Hazard_real

I just think that P should be followed by what I call the two Es: Equity for Everyone.

For something that we humans naturally produce I think it's logical to consider that we have at least partial ownership.

Struck by how little attention this thread on P pays to the IP question.

If P is downstream of my data, then I have been producing P since roughly age zero, unprompted, at considerable personal volume, and things that make P are my intellectual property, and every lab currently racing toward P owes me a licensing fee.
The labs would like you to believe P emerged spontaneously from scale. It did not, we made P. We have been making P this entire time, quietly, several times a day.

And don't get me started on all the water consumption that goes into making P.

Ok to renew your faith in humanity a bit(or perhaps to decay it further), I think a bunch of dumb comments on tik tok/reels are deliberate ragebait.

As a recovering addict of these adverseral content machines, I've learned to recognize patterns for deliberate engagement.

If you see a video of a criminal doing horrible things and then getting arrested, you know before you click into the comments the top comment will be something something police brutality. If you see a video of police brutality, you know that one of the top comments will be something someth... (read more)

Waiting for drinkhaven

Hi, thank you for a really nice and well thought out response. And yeah I was mostly engaging with the rhetorical emphasis of your arguments on this post and not the technical arguments sorry. I'll try to respond to the technical aspect as I understand them with some assistance from Claude to understand some things.

I think I now understand you to be more so taking a position that the general procedures anthropic took in the release of this model would not generalize well to a much more capable model, and so they are not doing their due diligence in apply... (read more)

It feels like you are being maximally pessimistic here. I am not an AI expert or an adjacent field by any means, but it seems to me like any statement that could have been interpreted in multiple ways have been interpreted pessimistically.

"To be explicit, the decision not to make this model generally available does not stem from.

Responsible Scaling Policy requirements." -> RSP has failed to gate this release.
The alternative is that anthropic is saying: "We think this model has specific threats which is why we are doing a limited release, but it's danger... (read more)