https://www.youtube.com/watch?v=CGsjE43iDa8
|
0:05 [Music] 0:07 hi it's dave so elon musk made some 0:09 comments on twitter this past week that 0:11 i think are super bullish for tesla 0:13 however i think they got lost in the 0:15 noise by most people so let's take a 0:17 look and reflect on some of these 0:18 comments here so elon replied to a two 0:20 by jack dorsey and elon says 2020 feels 0:23 like a pivotal year i'd be surprised if 0:25 we don't have agi by then hopefully 0:27 people on mars too so i want to focus on 0:30 this comment here um before actually um 0:33 on march 14th a few months ago 0:35 someone asked what's your guess um on 0:38 when people will land on mars and elon 0:40 says 2029. so we know elon has a goal of 0:42 sending people to mars in 2029 but this 0:45 is the first time from this tweet that 0:46 we know that elon thinks that actually 0:48 ati or artificial general intelligence 0:50 will be here by 2029. so he says i'd be 0:53 surprised if we don't have agi or 0:55 artificial general intelligence by 2029 0:58 all right so agi what is it artificial 1:00 general intelligence according to 1:01 wikipedia is the ability of an 1:03 intelligent agent to understand or learn 1:05 any intellectual task that a human being 1:07 can now there's varying different views 1:10 on what agi is if it is even possible 1:13 exactly what it means and by when it 1:14 will come i wrote last year in may i 1:16 said we're on the cusp of something big 1:18 really big namely the emergence of ai 1:20 powered robots that will be able to do 1:21 most of what humans do but better this 1:23 is both exciting and concerning all 1:25 right so 2029 this year is much sooner 1:28 than most everyone projects regarding 1:30 the advent of artificial general 1:32 intelligence and i think it's probably 1:34 because elon is seeing a lot of progress 1:36 in artificial intelligence that's 1:38 happening sooner and faster and as 1:40 evidenced by the progress we're seeing 1:42 in fsd beta and also with tesla's just 1:45 ai team in general all right so one of 1:47 the questions of debate is whether hgi 1:49 artificial general intelligence will be 1:51 physical kind of robot form or will it 1:53 be just virtual right so non-physical 1:56 most ai research and development is 1:58 happening actually in virtual 1:59 environments this is software internet 2:01 computers however tesla is a unique 2:03 company in sciences they're pushing 2:05 physical ai right and they are becoming 2:07 one of the leaders if not the defective 2:09 leader in physical ai meaning combining 2:12 hardware and software together their 2:13 first physical ai product is basically 2:16 robot cars or cars that dry themselves 2:18 this is still in development their 2:20 second physical ai product is the 2:22 humanoid ai robot that eventually can do 2:24 almost everything that a human can do 2:26 all right so this humanoid ai robot is a 2:29 product i think that practically has no 2:31 limits see once a humanoid ai robot the 2:34 intelligence of that robot reaches that 2:36 of a human and passes that then it can 2:38 only and will only increase rapidly 2:41 right in the future and the question is 2:42 what will humanoid ai robots not be 2:44 capable of at a certain point they can 2:47 mine right minerals themselves they can 2:49 build their own factories they could 2:51 repair their own bodies they can make 2:53 new forms right and better robots in the 2:55 future it kind of is mind-boggling in a 2:57 sense it is all concerning in many i 3:00 think rightful ways but also promising 3:01 if it's in the right hands and if it's 3:03 used for good alright so another big 3:05 question mark is timing when will tesla 3:08 be able to make a truly usable humanoid 3:10 ai robot so when is this all starting 3:13 and how fast can tesla improve their 3:15 humanoid robot to the point right that 3:17 you can actually become as smart as a 3:19 human and this is why elon's 2029 agi 3:22 comment is really significant because 3:24 it's showing that elon is seeing rapid 3:27 progress with artificial intelligence 3:29 also with robotics and it's showing that 3:31 actually elon is super bullish about 3:33 tesla's prospects and actually not just 3:35 solving ai but actually with tesla's 3:37 humanoid robot i think there's a 3:38 divergence of beliefs that we're going 3:40 to see more clearly in coming years most 3:42 people won't believe the rapid progress 3:45 and development of tesla's humanoid 3:46 robot it just is too mind-boggling for 3:48 most people i think this divergence of 3:50 beliefs meaning there'll be more people 3:52 who actually think it's coming and 3:54 there'll be more people who thinks it's 3:55 just 50 100 years or it's never going to 3:57 come and i think this divergence of 3:59 belief is actually presents some 4:00 opportunity all right so overall i think 4:03 robots are developing fast i want to 4:04 give an example here of agility robotics 4:07 they're based in 4:08 oregon and they've got a robot called 4:10 digit so here is their homepage here and 4:13 digit is a work robot and they're still 4:16 kind of developing it and they're going 4:18 to release supposedly their first 4:19 version this year here's the team at 4:21 agility robotics right you've got 4:23 probably a 4:24 few to several dozen people so it's not 4:26 a huge company but they're doing 4:28 actually some really interesting stuff 4:30 so here is kind of the version or what 4:32 they have of this digit a work robot you 4:35 see that pedaling 4:36 walking waving his hands it's got a 4:39 quite lean body 4:41 yeah and they've got actually very 4:42 ambitious uh plans for this robot so 4:45 they want to 4:46 for initial tasks they want to move 4:49 totes and packages right so this is 4:50 basically 4:52 a robot that work in factories they want 4:54 to unload trailers right this would be a 4:56 huge i think use case for robots um 4:59 there's a ton of cargo right at docks 5:01 not just docks but 5:02 warehouses etc 5:04 but not just unloading trailers but you 5:06 could actually do last mile delivery so 5:08 you basically take the boxes right from 5:10 a car a van and actually walk the boxes 5:13 up to a person's house lay it down 5:15 and it presents a really super 5:17 fascinating kind of um yeah initial task 5:20 or use cases for a humanoid robot right 5:23 so according to agility robotics their 5:26 kind of model or their robot actually is 5:29 about five feet two inches um and it's 5:32 very similar actually in kind of 5:34 dimensions to tesla's humanoid robot and 5:36 it's very interesting um weight 45 5:39 kilograms max speed 1.5 meter per second 5:42 you have a light duty run time of three 5:44 hours and a heavy duty run time of 1.5 5:47 hours also in terms of walking or 5:49 mobility you can go forward backwards 5:51 side by side turn in place crouching 5:53 walk up and down inclines across 5:56 unstructured territory uh terrain like 5:58 grass rocks and curbs other capabilities 6:00 are it pauses when perceives a person or 6:02 barrier in its path and navigates around 6:04 it can crouch and squat it rotates upper 6:06 body to gaze and look in different 6:08 directions it uses arms and hands and 6:10 feet to balance when bumped now this is 6:12 pretty cool ability to pick up and put 6:13 down objects of varying size of and 6:15 weight steps up and down from curbs so 6:18 agility robotics they've raised 150 6:20 million dollars just um last month april 6:23 22nd 6:24 and they raised it from dc vc and 6:26 playground global also amazon industrial 6:29 innovation fund joined in as an investor 6:31 here's a video of their digit work robot 6:34 in action um it's quite interesting 6:38 yeah so you see that we're moving boxes 6:41 kind of in a warehouse 6:43 loading boxes onto right a trailer 6:46 taking out products 6:48 from bins 6:50 placing the bins in a different place 6:54 loading up a truck here 6:57 actually taking up taking boxes 6:59 and 7:00 um yeah quite impressive stuff actually 7:03 by agility robotics here 7:08 all right so i think there are some huge 7:10 ramifications of what's going on here 7:12 tesla is recruiting the world's best 7:15 robotic experts and they've got their 7:17 choice of the cream of the crop tesla 7:19 already has one of the best ai teams out 7:21 there and i think we're going to see 7:23 progress far faster than any other 7:24 robotic teams in the world right now 7:26 second we're probably seeing the 7:28 development of the greatest product in 7:30 human history yes it has a lot of 7:31 dangers and risks so hopefully they'll 7:34 be the right regulation it'll be the 7:35 right companies that are that will 7:36 steward this technology to actually help 7:38 people not to harm write the greater 7:40 good third when you ever accompany or 7:43 people create great value right through 7:45 innovation technology there's going to 7:46 be a realization of some of that value 7:49 if not a lot of that value via revenue 7:51 and profits and through market cap right 7:53 and through stock price so i think this 7:55 becomes a backdrop of tesla stock for 7:57 the next decade and many decades to come 7:59 and again what's interesting is elon 8:01 says his expectations 2029 for hei 8:04 personally i think that's way early i 8:06 think this can take longer than that but 8:07 hey i take it as a sign of confidence 8:09 and a sign of optimism that things are 8:11 going very well with the development of 8:13 tesla ai fsd beta and also 8:16 tesla's humanoid robot all right hope 8:17 this has been helpful if it has good 8:19 like and subscribe all my videos can be 8:20 found as an audio podcast as well just 8:22 search for dave lee on investing in your 8:24 favorite podcast player i'm also on 8:25 twitter heyday7 alright we'll see you 8:27 guys in my next video thanks AllListenable
|
0:07 嗨，戴夫，伊隆馬斯克做了一些 0:09 上週在推特上的評論 0:11 我認為對特斯拉超級看好 0:13 但是我認為他們迷路了 0:15 大多數人的噪音，所以讓我們來看看 0:17 看看並反思其中的一些 0:18 這裡的評論所以埃隆回復了兩個 0:20 傑克·多爾西和埃隆說 2020 年的感覺 0:23 如果是關鍵的一年，我會感到驚訝 0:25 希望到那時我們還沒有 agi 0:27 火星上的人也一樣，所以我想專注於 0:30 這個評論在這裡嗯實際上之前 0:33 幾個月前的 3 月 14 日 0:35 有人問你猜怎麼著 0:38 當人們登陸火星和埃隆時 0:40 說 2029。所以我們知道埃隆的目標是 0:42 2029 年將人送上火星，但這 0:45 這是這條推文中的第一次 0:46 我們知道埃隆實際上認為 0:48 ati 或通用人工智能 0:50 到 2029 年會在這裡。所以他說我會 0:53 如果我們沒有敏捷或 0:55 到 2029 年的通用人工智能 0:58 好吧，所以 agi 是什麼人造的 1:00 一般智力根據 1:01 維基百科是一個能力 1:03 智能代理理解或學習 1:05 人類的任何智力任務 1:07 現在可以有不同的看法 1:10 如果可能的話，關於 agi 是什麼 1:13 究竟是什麼意思，什麼時候 1:14 會來我去年五月寫的 1:16 說我們正處於大事的風口浪尖 1:18 真的很大就是ai的出現 1:20 能夠做的動力機器人 1:21 人類所做的大部分事情，但更好的是這個 1:23 既令人興奮又關乎所有人 1:25 對，所以今年 2029 年要快得多 1:28 比大多數人的項目 1:30 人工將軍的出現 1:32 情報，我認為這可能是 1:34 因為埃隆看到了很多進步 1:36 在人工智能中 1:38 越來越快地發生 1:40 我們所看到的進展證明了這一點 1:42 在 fsd beta 和 tesla 中 1:45 ai團隊一般都還好所以之一 1:47 爭論的問題是hgi是否 1:49 通用人工智能將成為 1:51 物理類型的機器人形式或將它 1:53 只是虛擬的，所以非物理的 1:56 大多數人工智能研發是 1:58 實際發生在虛擬 1:59 環境 這是軟件互聯網 2:01 電腦，但特斯拉是獨一無二的 2:03 他們正在推動的科學公司 2:05 身體健康，他們正在成為 2:07 如果不是有缺陷的領導者之一 2:09 物理人工智能的領導者意味著結合 2:12 硬件和軟件在一起 2:13 第一款實體ai產品基本上是 2:16 機器人汽車或自動干燥的汽車 2:18 這仍在開發中 2:20 第二個物理人工智能產品是 2:22 最終能做到的人形人工智能機器人 2:24 人類能做的幾乎所有事情 2:26 好的，所以這個人形人工智能機器人是 2:29 我認為幾乎沒有的產品 2:31 極限見一次人形人工智能機器人 2:34 那個機器人的智能達到了 2:36 一個人，然後通過它可以 2:38 只會而且只會迅速增加 2:41 就在未來，問題是 2:42 人形人工智能機器人不會是什麼 2:44 能夠在某一點上，他們可以 2:47 他們可以自己開採正確的礦物 2:49 他們可以建立自己的工廠 2:51 他們可以修復自己的身體 2:53 新形式正確和更好的機器人在 2:55 未來它有點令人難以置信 2:57 感覺這一切都與我有關 3:00 思考正確的方法，但也有希望 3:01 如果它在正確的手中，如果它是 3:03 用得很好，所以另一個大 3:05 問號是時間特斯拉什麼時候 3:08 能夠製造出真正可用的人形機器人 3:10 人工智能機器人所以這一切什麼時候開始 3:13 特斯拉能以多快的速度提高他們的 3:15 人形機器人到正確的點 3:17 你實際上可以變得像 3:19 人類，這就是為什麼 Elon 的 2029 敏捷 3:22 評論真的很重要，因為 3:24 這表明埃隆正在迅速 3:27 人工智能的進步 3:29 也有機器人技術，它表明 3:31 實際上，埃隆非常看好 3:33 特斯拉的前景，實際上不僅僅是 3:35 解決 ai 但實際上是用特斯拉的 3:37 人形機器人我認為有一個 3:38 我們要去的信念的分歧 3:40 在未來幾年更清楚地看到最 3:42 人們不會相信快速的進步 3:45 和開發特斯拉的人形機器人 3:46 機器人太令人難以置信了 3:48 大多數人我認為這種分歧 3:50 信仰意味著會有更多的人 3:52 誰實際上認為它即將到來 3:54 會有更多的人認為 3:55 只有 50 100 年，否則永遠不會 3:57 來吧，我認為這種分歧 3:59 信念實際上是呈現一些 4:00 機會很好所以總的來說我認為 4:03 機器人發展很快我想 4:04 在這裡舉一個敏捷機器人的例子 4:07 他們位於 4:08 俄勒岡，他們有一個機器人叫 4:10 數字所以這裡是他們的主頁 4:13 digit 是一個工作機器人，他們仍然 4:16 有點開發它，他們會去 4:18 釋放他們的第一個 4:19 今年的版本這裡的團隊在 4:21 你擁有的敏捷機器人技術 4:23 可能是 4:24 幾到幾十人，所以不是 4:26 一家大公司，但他們正在做 4:28 實際上是一些非常有趣的東西 4:30 所以這是一種版本或什麼 4:32 他們有這個數字的工作機器人你 4:35 看到那踩踏板 4:36 走路揮手它有一個 4:39 很瘦的身體 4:41 是的，他們實際上非常 4:42 雄心勃勃的呃計劃這個機器人所以 4:45 他們想要 4:46 對於他們想要移動的初始任務 4:49 手提包和包裹是對的，所以這是 4:50 基本上 4:52 在他們想要的工廠工作的機器人 4:54 正確卸載拖車，這將是一個 4:56 巨大的我認為機器人的用例 4:59 碼頭上有一噸貨物 5:01 不僅是碼頭，而且 5:02 倉庫等 5:04 但不僅僅是卸載拖車，還有你 5:06 實際上可以做最後一英里的交付，所以 5:08 你基本上可以直接從 5:10 一輛車，一輛麵包車，實際上走箱子 5:13 到一個人的房子放下它 5:15 它提供了一個非常棒的 5:17 迷人的那種，嗯，是的，最初的任務 5:20 或人形機器人的用例對 5:23 所以根據敏捷機器人他們的 5:26 一種模型或他們的機器人實際上是 5:29 大約五英尺二英寸，嗯，它是 5:32 實際上非常相似 5:34 特斯拉人形機器人的尺寸和 5:36 很有趣 嗯 體重 45 5:39 公斤 最大速度 1.5 米每秒 5:42 您的輕型運行時間為 3 5:44 小時和 1.5 的重載運行時間 5:47 小時也可以步行或 5:49 移動性，您可以向後前進 5:51 並排 轉身 蹲伏 5:53 上下傾斜穿過 5:56 非結構化領土呃地形 5:58 草岩石和遏制其他能力 6:00 是在感知到一個人時暫停還是 6:02 在它的路徑上的障礙和導航 6:04 它可以蹲下和蹲下它可以向上旋轉 6:06 身體凝視和看不同 6:08 它使用手臂和手的方向和 6:10 腳在碰撞時保持平衡現在這是 6:12 非常酷的拾取和放置能力 6:13 不同大小的物體和 6:15 重量在路緣石上上下起伏，所以 6:18 他們已經籌集了 150 個敏捷機器人 6:20 上個月 4 月 10 萬美元 6:23 22日 6:24 他們從 dc vc 提出來 6:26 全球遊樂場也是亞馬遜工業 6:29 創新基金作為投資者加入 6:31 這是他們的數字工作機器人的視頻 6:34 在行動，嗯，這很有趣 6:38 是的，所以你看到我們正在移動盒子 6:41 有點像在倉庫裡 6:43 將箱子裝載到拖車的右側 6:46 取出產品 6:48 從垃圾箱 6:50 將垃圾箱放在不同的地方 6:54 在這裡裝卡車 6:57 實際上佔用了盒子 6:59 和 7:00 嗯，是的，實際上非常令人印象深刻的東西 7:03 通過敏捷機器人在這裡 7:08 好吧，所以我認為有一些巨大的 7:10 這裡發生的事情的後果 7:12 特斯拉正在招募世界上最好的 7:15 機器人專家，他們有他們的 7:17 特斯拉的精華選擇 7:19 已經擁有最好的人工智能團隊之一 7:21 在那裡，我想我們會看到 7:23 進步比其他任何人都快得多 7:24 現在世界上的機器人團隊 7:26 其次，我們可能會看到 7:28 開發最偉大的產品 7:30 人類歷史是的，它有很多 7:31 危險和風險，所以希望他們會 7:34 是正確的監管，它將是 7:35 正確的公司將 7:36 管理這項技術以實際提供幫助 7:38 人們不要傷害寫的更大 7:40 當你陪伴或 7:43 人們通過 7:45 創新技術將會 7:46 實現其中的一些價值 7:49 如果不是通過收入獲得的很多價值 7:51 和利潤以及通過市值權 7:53 並通過股價，所以我認為這個 7:55 成為特斯拉股票的背景 7:57 未來十年和未來幾十年 7:59 有趣的是elon 8:01 說他對 2029 年的期望 8:04 我個人認為這還為時過早 8:06 認為這可能需要更長的時間，但是 8:07 嘿，我認為這是自信的標誌 8:09 和樂觀的跡象，事情是 8:11 發展得很好 8:13 特斯拉 ai fsd beta 以及 8:16 特斯拉的人形機器人沒事的希望 8:17 這很有幫助，如果它有好的 8:19 喜歡和訂閱我所有的視頻都可以 8:20 也被發現是一個音頻播客 8:22 尋找 dave lee 投資於你的 8:24 最喜歡的播客播放器我也在 8:25 twitter heyday7 好的，我們再見 8:27 我下一個視頻中的伙計們謝謝 全部可聽
|
沒有留言:
張貼留言