为1.4万亿美元。

Honestly though, guys, all this note really is is Wall Street waking up to what many of us have now been talking about for months and years. He said, "We believe Dojo can add up to 500 billion dollars to Tesla's enterprise value, expressed through a faster adoption rate in mobility, robo taxi, and network services software as a service."

老实说，伙计们，这个笔记其实只是华尔街对我们许多人数月甚至数年来一直在谈论的事情的觉醒。他说：“我们相信Dojo可以为特斯拉的企业价值增加多达5000亿美元，表现在移动性、机器人出租车和网络服务软件方面的更快采用率上。”

The summary: With a highly experienced semiconductor team, Tesla has built a custom AIA sick chip that due to its core function of processing vision-based data for autonomous driving, it can operate more efficiently than the leading cutting-edge general-purpose chips on the market and at a fraction of the cost. Tesla pursues high vertical integration and key technology domains to enable high iteration and continual improvement while helping to diversify away from over-reliance on third-party suppliers that may not be able to provide an optimal solution for Tesla's specific needs.

摘要：凭借经验丰富的半导体团队，特斯拉构建了一款定制的AIA芯片，该芯片的核心功能是处理基于视觉的自动驾驶数据，相比市场上领先的尖端通用芯片，它能够以更高效的方式运行，并且成本只需一小部分。特斯拉追求高度垂直整合和关键技术领域，以实现高迭代和持续改进，同时帮助其摆脱对第三方供应商的过度依赖，这些供应商可能无法提供适合特斯拉特定需求的最优解决方案。

The more we looked at Dojo, the more we realized the potential for under-appreciated value in the stock. We believe Tesla can reasonably test its all-time highs of 400 over the next 12 months.

我们越深入研究道教，就越意识到这只股票被低估的价值潜力。我们相信特斯拉在接下来的12个月内有望合理测试其历史最高价400美元。

If you're not new to the channel, everything I just said really should just be a reiteration of what we've all known. They talk about Dojo changing things at Tesla going from targeting just automotive to being multi industry training speed going from months to weeks capital expenditure going from high single digit billions down to low single digit billions again all things that we already know but morgan saintly did say that they raised their bull case valuation to five hundred and fifty dollars up from 450 before they said we believe dojo can represent the next step change in market perception of tesla longer term we see scope for dojo to provide avenues for tesla software and hardware capabilities to extend well beyond the auto industry if dojo can help make cars see and react what other markets could open up with their calculations when comparing what tesla would have to spend on equivalent compute from in vidio dojo has the potential to drive six and a half billion dollars in cost savings for tesla over the next few years then despite those calculations they said we have not given tesla credit for specific cost savings from dojo versus current super computing budget nor have we given tesla credit for any non-auto related revenue streams they also made no changes to their assumptions for tesla energy or tesla insurance and finally they said dojo is designed to process visual data which can lay the foundation for vision based ai models like robotics health care and security in our view once tesla makes headway on autonomy and software third party dojo services can offer investors the next leg of tesla's growth story replying to this note elon said almost all of tesla's value long term will be from ai and robots both vehicle and humanoid believe it or not

如果你对这个频道并不陌生，我刚才说的一切只是对我们所知的事情的重申。他们谈论到道场将特斯拉的目标从仅仅针对汽车行业转变为多行业培训，速度从数月降至数周，资本支出从高个位数十亿美元降至低个位数十亿美元，这些都是我们已经知道的事情。但是，摩根士丹利确实表示他们将牛市估值提高到550美元，之前为450美元。他们说我们相信道场可以代表特斯拉在市场感知上的下一个飞跃，从长远来看，我们认为道场可以为特斯拉的软件和硬件功能提供途径，远远超出汽车行业。如果道场可以帮助汽车看到并做出反应，他们的计算会在比较特斯拉与英伟达等公司在等效计算上的支出时，道场有潜力在未来几年为特斯拉节省65亿美元的成本。尽管有这些计算，他们表示我们对特斯拉因道场与目前超级计算预算的差异而节省的具体成本并没有给予信任，也没有考虑特斯拉的非汽车相关收入来源。他们对特斯拉能源或特斯拉保险的假设也没有做出任何变化。最后他们表示，道场被设计用于处理视觉数据，这可以为基于视觉的人工智能模型如机器人、医疗保健和安全奠定基础。在我们看来，一旦特斯拉在自动驾驶和软件方面取得进展，第三方的道场服务可以为投资者提供特斯拉增长故事的下一步。对于这个讨论，埃隆回复说，从长远来看，特斯拉几乎全部的价值都将来自人工智能和机器人，包括车辆和仿人。无论你信不信。

parts of wall street are clearly waking up to the real tesla story bit by bit in the past five years evs have gone from 2% of new car sales in california to 22% and i just wanted to show you this chart if california were a country it would be the fourth largest in terms of quarter to ev sales this year so here we have the top 10 global ev markets and we can see their ev share of new cars of course norway still leading the way but california should be leading the way and a sign of things to come for the rest of the united states especially once we cross that 10 threshold

华尔街的一些人正在逐渐认识到真实的特斯拉故事。在过去的五年里，电动汽车在加利福尼亚的新车销量中所占比例已经从2%上升到22%。我只是想给你展示一下这张图表。如果加利福尼亚是一个国家，以该季度的电动汽车销量来计算的话，它将成为全球第四大市场。下面是全球电动汽车市场的前十名，我们可以看到它们在新车销量中的电动汽车份额。当然，挪威仍然领先，但加利福尼亚应该成为引领者，并且这将成为美国其他地区未来趋势的预兆，尤其是当我们突破10%的阈值之后。

this one was actually from gm authority just showing us the luxury ev cuv sold in the quarter two this year in the united states 108.7 000 vehicles total 105.5 000 of which were the tesla model y the Cadillac lyric in second place at a distant 1.3 000 from car sales in australia the model y has been crowned the winner of best electric car for 2023 they narrowed down the field to six new or upgraded finalists price under 89.3 000 dollars the judges then analyze these vehicles based on 100 criteria across five areas safety tech comfort convenience driving performance and ownership cost the model y came up trumps after solid scores in all assessment areas the model y is also the most searched electric car on car sales.com.au followed by the model three hydrag zane gave us an interesting scoop on an update for the tesla semi-production and the giga navada expansion we do know most of the megapack production has now been moved out of giga factory navada into lay throat

这是从gm authority来的消息，展示了今年第二季度在美国销售的豪华电动跨界车（EV CU）共有108.7万辆车，其中105.5万辆是特斯拉Model Y，而凯迪拉克Lyriq以遥远的1.3万辆排名第二。在澳大利亚的汽车销售中，Model Y被评为2023年最佳电动汽车的冠军。他们将候选车辆限定在价格低于8.93万美元的六款新车或升级车型上，评委们根据安全、科技、舒适性、便利性、行驶性能和拥有成本等五个方面的100项评判标准分析这些车辆。经过在所有评估领域都获得扎实分数后，Model Y脱颖而出。Model Y也是car sales.com.au上搜索量最多的电动汽车，其次是Model 3。Hydrag Zane给我们带来了关于特斯拉半挂卡车生产和Giga Nevada扩张的有趣消息。我们了解到，目前大部分Megapack的生产已经从Giga Nevada工厂转移到了Lathrop。

i have a high degree of certainty that with this information i'm about to share first week of august or so in the midnight hour the cover of darkness 17 semis arrived at the tesla giga factory navada and they brought in production equipment into these doors we're about to see 17 full trucks full truck loads trailer loads of semis and the where you see the three the two trailers backed up to the three loading doors and we'll get back to it here in a second and the wide door a lot of semi a lot of production equipment went in there and the rumor is is that it's for a semi-production sort of another prototype production line but maybe more automated semi-production has been moved up to the compound i don't believe they're starting it i believe they're waiting for the new version and then they will so we're getting closer to actual semi-production from giga navada because to date the pilot tesla semi-line is not actually in this building but i believe it's within a few miles of this building and sometime soon we should actually begin to see construction at this site where they expand this facility to allow for that semi-production

我非常确定，在这些信息即将揭示的情况下，八月初的午夜时分，17辆卡车抵达了特斯拉终装厂内华达，并且他们把生产设备带入了即将见到的这些门。我们将看到17辆装满车的卡车和三个装载门后倒退的两辆拖车，稍后我们将回到这点上。那个宽敞的门有很多半拆车，大量的生产设备被送入其中。传言是这是为了生产半拆车的另一种原型生产线，但可能会更加自动化。半拆车的生产已被提前移至该基地，我不相信他们是在开始生产，我相信他们在等待新版本，然后才会开始。所以我们离终装厂实际生产半拆车越来越近了，因为迄今为止，试制的特斯拉半拆车生产线实际上并不在这座建筑物内，但我相信它就在这座建筑物的几英里之外，很快我们应该会看到这个工地开始扩建设施，以容纳半拆车的生产。

here we have another excerpt from the Walter Isaacson book this devol shroff is a member of tesla's autopilot team and by the early 2023 the neural network planner project the end to nv12 had to analyze 10 million clips of video collected from the cars of tesla customers and yes tesla is still using human labelers many of them based in buffalo new york and they assess the videos and give them grades more on this in a second shroff had been showing elon instances in where a neural network planner nt nv12 would work better than a rules based approach they had a demo littered with trash cans fallen traffic cones etc and he said here's what happens when we move from rules based to a network path based the car will never get into a collision if you turn this thing on even in unstructured environments elon got pumped about the idea and then said I want the latest data on miles per intervention to be the starting slide at each of our meetings and he told the team to make it like a video game where they could see their score every day so it would be motivating to watch each day as the miles per intervention increases so members of the team installed a massive 85 inch television monitor in their workspace that displayed in real time how many miles the fsd cars were driving on average without interventions they put a gong near their desks and whatever they successfully solved the problem causing an intervention they got to bang the gong

在这里，我们有来自Walter Isaacson的另一段摘录。这个Devon Shroff是特斯拉自动驾驶团队的成员，到2023年初，神经网络规划师项目NV12必须分析来自特斯拉客户车辆的1000万个视频剪辑，并且是的，特斯拉仍在使用人工标注者，其中许多人位于纽约州的布法罗，并对视频进行评分。稍后会更详细介绍这一点。Shroff曾向埃隆展示了神经网络规划师NV12相较于基于规则的方法更好的案例，演示中有垃圾桶、倒下的交通锥等。他说，当我们从基于规则的方法转为基于网络路径时，即使在非结构化环境下，汽车也永远不会发生碰撞。埃隆对这个想法非常兴奋，然后他说，我希望每次会议开始时都能看到最新的干预里程数据，并告诉团队让它像一个视频游戏，他们每天都能看到自己的分数，这样每天看着干预里程增加会有动力。于是团队成员在工作区安装了一台巨大的85英寸电视显示屏，实时显示FSD车辆平均无干预行驶了多少英里，他们在桌子附近放了一个锣，每当成功解决导致干预的问题时，他们就可以敲锣。

in a conversation about this end to end model the team found that the neural network did not work well until it had been trained on at least a million video clips luckily on that point james downma just talked to herbert full video will be linked below and james said that as the system gets better the number of clips that tesla needs to train on will be fewer and fewer it just becomes a game of finding the highest quality and curating what data tesla is actually using to train the neural nets and what's cool with the end-to-end system according to james is let's say tesla is trying to solve for one specific stoplight situation well in those videos the neural network is still going to learn from everything else in the scene that's going on not just that one little instance james did clarify that v12 is being built on top of a lot of the framework from v11 and prior version so no tesla has not thrown away everything that it did before v12 is built on a legacy of the system that tesla had but they've impressively found a way to take out most of the remaining code and glue all of the networks together so now the neural nets can begin to solve problems that engineers couldn't solve on their own and james did say the competition can't start out end-to-end you can only get there incrementally like tesla has done

在关于这种端到端模型的对话中，团队发现神经网络直到受训于至少一百万个视频剪辑之后才能正常运行，幸运的是，詹姆斯·唐马刚刚与赫伯特通话全视频将在下面链接，詹姆斯说，随着系统的改进，特斯拉需要训练的剪辑数量会越来越少，只需要找到最高质量的剪辑，并进行精心筛选，从而让特斯拉实际使用来训练神经网络的数据更好。根据詹姆斯的说法，端到端系统的有趣之处在于，比如特斯拉正在尝试解决某个特定的红绿灯情况，那么在这些视频中，神经网络还将从场景中发生的其他事物中学习，而不仅仅是那个小小的瞬间。詹姆斯澄清了v12是在v11和之前版本的框架基础上构建的，并没有抛弃之前的所有工作，v12是建立在特斯拉现有系统的基础上的，但他们令人印象深刻地找到了一种方法来删除大部分剩余的代码，并将所有网络粘合在一起，所以现在神经网络可以开始解决工程师无法独立解决的问题。詹姆斯还表示，竞争对手无法一开始就采用端到端方法，只能像特斯拉一样逐步到达这一点。

And we know Tesla is planning to release shadow mode if they haven't already. That'll be pushed most likely to all cars to gather data even when those cars are driving manually. Pair that with Tesla hiring FSD drivers globally to then collect specific data for v12, and the data collection will be different with v12 focusing on gathering different types of video data. Because before, Tesla didn't care as much about what the driver was doing, they were just capturing the video because the manual code would ultimately control the output. So now Tesla needs to collect a new video database because now they need to train for the part that goes all the way to the end user controls what the human is doing in these situations. And now the human labelers that Tesla is still using, they're actually curating the dataset rather than labeling or approving the auto labeling of objects like before because again it's all about quality video data.

我们知道特斯拉计划推出影子模式，如果他们还没有的话。很可能这将被推广到所有车辆中，即使这些车辆是手动驾驶。再加上特斯拉全球招聘全自动驾驶(FSD)驾驶员来收集特定的v12数据，数据收集将与v12有所不同，重点是收集不同类型的视频数据。因为在此之前，特斯拉并不太关心驾驶员在做什么，他们只是捕捉视频，因为最终手动代码将控制输出。所以现在特斯拉需要收集一个新的视频数据库，因为现在他们需要为最终用户控制在这些情况下人类所做的部分进行训练。而现在特斯拉仍在使用的人类标注员，实际上是在策划数据集，而不是像以前那样标注或批准对象的自动标注，因为这一切都关乎优质视频数据。

You know, we'll see how they do with this stuff but I don't think they're anywhere near where Tesla is right now and I don't think the infrastructure they put in place or the compute they have in the vehicles is capable of doing it like they're going to have to come up with a different way to do it and they're going to have to prove it. It's a research project, it's not a term the crank thing. So if so, someone's asking a question, well you know how soon after Tesla does what they're doing well somebody else just do what they did okay how long will it take a competitor to put four million instrumented vehicles on the road and then run them for two years to gather all data and then train the neural network. Even if even if you knew everything Tesla is doing even if they published the source code and shared what they were doing, a competitor's got to build a fleet of vehicles, they got to get them out there, they got to run them for years because they got to turn the data engine crank to get the thing going. And then at the end of that, you know then they get results that the Tesla's doing now. It might be a little faster, you know three years from now, five years from now, maybe the computers are faster than that kind of stuff but you know how long does it take to build four million cars or a million cars and put all the instrumentation on them out there, that's not something that happens in a year right? That's not, that's probably not something that happens in three or four years given you know the build rates at a lot of the competitors. So you know it, the idea that someone is going to see what Tesla did and then they're going to go do it a year later, like that's not going to happen. Something else could happen, it's certainly possible, I don't think it's going to happen because I don't see any other players that are capable of putting a system in an OEM vehicle, a privately owned vehicle, and putting it out there. I don't see any of them building the infrastructure necessary to close the gap in less than, I don't know, three years, five years, something like that. I mean, I would say that that would kind of be the minimum and I'd be surprised that somebody did it five years after Tesla but it'll depend on how the technology moves. The tech moves really fast every year and surprising things happen. Tesla adapts to these surprises, everybody else adapts to the surprises, and some of the surprises they may, you know, they may open the gap with Tesla and some of them might close the gap. It's hard to say.

你知道，我们将看看他们如何处理这个问题，但我不认为他们与特斯拉现在的水平有任何接近，我也不认为他们所建立的基础设施或车辆中的计算能力能够像他们将要做的那样完成任务，他们将不得不找到另一种方式来完成，并且他们需要证明它的可行性。这是一个研究项目，不是一个急需解决的问题。因此，如果有人问一个问题，假设特斯拉完成了他们正在做的事情后，多久后会有其他公司也做类似的事情，那么竞争对手需要多长时间来将四百万辆配备有设备的车辆上路并运行两年以收集数据并训练神经网络？即使你知道特斯拉正在做的一切，即使他们公开了源代码并分享了他们正在做的事情，竞争者也必须建立一支车队，将其推出去，并运行数年，因为他们必须启动数据引擎以使之起作用。在此之后，他们将获得与特斯拉现在所做的相同的结果。可能会快一点，你知道，从现在开始三年后或五年后，也许计算机速度会快一些，但你知道要建造四百万辆汽车或一百万辆汽车，并在其中配备所有仪器，这不是一年的时间可以完成的，对吗？考虑到许多竞争对手的建造速度，这可能不会在三到四年内完成。所以你知道，认为有人会看到特斯拉的成就，然后一年后去复制，这是不可能的。可能会发生其他事情，这是有可能的，但我不认为会发生，因为我没有看到其他任何一个能够将系统装入原始设备制造商（OEM）的车辆并将其投放市场的人。我没有看到任何一个人构建起必要的基础设施来在不到-年、不到5年的时间里消除差距。我的意思是，那可能是最低限度的，我会惊讶地看到有人在特斯拉后五年完成这些任务，但这将取决于技术的发展。每年技术发展非常快，会有令人吃惊的事情发生。特斯拉适应这些变化，其他人也适应这些变化，其中一些变化可能会拉大与特斯拉之间的差距，而其他变化可能会缩小差距。很难说。

You not only do you need a, you know, a million cars out there, a really large population of cars driving this stuff, and, and just to be clear it's not the 400 000 FSD cars that are important one that all of the, you know, there's four million Teslas out there driving around. They have the hardware that they can gather data with this stuff because for this kind of stuff and for a lot of the regular FSD stuff as well the FSD 11, FSD 10 perception whatnot, any car that's instrumented, you can gather data from it. It doesn't have to be driving on FSD. um you as we're mentioning before the quality of the data really matters so you can't just hoover up you know it i mean for instance say that you had a million cars out there and they were just driving around recording stuff but they didn't have the capacity that they didn't have the capacity to run fsd you just had cameras on cars and you just have all the state well first of all you can't ingest all that data you can't store all that data video is really big it takes insane amounts of storage i think as you know i had commented today just storing the stuff that they want to store the stuff that's actually useful takes you know mind-boggling amounts of of storage and if you're just going to store everything that a lot of cars come from like it's not doable you can't do that there's just there's not enough storage they're not enough bandwidth that's not a practical approach to trying to solve this problem tessell's approach was to get filters out there the cars are smart you can tell.

你不仅需要一百万辆车，哦，我是说有大量汽车在驾驶这些东西，而且，明确一下，重要的不是那40万辆FSD车辆，而是那四百万辆在路上行驶的特斯拉。它们有硬件可以收集这些数据，因为对于这种东西以及很多常规FSD的东西，FSD 11，FSD 10 perception之类，只要是有仪器的车辆，都可以收集数据。它们不必开启FSD。我们之前提到过，数据质量真的很重要，所以你不能只是汇集所有的数据，比如说你有一百万辆车在外面行驶，只是记录一些东西，但它们没有运行FSD的能力，只是在车上装了摄像头，只有全部这些数据储存下来。首先，你无法摄取所有数据，视频文件非常大，需要大量的存储空间，我想你知道的，他们今天只是储存了他们认为有用的数据，就需要惊人的存储量，如果你只是要储存所有来自许多车辆的全部数据，那是不可行的，你无法做到，存储空间不足，带宽也不够，这不是一个解决这个问题的实际方法。特斯拉的方法是让智能车上的过滤器出场，你可以觉察到。

the cars what you want and they go find just that stuff and they send it to you so you don't get if a car drives for an hour you don't get an hour video from all eight cameras you get 10 seconds of video of the one thing that you really wanted to see that you needed to add to your data so that that's a really critical component of this that the fleet is a search engine it's not a tape recorder right it's not just recording stuff and one last big takeaway from me from this conversation was james was talking about even if a competitor hypothetically were to solve fsd shortly after tesla does it's not like they're really going to be eating into tessell's market share the market is so huge the tam is massive so it's going to be a long time before we start to see any competition eating away at each other's market share in a robo taxi market so there will undoubtedly be a major first mover advantage for presumably tesla but even if somebody else were to figure it out a year or two later which is very unlikely it's not really going to matter that much for tessell's market share because the market is so big even the mighty tesla can't produce enough robo taxis fast enough to saturate this market for many many years to come just like the model s and x the new model three when this feature is enabled has the ability to allow it to use inputs from sensors to select drive or reverse automatically but yes this feature can be left off entirely and it can be overridden by pressing the break and then using the actual manual buttons to select your gear hope you guys have a wonderful day you can find me on x linked below please like the video if you did and a huge thank you to all of my patreon supporters.

你想要的车辆，它们会去找到你需要的东西，并将其寄给你，这样当车辆行驶一个小时时，你就不会得到八个摄像头一小时的视频，而是得到你真正想要看到的那个关键时刻的10秒视频，以便将其添加到你的数据中。所以这是很关键的一点，车队是一个搜索引擎，而不是一个录像机。另外，通过这次交流的一点非常重要的启示是，即使竞争对手在特斯拉之后不久就解决了全自动驾驶的问题，也不会真的侵蚀到特斯拉的市场份额，市场是如此巨大，市场的竞争将需要很长时间才会开始出现。因此，可以确定的是，在自动出租车市场上，无疑会有一个巨大的先发优势，预计是特斯拉，但即使其他公司在一两年后解决了这个问题（这种可能性非常小），对特斯拉的市场份额影响也不会太大，因为市场是如此之大，即使强大的特斯拉也无法生产出足够快速的自动出租车来饱和这个市场，这将需要很多年的时间，就像Model S和X以及新的Model 3，在启用此功能后，它具有使用传感器输入来自动选择驾驶或倒车的能力，但是是的，这个功能可以完全关闭，并且可以通过按下刹车然后使用实际的手动按钮来选择档位，希望你们有个美好的一天，你可以在下面的链接中找到我，请给视频点赞，如果你喜欢的话，还要非常感谢我所有的赞助者。