4 ms·
A lot of his talk was wasted on irrelevant complaining about lack of capex in R&D. That's only partly correct. Any one of us can afford to rent a crazy amount o
by Zigurd 13y ago
A lot of his talk was wasted on irrelevant complaining about lack of capex in R&D. That's only partly correct. Any one of us can afford to rent a crazy amount of computing power and storage on demand. Pfft.
In short, skip the first 20 minutes. He's being a grumpy old man. In the second part, he's a pissed-off genius and revolutionary.
- dropit_sphere 13y agoSure, but do they have money to live on while they're experimenting?
- Zigurd 13y agoHis argument that 5 year timespans are minimal for invention is on target, but that's opex, even when you are cash flow negative. Where he is wrong is that you need large equipment capex for invention. Unless you are building a novel special purpose computing device, like an giant FPGA cryptocurrency miner, you really don't need more than $5k capex per coder for anything anymore, and renting Web-scale power is cheap.
- nkuttler 13y agoComputing power and storage makes the computer of the future? If you think like that, "new" things can only come from storing more data and processing it in new ways. That seems rather limiting... Off the top of my head, I'd expect the "computer" of the future to know a lot more about my immediate surroundings. That means sensors, not big data, computing power and storage are useless. Can I get sensors that are 20 years ahead of the stuff that's mass-produced today? Maybe, but it's certainly not something I can rent for cheap.