我都算係做過工業機械人的公司,覺得成個假設有幾點值得關注:1)好多人的工作根本人型係無非人形甘好,呢個就係一直以來非AI機械人大部分係單手機械人,固定落個地基或者路軌甘,呢個做法會降低成本集中發展向重要的地方(例如感應器),效果會好一的。2)你推介的公司例如Timkin / NSK,本身就係做工業設備的機械公司,主要業務亦係工業設備爲主(汽車同硬盤bearing),同樣產品唔會因爲你賣比AI人形機器人就特別唔同特別貴過賣比汽車同硬盤業,利潤其實有得計算同AI唔會增加好多(相反因爲大家少買硬盤多買SSD,或者少買車多Work From Home 而減少左生意)。3)個人感覺係,真正需要人形,主要係因爲幫人做野時見到個人型會舒服過見到一只怪手轆來轆去,所以可能係會先出現在客服方面,工業方面就會係本來的工業機器人加AI方向繼續進化,用數據來調教而非人手調教。4)個人感覺AI機器人最大得益者會係各種感應器公司(例如Lidar),個人覺得潛力最大係本來就做工業設備的例如Keyence, CyberOptics.
I also believe in Humanoid robot. But you have to factor in the maintenance costs on these device is fairly expensive. I think it will still take another decade to produce them in scale with reasonable maintenance costs, which might be cheaper than labour cost. PS. Kinda in this industry
Ware and tear cost has not been mentioned also part of running a robot cost. Like car dealerships make “most”money on your car maintenance, not initial sale. non replaceable job/work is the engineer for maintaining a robot.
Waymo is way ahead of Tesla and is really already driving around with no human inside in san francisco, not even backup driver. Tesla is only ahead if you want a car with no lidar sensor and only camera based. I would say even Cruise with Lidar sensors are ahead of Tesla at this point as they have been really driving around without human in San Francisco for a couple of years.
Пікірлер: 260