https://www.youtube.com/watch?v=sZK9u28Ig5M&t=11s
|
0:01 typical uh 0:03 car company do um what what they do is 0:06 they they um 0:08 they assemble vehicles 0:10 um and they send them to dealers and 0:13 they manage the supply chain uh the the 0:16 they might make the engine uh or 0:18 typically we'll make the engine but most 0:20 of the parts are made by suppliers hey 0:22 it's dave so elon musk was interviewed 0:24 at the all in summit today and he had a 0:26 bunch of things to say about tesla about 0:28 twitter and about a possible recession 0:31 in this video i want to focus on just 0:32 one part of the interview and that is 0:34 elon diving into this concept of tesla 0:37 is actually many companies in one 0:39 actually many startups in one and how 0:40 that is a competitive advantage for 0:42 tesla 0:44 and a lot of the actual technology 0:46 development is done by suppliers and 0:47 most most of the vehicle software is 0:49 done by suppliers so the actual amount 0:52 of uh real work done by car companies 0:54 that what you think of sort of like a gm 0:56 ford is not actually that much here elon 0:59 is contrasting tesla which develops a 1:01 lot of their technology and software 1:03 in-house and he compares it to actually 1:05 oems a typical oem which basically they 1:07 might make the engine but they basically 1:08 assemble all these parts from different 1:10 suppliers and a lot of the technology 1:12 development is handled by suppliers and 1:14 not the oems so this is a key difference 1:16 between tesla and other companies so so 1:18 in the case of tesla for example we we 1:20 do we we do our own sales and service 1:23 we don't have dealerships 1:25 then 1:26 tesla also has by far the biggest 1:28 network of superchargers sort of the 1:31 electric equivalent of gas stations so 1:33 we built an entire global supercharger 1:35 network which is still the most advanced 1:37 and by far the best 1:39 way to charge your car when traveling 1:40 long distance or if you live in a city 1:42 um and uh and don't have the ability to 1:45 charge your car with a street parking or 1:47 an apartment so elon starts out in 1:49 details three advantages that tesla has 1:51 from the get-go so first is tesla 1:53 handles their own sales right most oems 1:56 they basically outsource their sales to 1:59 franchises right and they don't have 2:00 control over how the actual car is being 2:02 sold this causes a big problem when a 2:05 dealership is selling both electric cars 2:07 and ic cars and there's a big kind of 2:09 tension because these dealerships 2:11 actually make more money off of the 2:12 service than actually the 2:14 the sale of the car often in tesla's 2:16 case they handle the sales themselves 2:17 most of the cars are ordered online and 2:19 then the second piece is that tesla 2:21 handles the service right and because 2:23 they handle the service they have all 2:24 the data of what problems are going 2:26 wrong and they can integrate that and 2:27 actually fix the problems much quicker 2:29 than a typical oem does and then the 2:31 third advantage that elon details is the 2:33 supercharger network now this is kind of 2:34 crazy but tesla's had a supercharger 2:37 network for basically a decade and 2:38 there's yet to be oem as aggressive as 2:41 tesla to set up a nationwide 2:43 supercharging network so this becomes a 2:45 big disadvantage to other oems which are 2:47 piecemealing charger networks together 2:48 versus tesla's kind of 2:50 easy to use 2:52 pervasive and prevalent supercharger 2:54 network that just basically is a huge 2:55 selling point for tesla cars then uh in 2:58 terms of vertical integration uh we 3:02 we make the the battery pack uh the the 3:05 power electronics the drive unit um 3:07 we 3:08 uh 3:09 we actually make we're we're more 3:11 integrated in in the parts we actually 3:13 make so much of the car uh internally uh 3:16 we're vertically integrated um not 3:18 necessarily because we just we think 3:19 that there's some religious reason to be 3:21 a product integrator but because uh the 3:23 pace that we needed to move was just 3:25 much faster than the supply chain could 3:26 move and to the degree that you inherit 3:28 the legacy supply chain and 3:30 hurt the legacy constraints including 3:32 their speed uh cost and uh and 3:35 technology here elon details vertical 3:37 integration as a key differentiating 3:39 factor for tesla and tesla is laser 3:41 focused right on driving down costs and 3:43 really ramping production this is one of 3:45 the main reasons why tesla builds their 3:46 own battery packs they build their own 3:48 drive trains power electronics um they 3:51 have their own glass department plastics 3:52 etc and they're even making their own 3:54 seats so for tesla this vertical 3:56 integration is a necessary part of 3:58 driving down costs which actually 3:59 increase demand they increase quality 4:01 also by making their own parts and while 4:02 it's super ambitious and also risky for 4:04 a startup or new company to be so 4:06 vertically integrated if it actually 4:08 works and the company gets to scale it 4:10 becomes a huge advantage because you're 4:12 able to produce better products at a 4:14 cheaper cost and move faster and then 4:16 tesla is as much a software company as 4:18 it is a hardware company so the software 4:20 that runs in tesla operates the car 4:23 operates the screen 4:25 does the charging 4:26 all of that stuff is developed by tesla 4:29 and 4:29 so we have sort of a car a tesla os in 4:32 the car building software right and 4:33 actually on a world-class level is 4:35 actually super hard especially in 4:37 automotive applications you need the 4:39 software actually to work and work 4:40 reliably i think for tesla they've spent 4:43 many many years building up 4:44 a world-class software team that rivals 4:46 any silicon valley company this has 4:48 become one of the major competitive 4:50 advantages for tesla most oems don't 4:52 control like the overall software 4:54 experience like each of their suppliers 4:56 actually controls the software for their 4:57 parts or controllers because of that 4:59 when there's a bug it's hard to find out 5:01 which supplier right is responsible for 5:03 the bug it's a huge endeavor for auto 5:05 oem to try to transform themselves from 5:07 kind of this assembler supply chain 5:09 manager into actually a modern day 5:11 software company and i just don't think 5:12 it's a realistic expectation to have 5:15 tesla has built uh 5:18 an autopilot ai team from scratch that 5:21 is the best real world ai team on earth 5:24 and if anyone else has got a better one 5:25 i'd like to see it demonstrated in a car 5:27 um 5:29 the full self-driving beta at this point 5:32 can 5:34 very often take you with zero 5:35 interventions across the bay area from 5:38 san jose to marin 5:40 so 5:41 through complex traffic it's really 5:43 quite sophisticated um and i invite 5:45 anyone to to join the beta or or look at 5:48 the videos of those who are in the beta 5:49 we've got like 100 000 people in the 5:50 beta so it's not tiny and we'll be 5:52 expanding that to 5:54 probably a million people or a million i 5:56 don't know on that order by the end of 5:58 the year tesla's full self-driving beta 6:00 program already has a hundred thousand 6:02 beta participants and elon is saying 6:03 that their plan is to scale to about a 6:06 million right beta participants by the 6:08 end of the year which is huge ambitious 6:09 growth and elon here is starting to 6:11 detail tesla's ai advantage right they 6:14 have the world's best full self-driving 6:15 autonomous driving team and they've 6:17 deployed their software in real cars 6:19 right different by tens of thousands of 6:21 people and accumulating a massive amount 6:23 of miles each day and each week which is 6:25 going back into the system to improve 6:27 the overall right fsd beta experience 6:30 we also we also built a chip team to 6:32 because there wasn't it wasn't hardware 6:34 that we could run the friggin uh ai on 6:37 uh we couldn't just uh fill the trunk 6:39 with a whole bunch of gpus um and 6:42 you know 6:44 they would would have taken a trunk full 6:46 of gpus that would have been very 6:48 expensive and take massive amount of 6:49 power and cooling uh just to be able to 6:51 do what the tesla designed uh full self 6:54 driving computer can do 6:55 so 6:56 and we started a chip team from scratch 6:58 designed it it was the best in the world 6:59 and still is the best in the world 7:01 several years later so here's an 7:03 advantage of tesla that is not talked 7:05 about a lot and that is the power 7:06 efficiency of tesla's fsc um hardware 7:09 computer inside their cars so a lot of 7:11 the comparable systems with gpus they 7:14 just suck up a lot of energy and they 7:16 aren't practical for use inside the car 7:18 so what tesla had to do is they created 7:19 their own chip they put it together into 7:21 their own fsd hardware they stuck it in 7:23 their cars ever since 2017 right and 7:26 since then tesla has been accumulating 7:27 building up this fsd beta fleet for data 7:30 purposes and to improve their systems 7:32 now this is something i think wouldn't 7:33 have been possible if tesla just waited 7:35 for an off-the-shelf solution it just 7:37 would have been too power hungry and 7:39 tesla wouldn't be able to achieve what 7:40 they set out to do so it's one of those 7:42 stories of tesla creating and owning 7:44 their own destiny when it comes to 7:46 autonomous driving we're designing a 7:48 dojo uh supercomputer to be able to 7:50 process the 7:51 um all the video that's coming in from 7:54 billions of miles of data because just 7:56 sort of like the way that it's critical 7:57 to compete with google because they have 7:59 so much data and they have all this 8:00 people doing searches all the time and 8:02 humanity is training it but the same is 8:04 true of of tesla you really need 8:06 billions of miles and ultimately tens of 8:08 billions of miles of training data 8:09 combined with a sort of a vast training 8:12 computer and then 8:14 optimize 8:16 uh inference hardware in the car 8:18 and say the ai and training and 8:20 specialized software across the board to 8:22 be able to achieve a full self-driving 8:24 solution so elon shares how tesla's 8:27 building their own dojo ai chip but then 8:29 he goes on and says actually that tesla 8:31 has this huge data advantage just like 8:33 how the whole world is training like 8:35 google's supercomputer and the neural 8:37 nets with google search so every search 8:39 query you enter into google and the 8:40 results that you click is actually 8:42 training their neural nets to present 8:43 better results to you in a similar way 8:45 elon is saying that tesla has this huge 8:47 data advantage where they have billions 8:49 of miles right driven by the tesla fleet 8:52 and that data is being put back into the 8:54 system and you need a super computer 8:56 that's um powerful enough to handle that 8:58 data so it's all this system that 8:59 actually is dependent on each other you 9:01 need a massive super computer you need 9:02 the data in massive amounts you need it 9:05 labeled right in the right way you need 9:06 the inference hardware inside the cars 9:08 and you need a super talented ai team 9:10 right to oversee all of this the car 9:12 insurance thing is a bigger deal it may 9:14 seem a lot of people are paying um you 9:16 know 30 40 percent as much as their 9:18 lease payment for the car in in car 9:21 insurance um so the car insurance 9:23 industry is incredibly inefficient 9:24 because they they're just 9:26 first of all you got like so many um 9:28 sort of middle entities you've got from 9:30 an insurance agent all the way to 9:32 the final sort of reinsurer there's like 9:34 a half dozen companies each taking a cut 9:36 um and then 9:38 the 9:38 [Music] 9:39 it's all very statistical so that this 9:42 um even if you're a very good driver 9:44 like you could be like you know 9:46 20 years old and a great driver but they 9:48 it's a little statistical so you can't 9:50 get you either can't get insurance or 9:52 it's extremely expensive um so 9:55 but tesla allows for real-time insurance 9:57 based on you how you actually drive the 9:58 car um you can actually if you drive the 10:01 car 10:02 in a safer way you actually uh have 10:05 lower insurance so ours is insurance is 10:08 based on how you actually drive not how 10:11 you know 10:12 historically people that you know 10:15 fit your whatever demographic have drive 10:16 it's and then you can close the loop 10:18 around your uh insurance rate by simply 10:21 driving better and looking at your score 10:23 and and and lowering your insurance in 10:25 real time and people do it actually 10:26 promotes safer driving the more i think 10:28 about tesla insurance the more i think 10:30 it's just a revolutionary product and 10:32 it's a combination of it's integrated 10:34 into the car right and it actually 10:36 tracks how you drive so you can actually 10:38 adjust right the insurance rates that 10:40 you give to customers real time based 10:42 upon their driving you're showing them a 10:44 safety score so you can get feedback and 10:46 they can actually adjust and make their 10:48 driving safer if they want a lower rate 10:50 and by doing so you could probably 10:52 undercut a lot of other insurance 10:53 providers because you actually have the 10:55 data and you can actually insure people 10:57 accordingly with the right prices over 10:59 time this can be a competitive advantage 11:01 because you can think that people buy 11:03 and test a car they can say hey i can 11:05 actually save money when i actually also 11:06 get tesla insurance on top of that and 11:08 then you add on the gas savings right of 11:10 not buying gas and just using 11:12 electricity to fill the car and you get 11:14 this kind of overall cost advantage and 11:16 then you add on tesla solar and tesla 11:18 energy right you have a powerwall tesla 11:20 solar or tesla solar roof and all 11:22 combined right it just is a huge 11:24 ecosystem where you're just saving money 11:26 you have a better product experience and 11:28 the product that you're driving which is 11:30 this car is getting better and better at 11:31 data and eventually right the goal is 11:33 this car drives itself and then you have 11:35 further feels that tesla is expanding 11:36 into namely you have robo taxi a huge 11:39 fleet of autonomously driven vehicles 11:41 that can basically obliterate uber and 11:44 lyft in the future likely and then moves 11:46 on to other fields like humanoid robots 11:48 which could be the ultimate market of 11:50 all and not to forget a lot of other 11:51 divisions of tesla that are really doing 11:53 amazing work you have people in tesla or 11:55 huge divisions actually focused on 11:57 building factories you have 11:58 big division departments building 12:00 machinery um these machineries are used 12:03 in their factories to build not just the 12:05 cars and parts but now more and more 12:06 tesla is developing their own 4680 12:08 battery cells as well so overall you've 12:10 got this huge conglomerate of companies 12:12 under one umbrella of tesla and when you 12:14 put them all together it's actually an 12:16 extremely focused company that is really 12:18 focused on driving down costs for 12:20 electric vehicles increasing adoption 12:22 making the product better using ain 12:25 software right to really leverage neural 12:27 nets to basically transform the driving 12:30 experience from an active driving 12:31 experience into a driving experience 12:33 where ai and artificial neural nets are 12:35 driving for you and eventually all this 12:38 can be placed in a humanoid robot it's 12:39 kind of insane and crazy anyways i 12:41 thought this was a great interview by 12:42 elon kind of giving an overview on what 12:45 makes tesla so unique and distinct 12:47 there's still a lot of confusion i think 12:49 by a lot of people they still compare 12:51 tesla with the typical oem and they 12:53 complain saying that tesla is just an 12:55 auto company right deserving what other 12:56 auto companies deserve in terms of 12:58 investment multiples anyways hope this 13:00 has been helpful if it has good like and 13:01 subscribe all my videos can be found as 13:03 an audio podcast as well just search for 13:05 dave lee on investing in your favorite 13:06 podcast player i'm also on twitter at 13:08 heyday7 alright we'll see you guys my 13:10 next video thanks
|
0:00 如果你看著並說出什麼是什麼 0:01 典型的呃 0:03 汽車公司做的，他們做的是什麼 0:06 他們他們嗯 0:08 他們組裝車輛 0:10 嗯，他們把它們寄給經銷商和 0:13 他們管理供應鏈 0:16 他們可能會製造引擎 0:18 通常我們會製造引擎，但大多數 0:20 零件由供應商製造嘿 0:22 是戴夫，所以伊隆馬斯克接受了採訪 0:24 在今天的峰會上，他有一個 0:26 一堆關於特斯拉的事情要說 0:28 推特和關於可能的經濟衰退 0:31 在這個視頻中，我只想專注於 0:32 面試的一部分，那就是 0:34 伊隆潛入特斯拉的這個概念 0:37 實際上是許多公司合而為一 0:39 實際上有很多初創公司以及如何 0:40 這是一個競爭優勢 0:42 特斯拉 0:44 和很多實際的技術 0:46 開發由供應商和 0:47 大多數車輛軟件是 0:49 由供應商完成，因此實際金額 0:52 汽車公司所做的真實工作 0:54 你的想法有點像gm 0:56 福特在這裡實際上並沒有那麼多埃隆 0:59 對比特斯拉，它開發了一個 1:01 很多他們的技術和軟件 1:03 在內部，他將其與實際進行比較 1:05 oem 一個典型的 oem 基本上他們 1:07 可能會製造引擎，但他們基本上 1:08 從不同的地方組裝所有這些零件 1:10 供應商和很多技術 1:12 開發由供應商處理，並且 1:14 不是原始設備製造商，所以這是一個關鍵的區別 1:16 特斯拉和其他公司之間如此 1:18 以特斯拉為例，我們 1:20 我們做我們自己的銷售和服務嗎 1:23 我們沒有經銷商 1:25 然後 1:26 特斯拉也擁有迄今為止最大的 1:28 增壓器網絡之類的 1:31 加油站的電當量所以 1:33 我們建造了一個完整的全球增壓器 1:35 仍然是最先進的網絡 1:37 到目前為止最好的 1:39 旅行時給汽車充電的方法 1:40 長途或者如果你住在城市 1:42 嗯，嗯，沒有能力 1:45 使用路邊停車位為您的汽車充電或 1:47 一個公寓，所以埃隆開始在 1:49 詳述特斯拉的三大優勢 1:51 從一開始就是特斯拉 1:53 大多數 OEM 處理自己的銷售權 1:56 他們基本上將銷售外包給 1:59 特許經營權，他們沒有 2:00 控制實際汽車的運行方式 2:02 賣了這會導致一個大問題，當 2:05 經銷商正在銷售這兩款電動汽車 2:07 和ic汽車，還有一種很大的 2:09 緊張，因為這些經銷商 2:11 實際上賺更多的錢 2:12 服務比實際上 2:14 賣車的時候經常在特斯拉 2:16 如果他們自己處理銷售 2:17 大多數汽車都是在線訂購的 2:19 然後第二塊是特斯拉 2:21 處理服務權，因為 2:23 他們處理他們擁有的服務 2:24 正在發生什麼問題的數據 2:26 錯了，他們可以整合 2:27 實際上更快地解決問題 2:29 比典型的 oem 做的，然後 2:31 elon 細節的第三個優勢是 2:33 增壓器網絡現在這是一種 2:34 瘋了，但特斯拉有一個增壓器 2:37 網絡基本上十年和 2:38 還沒有像 2:41 特斯拉將在全國范圍內建立 2:43 增壓網絡，所以這變成了一個 2:45 對其他 oem 的巨大劣勢 2:47 零碎的充電網絡 2:48 與特斯拉的那種 2:50 便於使用 2:52 無處不在的增壓器 2:54 基本上是一個巨大的網絡 2:55 特斯拉汽車的賣點然後呃 2:58 垂直整合的條款，呃，我們 3:02 我們製造電池組 3:05 電力電子驅動單元 um 3:07 我們 3:08 呃 3:09 我們實際上讓我們更 3:11 集成在我們實際上的部分中 3:13 做這麼多的車呃內部呃 3:16 我們是垂直整合的，嗯不是 3:18 必然是因為我們只是我們認為 3:19 有一些宗教原因 3:21 產品集成商，但因為呃 3:23 我們需要移動的速度只是 3:25 比供應鏈快得多 3:26 移動到你繼承的程度 3:28 傳統的供應鍊和 3:30 損害遺留約束，包括 3:32 他們的速度呃成本和呃和 3:35 技術在這裡 elon 細節垂直 3:37 整合是一個關鍵的區別 3:39 特斯拉和特斯拉的因素是激光 3:41 專注於降低成本和 3:43 真正提高產量這是其中之一 3:45 特斯拉建立他們的主要原因 3:46 擁有自己的電池組 3:48 傳動系統 電力電子 嗯，他們 3:51 擁有自己的玻璃部門塑料 3:52 等等，他們甚至自己做 3:54 所以對於特斯拉來說，這個垂直的座位 3:56 整合是必不可少的一部分 3:58 降低成本，實際上 3:59 增加需求，他們提高質量 4:01 也通過製作自己的零件，而 4:02 這是超級雄心勃勃的，也有風險 4:04 一家初創公司或新公司 4:06 如果它實際上是垂直整合的 4:08 工作，公司可以擴大規模 4:10 成為一個巨大的優勢，因為你 4:12 能夠生產出更好的產品 4:14 成本更低，移動速度更快，然後 4:16 特斯拉是一家軟件公司 4:18 它是一家硬件公司，所以軟件 4:20 在特斯拉運行的汽車 4:23 操作屏幕 4:25 充電嗎 4:26 所有這些東西都是由特斯拉開發的 4:29 和 4:29 所以我們有一輛裝有特斯拉操作系統的汽車 4:32 汽車製造軟件的權利和 4:33 實際上在世界級水平上是 4:35 實際上超級難，尤其是在 4:37 您需要的汽車應用 4:39 軟件實際工作和工作 4:40 可靠地，我認為他們已經為特斯拉花費了 4:43 多年積累 4:44 世界一流的軟件團隊，可與對手媲美 4:46 任何一家矽谷公司都有 4:48 成為主要競爭者之一 4:50 特斯拉的優勢大多數 OEM 沒有 4:52 像整體軟件一樣控制 4:54 像他們的每個供應商一樣體驗 4:56 實際上控制他們的軟件 4:57 因此，零件或控制器 4:59 當有錯誤時很難發現 5:01 哪個供應商權利負責 5:03 對汽車來說這是一個巨大的努力 5:05 oem 試圖改變自己 5:07 這種組裝商供應鏈 5:09 經理實際上進入了現代 5:11 軟件公司，我只是不認為 5:12 這是一個現實的期望 5:15 特斯拉已經建立了呃 5:18 一個從零開始的自動駕駛人工智能團隊 5:21 是地球上最好的真實世界人工智能團隊 5:24 如果其他人有更好的 5:25 我想看它在汽車上演示 5:27 嗯 5:29 此時是完整的自動駕駛測試版 5:32 能夠 5:34 經常帶你零 5:35 整個灣區的干預措施 5:38 聖何塞到馬林 5:40 所以 5:41 通過複雜的交通 5:43 相當成熟的嗯，我邀請 5:45 任何人都可以加入測試版或查看 5:48 那些處於測試階段的人的視頻 5:49 我們有大約 100 000 人 5:50 測試版，所以它不是很小，我們會 5:52 將其擴展到 5:54 可能有一百萬人或一百萬我 5:56 到年底不知道那個訂單 5:58 當年特斯拉的全自動駕駛測試版 6:00 程序已經有十萬 6:02 beta 參與者和 elon 說 6:03 他們的計劃是擴大到大約 6:06 百萬權利測試參與者 6:08 雄心勃勃的年底 6:09 這裡的增長和 elon 開始 6:11 詳細說明特斯拉的人工智能優勢對他們 6:14 擁有世界上最好的全自動駕駛 6:15 自動駕駛團隊，他們已經 6:17 在真實汽車中部署他們的軟件 6:19 相差數万 6:21 人和積累了大量的 6:23 每天和每週的英里數 6:25 重新進入系統以改進 6:27 整體正確的 fsd beta 體驗 6:30 我們還建立了一個芯片團隊 6:32 因為沒有它不是硬件 6:34 我們可以運行 friggin uh ai on 6:37 呃，我們不能只是呃填滿後備箱 6:39 用一大堆 gpus um 和 6:42 你知道 6:44 他們會裝滿一個行李箱 6:46 gpus 本來會很 6:48 昂貴並且需要大量的 6:49 電源和冷卻，呃，只是為了能夠 6:51 做特斯拉設計的，完全自我 6:54 行車電腦可以做 6:55 所以 6:56 我們從零開始組建了一個芯片團隊 6:58 設計它是世界上最好的 6:59 仍然是世界上最好的 7:01 幾年後所以這裡有一個 7:03 特斯拉不為人知的優勢 7:05 大約很多，這就是力量 7:06 特斯拉 fsc um 硬件的效率 7:09 他們的車裡有很多電腦 7:11 具有 gpus 的可比系統，它們 7:14 只是吸收很多能量，他們 7:16 不適合在車內使用 7:18 所以特斯拉要做的是他們創造了 7:19 他們自己的芯片，他們把它放在一起 7:21 他們自己的 fsd 硬件，他們把它粘在裡面 7:23 他們的車從 2017 年開始就對了 7:26 從那以後，特斯拉一直在積累 7:27 為數據建立這個 fsd beta 艦隊 7:30 目的並改進他們的系統 7:32 現在這是我認為不會 7:33 如果特斯拉只是等待，就有可能 7:35 對於現成的解決方案，它只是 7:37 本來太耗電了 7:39 特斯拉將無法實現什麼 7:40 他們開始這樣做，這是其中之一 7:42 特斯拉創造和擁有的故事 7:44 當涉及到他們自己的命運時 7:46 自動駕駛我們正在設計一個 7:48 dojo uh 超級計算機能夠 7:50 處理 7:51 嗯，所有來自的視頻 7:54 數十億英里的數據，因為只是 7:56 有點像它的關鍵 7:57 與穀歌競爭，因為他們有 7:59 這麼多數據，他們擁有這一切 8:00 人們一直在搜索 8:02 人類正在訓練它，但同樣是 8:04 真正的特斯拉你真的需要 8:06 數十億英里，最終數十億英里 8:08 數十億英里的訓練數據 8:09 結合一種廣泛的培訓 8:12 電腦然後 8:14 優化 8:16 呃車內的推理硬件 8:18 說人工智能和訓練和 8:20 全面的專業軟件 8:22 能夠實現全自動駕駛 8:24 解決方案所以埃隆分享了特斯拉的 8:27 建立自己的道場人工智能芯片，然後 8:29 他繼續說實際上特斯拉 8:31 擁有巨大的數據優勢，就像 8:33 全世界都在訓練 8:35 谷歌的超級計算機和神經網絡 8:37 帶有谷歌搜索的網，所以每次搜索 8:39 查詢您輸入谷歌和 8:40 您點擊的結果實際上是 8:42 訓練他們的神經網絡呈現 8:43 以類似的方式為您帶來更好的結果 8:45 埃隆是說特斯拉有這麼大 8:47 他們擁有數十億美元的數據優勢 8:49 由特斯拉車隊駕駛的英里數 8:52 並且這些數據正在被放回 8:54 系統，你需要一台超級計算機 8:56 這足夠強大，可以處理 8:58 數據，所以這一切都是這個系統 8:59 其實是互相依賴的你 9:01 需要你需要的大型超級計算機 9:02 你需要的海量數據 9:05 以您需要的正確方式標記 9:06 車內的推理硬件 9:08 你需要一個超級有才華的人工智能團隊 9:10 監督所有這一切的權利 9:12 保險是一件大事，它可能 9:14 好像很多人都在付錢給你 9:16 知道他們的 30% 40% 9:18 車內車的租賃付款 9:21 保險嗯所以汽車保險 9:23 工業效率低得令人難以置信 9:24 因為他們只是 9:26 首先你得到了很多 9:28 您從中獲得的那種中間實體 9:30 一路保險代理人 9:32 最後一種再保險公司就像 9:34 六家公司各分一杯羹 9:36 嗯然後 9:38 這 9:38 [音樂] 9:39 這都是非常統計的，所以這 9:42 嗯，即使你是一個很好的司機 9:44 就像你知道的那樣 9:46 20 歲，一個很棒的司機，但他們 9:48 這有點統計，所以你不能 9:50 讓你要么無法獲得保險，要么 9:52 它非常昂貴，所以 9:55 但特斯拉允許實時保險 9:57 根據您實際駕駛的方式 9:58 嗯，如果你開車，你實際上可以 10:01 車 10:02 以一種更安全的方式，你實際上有 10:05 較低的保險，所以我們的保險是 10:08 基於你實際駕駛的方式而不是駕駛方式 10:11 你知道 10:12 歷史上你認識的人 10:15 適合您的任何人群有驅動力 10:16 它是，然後你可以關閉循環 10:18 簡單地圍繞你的保險費率 10:21 更好地駕駛並查看您的分數 10:23 並降低您的保險 10:25 實時的，人們實際上是這樣做的 10:26 我認為越提倡安全駕駛 10:28 關於特斯拉保險，我想得越多 10:30 它只是一個革命性的產品 10:32 它是集成的 10:34 進入車裡，它實際上 10:36 跟踪您的駕駛方式，以便您實際上可以 10:38 調整正確的保險費率 10:40 你給客戶實時基於 10:42 在他們開車時，您正在向他們展示 10:44 安全評分，以便您獲得反饋和 10:46 他們實際上可以調整併使他們的 10:48 如果他們想要更低的利率，駕駛更安全 10:50 這樣做你可能會 10:52 削弱了許多其他保險 10:53 提供者，因為您實際上擁有 10:55 數據，您實際上可以為人們提供保險 10:57 因此，價格合適 10:59 時間這可以成為競爭優勢 11:01 因為你可以認為人們購買 11:03 並測試一輛車，他們可以說嘿，我可以 11:05 其實我也省錢的時候 11:06 在此之上獲得特斯拉保險， 11:08 然後你添加天然氣儲蓄權 11:10 不買汽油，只用 11:12 電力充滿汽車，你得到 11:14 這種綜合成本優勢和 11:16 然後你加上特斯拉太陽能和特斯拉 11:18 能源對，你有一個 powerwall 特斯拉 11:20 太陽能或特斯拉太陽能屋頂和所有 11:22 結合起來，它只是一個巨大的 11:24 您只是在省錢的生態系統 11:26 你有更好的產品體驗和 11:28 您駕駛的產品是 11:30 這輛車越來越好 11:31 數據和最終正確的目標是 11:33 這輛車自動駕駛然後你有 11:35 進一步感覺特斯拉在擴張 11:36 進入即你有一個巨大的機器人出租車 11:39 自動駕駛車隊 11:41 基本上可以消滅優步和 11:44 未來可能會飛然後移動 11:46 到人形機器人等其他領域 11:48 這可能是最終的市場 11:50 所有，不要忘記很多其他 11:51 真正在做的特斯拉部門 11:53 令人驚嘆的工作，你有特斯拉的人或 11:55 巨大的分歧實際上集中在 11:57 建造你擁有的工廠 11:58 大司部門大樓 12:00 機器嗯，這些機器是用的 12:03 在他們的工廠中建造的不僅僅是 12:05 汽車和零件，但現在越來越多 12:06 特斯拉正在開發自己的4680 12:08 電池也是如此，總體而言，您已經 12:10 得到了這個龐大的公司集團 12:12 在特斯拉的一把保護傘下，當你 12:14 把它們放在一起它實際上是一個 12:16 非常專注的公司 12:18 專注於降低成本 12:20 電動汽車越來越多的採用 12:22 使用 ain 使產品更好 12:25 真正利用神經網絡的軟件權利 12:27 網基本改造驅動 12:30 主動駕駛的經驗 12:31 體驗到駕駛體驗 12:33 人工智能和人工神經網絡在哪裡 12:35 為你開車，最終這一切 12:38 可以放在人形機器人中 12:39 反正我有點瘋狂和瘋狂 12:41 認為這是一次很棒的採訪 12:42 埃隆有點概述了什麼 12:45 讓特斯拉如此獨特和與眾不同 12:47 我認為仍然有很多困惑 12:49 很多人還在比較 12:51 特斯拉與典型的 oem 和他們 12:53 抱怨說特斯拉只是一個 12:55 汽車公司有什麼其他權利 12:56 汽車企業應得的 12:58 無論如何，投資倍數希望這個 13:00 如果它有很好的喜歡和 13:01 訂閱我所有的視頻可以找到 13:03 一個音頻播客也只是搜索 13:05 戴夫李投資你最喜歡的 13:06 播客播放器我也在推特上 13:08 heyday7 好吧，我們會再見的，我的 13:10 下個視頻謝謝
|
沒有留言:
張貼留言