Ok, twenty years may seem like a long time. But this, to me, is an example of why it is really hard to push through inventions in large companies. Management doesn’t think like investors do. And even when you talk to investors, you often have to talk to a lot of them to get some traction.
How would you run R&D and new business in a large company?
Facebook has spent $10 billion on Metaberse. I think running an internal accelerator, just like Y-Combinator, with internal startups (with their own budget, own teams etc, just like any startup) would have a bigger chance at getting a good result than what they did.
Like Xerox Parc, which is where much of our modern stuff came from. Mice, touch pads, GUI, etc. It produced all sorts of marvelous tech. The problem was Xerox didn't realize what they had and companies like Apple and Microsoft just copied it and ran.
> The problem was Xerox didn't realize what they had
I'd contest this (quite popular narrative): Xerox developed the Star as an highly integrated product and introduced it in 1981, 2 years before the Lisa and 3 years before the Mac. It may have been too early, though, as this was before capable microprocessors and the hardware was somewhat limited. Development started in 1977, which was about as early as PARC-based software became sufficiently optimized to make this a viable option, and the first D-processors, Dorado (rejected for its complexity and price point), became ready for evaluation. Arguably, it would have been hard to release a thoroughly engineered system like the Star significantly sooner. (There were also other GUI-driven spin-offs, outside Xerox, like PERQ, which also failed to make the proverbial dent in the universe.)
This is not all management, it is special to hereditary businesses. There is no rational justification for a series of CEOs to be descendants of the prior CEO. Motorola's third-generation heir crashed it into the ground, because he had no particular talent or qualification for the role.
Internal accelerators don't tend to work that well anyway, but you also need a different corporate structure than Facebook has. You need something where there's a lot more of distinct internal businesses, then it's easier to manage more of those. More like Yahoo or Google where there's clear groups. You also need to figure out how to let these internal startups use at least some of the corporate resources, but make their product public without mentioning the brand.
Having a big brand on a new 'startup' makes a lot of expectations and meeting those expectations reduces the flexibility of being a startup. There needs to be three clear paths for these things: shut it down, subsume it into the brand, spin it off as independent. But only one of those has clear value for the parent.
Is anyone still running the old shunkworks departments to launch new ideas/technologies? A small group off the side left to do there own thing seems like it could work
I guess the old Bell Labs or Xerox PARC are the examples that most people here might be familiar with. But just because it sometimes works well doesn't mean it always does, or even usually does. It's a risk; a lot of academic research doesn't produce anything very practical, and the same is probably true with corporate research, only in the corporate case there are a lot more people expecting a return on their investments.
Can we spin a story about Google and their failing to take advantage of AI opportunities, open AI plus Microsoft coming up crush them in search? Apparently Google hasn't really tried that hard to commercialize their AI stuff and they had worries that it would tell lies and talk about uncomfortable things like say Nazis. So Google fudged their future kind of like Xerox... and this could be what finally seriously hurts them.
How would you run R&D and new business in a large company?
Facebook has spent $10 billion on Metaberse. I think running an internal accelerator, just like Y-Combinator, with internal startups (with their own budget, own teams etc, just like any startup) would have a bigger chance at getting a good result than what they did.