![](/static/66c60d9f/assets/icons/icon-96x96.png)
![](https://fry.gs/pictrs/image/c6832070-8625-4688-b9e5-5d519541e092.png)
I’ve always heard them described as seagull managers. Screams loudly, shits everywhere, leaves.
I’ve always heard them described as seagull managers. Screams loudly, shits everywhere, leaves.
Copyright has little to say in regards to training models - it’s the published output that matters.
The UNIX philosophy isn’t about having only one way to do things - it’s about being able to use tools together. The deliberately simple interface is what makes it so powerful - almost any existing too can become part of a pipeline. It’s adaptable.
Something transformative from the original works. And arguably not being being distributed. The model producing and distributing derivative works is entirely different though. No one really gives a shit about data being used to train models - there’s nothing infringing about that which is exactly why they won their case. The example in the post is an entirely different situation though.
Using it to train on is very different from distributing derived works.
I thought the point of the LGPL was to allow this sort of usage without requiring the release of source code. It’s an extension of the GPL to remove those requirements isn’t it?
Why does the prompting matter? If I “prompt” a band to play copyrighted music does that mean they get a free pass?
I don’t think so. Those users had opted in to share information within a certain group. They’ve already accepted the risk of sharing info with someone who might be untrustworthy.
Plenty of other systems do the same thing. I can share the list of games on my Steam account with my friends - the fact that a hacker might break into one of their accounts and access my data doesn’t mean that this sharing of information is broken by design.
If you choose to share your secrets with someone, you accept the risk that they may not protect them as well as you do.
There may be other reasons to criticise 23andMe’s security, but this isn’t a broken design.
Sure, there’s always going to be outliers. Most people live and work in the same metropolitan area though - they’re not driving 50,000km+ a year. Besides, having a vehicle with 5 times the effective lifetime is going to be a big win regardless of how much you drive it.