3 ms·
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
by toast0 4y ago
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.
- horeszko 4y agoIs 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
- SoftTalker 4y agoI 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.
- rnk 4y agoCan 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.
- _ea1k 4y agoIts a difficult thing to do, tbh. Too many drive to produce results and its just a managed entity that often fails for the wrong reasons. Too little and it ends up being a skunkworks that has fun, produces curious things, and never makes money. Simulating real market dynamics is pretty difficult.