Ediz Ucar(Ediz Ucar)
Thanks for this summary. I didn’t find the examples of irreducible categories all that convincing. I think that, although more difficult than the average category (e.g. certain physical objects), there are underlying traits that we are pointing at for these.
For example:
Teams or clubs: BG give the particularly crisp example of Pokemon GO, where players choose to associate with one of three teams that are functionally indistinguishable except for their respective color.
The trait here is not within the group but in how it interacts with other groups. The trait is that they collaborate against other teams.
More generally, this argument for self id gender seems to also work for self id racial/ethnic categories. The received narrative would rebut this by reference to the fact that this phenomena is not present among people to the extent that gender dysphoria is. What do you think BG’s position on this is?
The thing that we care about is how long it takes to get to agents. If we put lots of effort making powerful Oracle systems or other non-agentic systems, we must assume that agentic systems will follow shortly. Someone will make them, even if you do not.
if you don’t do RL or other training schemes that seem designed to induce agentyness and you don’t do tasks that use an agentic supervision signal, then you probably don’t get agents for a long time
Is this really the case? If you imagine a perfect Oracle AI, which is certainly not agenty, it seems to me that with some simple scaffolding, one could construct a highly agentic system. It would go something along the lines of
Setup API access to ‘things’ which can interact with the real world.
Ask the oracle ‘What would be the optimal action if you want to do <insert-goal> via <insert-api-functions>?’
Do the actions that are outputted.
Some kind of looping mechanism to gain feedback from the world and account for it.
This is my line of reasoning why AIS matters for language models in general.
Correlation or causation?
Could you site the studies that this section was based on. I would be interested in reading further as this seems to be the sticking point for most people when it comes to the topic of GM for embryos.
Perhaps I’ve missed the point of your post, but to me the whole confusion around Gender is not internal validity, after all circular definitions are valid—but not convincing to the outside view.