hexorg · HackerTrans

3 ปีที่แล้วdiscuss

Yes, but it’s much easier said than done. The article talks about FSF being shadowed by all of the funding for MIT licensed code. So funding means of making GPL code more attractive to companies probably should be one of the higher priorities. Though I’m sure companies also ensure FSF and GPL get bad publicity .

hexorg

3 ปีที่แล้วdiscuss

Yes but you can interface with Linux kernel without being required to use GPL license.

hexorg

3 ปีที่แล้วdiscuss

What the article fails to address is… We all need food. Yeah it’s great working on GPL code and ensuring it’s all open. But when companies consider your gpl library vs someone else’s mit library they will naturally go with mit. And then they’ll say “well we’re using this free library already might as well donate/fund it”. So suddenly this MIT dev is able to put way more time into the mit library than your gpl library because it becomes their job. Something that feeds them. Their library gets better faster… And more and more companies use it and fund it.

GPL is great if absolutely everyone is on board and everyone is fed. But that’s not the world we live in.

hexorg

3 ปีที่แล้วdiscuss

I generally like the approach, but you can implement sorting of fixed number of items in a Boolean circuit(combinatorial logic, a bunch of ANDs and ORs stringed together). There is no need for recurrence or memory at all to sort a fixed number or items. There are a handful of abstractions possible in Boolean logic than n-gram statistics and there are a lot of Boolean circuits that can be very useful. But I’d like to see neural networks to be able to figure out algorithm that require at least a finite state machine(FSM), but preferably something harder. The challenge is to find a size difference for a problem between FSM and combinatorial logic.

comments