Hey—this is Brian from Labelbox. Thanks for the mention.
Martin and Matt are incredibly sharp on the trends in AI and have inspired our work at Labelbox immensely. I particularly like the discussion on getting the operations right for building and iterating with ML. Like software development, iterating quickly is key to building a performant model on real world data. The average iteration time for ML is 2-4 weeks in the industry right now. Comparing this to software development averages is stark. Getting the development and operations of ML right can greatly speed up iteration and improve the likelihood of getting to market with performant ML systems.
I agree, although I think Musk is thinking much more fundamentally than cars with Tesla... we as a civilization need an efficient way to produce, store, and consume energy from the sun. The car is a logical near-term path/product to advance the 'store' and 'consume' technologies while having a near-term positive environmental impact and a sustainable business model.
I also suspect that Tesla is a battery company that happens to sell cars.
But there's a problem. There is just so much R&D being done in energy storage devices and battery technology that Tesla will need to do quite a bit of M&A in order to stay relevant. That's a weird thing to do whilst you are still selling cars.
Yes. I would associate Tesla with Samsung as much as with Apple. Samsung LCD/OLED technology is in many high-end mobile device for example, even though they have flagship products that directly compete.. they are selling the more fundamental technologies. In the same way, Tesla (Elon) would much rather be selling batteries to many other car manufacturers and home battery storage manufacturers (the patent release makes a lot of sense within this line of thinking).
The challenge with energy storage is not only in the R&D, but also in the manufacturing at scale, cars are a great avenue to approach this challenge.
In the commercial airplane world 'launching' an airplane just means it can be sold to customers, not that the airplane is done. Boeing adheres to a 'gated' design process, where one of the earlier gates is the 'launch' of the aircraft and the last is entry into service. The airplane is designed far enough to close the business case and then can be 'launched'.
Looking at this from an Aerospace Engineering perspective, the weight distribution / CG properties may significantly limit the lateral movement of this vehicle; looks as such in the videos as well. This is because in order for a helicopter to move laterally it must pitch the rotors (in this case meaning the entire body) and a cg far from the rotor location makes this difficult. Something to think about.
In fact a large fuselage hanging down from below the rotor can cause a helicopter to become unstable. If the helicopter rolls, the rotor will slip sideways, gaining horizontal speed. The drag on the fuselage below the rotor will cause the helicopter to roll further, causing increased slip and increased horizontal speed, hences increased roll. To avoid this the rotors will have dihedral (they're angled upwards slightly). This is what gives stability.
This application might not need that much lateral performance, and the cg as is gives you nice static stability margin where you can relax all sorts of other requirements (control loop can be slower, the nav sensors can be lower performance, it's less sensitive to wind gusts, etc).
From an aerodynamic efficiency and maneuverability standpoint, a traditional single rotor helicopter is far superior. However, the operational benefits, safety, and mechanical simplicity (way less expensive when they crash) of the multirotor configuration make it a very popular choice. The confluence of better battery technology and cheap/lightweight controllers has allowed for increased viabiliity. As the commercial drone industry matures and payload carrying missions become viable, I expect there to be a shift in popularity back to the single rotor configuration. For example, more than 30% of Japan's rice paddy fields are sprayed with Yamaha's single rotor drone [http://textually.org/drones/2013/06/032493.htm]. There is an exception here; the multirotor configuration is a more stable platform for shooting video.
Congrats on YC! I'm excited to see what you build next.