With intellectual labor; the effectiveness of an org scales roughly logarithmically with the number of people in it. This means 1000 people may only be 2x more effective than 100. And a company of 10,000 people can actually be 25x less effective than 100 companies of 100 people.
-
-
Replying to @DavidSHolz
Checks out empirically for most "innovative" software products, but scale also enables new kinds of projects. Communication scales logarithmically but infrastructure scales super-linearly. You couldn't build a 5th generation fighter jet with 100 orgs of 100 people.
4 replies 0 retweets 41 likes -
Replying to @fchollet
The 5th generation fighter jet example is provocative but also a bit turbulent for me! The aero industry seems to universally feel the f22 and f35 programs were boondoggles. Maybe would have been better with something more distributed?
1 reply 0 retweets 1 like -
Replying to @DavidSHolz
Distributed small orgs work best when each one can occupy a niche for which there is a reasonably-sized market. So you could build billion-dollar highrises with small orgs. You need a big org when you're building very large systems made of components for which there is no market.
1 reply 0 retweets 3 likes -
Replying to @fchollet
The key thing here is the importance of infrastructure and long term investments. I don't think this has to be bankrolled inside of giant companies. What if we taxed SW companies to fund public open source the same way we tax people to fund roads?
1 reply 0 retweets 2 likes -
Replying to @DavidSHolz @fchollet
I think this might sort itself out: all big companies rent out their datacenter to smaller companies to do cloud computing on them. Also big companies contribute to open source in various ways already (financing, code-contributions, support, etc).
1 reply 0 retweets 0 likes -
Replying to @ChrSzegedy @fchollet
It's true that there's a private-ish version of this already. I still wonder what would happen though if web hosting was a public utility and open source was funded by taxes. It's not obvious to me that we wouldn't be in a better place.
2 replies 0 retweets 1 like -
Replying to @DavidSHolz @fchollet
I have a distrust of politicians setting priorities of open-source software. I think it could work better with semi-private/non-profit entities building a better demand-based resource allocation system for open-source software.
1 reply 0 retweets 0 likes -
Replying to @ChrSzegedy @fchollet
Very true. I'd love to hear your thoughts about how better resource allocation systems could work. You're a prolific and inspiring researcher
1 reply 0 retweets 0 likes -
Replying to @DavidSHolz @fchollet
Thanks. I think of a system that pools money for a market place in which people could bid for tasks to be solved according to priorities of those people who have put money into the system. For example one could set a price on implementing some feature and many could chip in.
1 reply 0 retweets 4 likes
That's a cool idea. Kinda like OSS donations, but linked to specific features.
-
-
Replying to @fchollet @ChrSzegedy
How would this work for larger projects? For example how would a public resource allocation system fund a major infrastructure system from scratch?
1 reply 0 retweets 1 like -
Replying to @DavidSHolz @fchollet
This is speculation, but what I would envision is system which still requires some initiative without compensation. Alternatively high-profile, reputable contributors could make a public plan with clear deliverables and it would be crowd-sourced.
1 reply 0 retweets 1 like - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.