@connerdelights Add @fchollet's @ndeainc
This would be great - esp if it included a stream of recent updates or blog posts from each. Rn I just have their sites bookmarked and check regularly.
@connerdelights Taking this a step further - do you think we'll develop a spectrum of "guilt levels"?
If we can determine someone did it, but was only partially in control of themselves or felt so desperate they needed to commit said crime - would that change the sentencing? Should it?
Communication changes as it becomes faster (horse, telegram, phone...) and alters society in significant ways.
As technology is further intertwined with our lives, how might it systemically change the nature of society and how humans operate and interact?
As brain scans increase in resolution, will we be able to identify the seat of "identity"? If so...
Is there a specific place for it?
Is it a unique set of connections?
Or does it differ from person to person?
Are there mental equivalents of affordances?
Could we build them into our day to habituate healthy mental habits that could support things like an unshakeable identity?
What benchmark AI achievements will most challenge our identity as humans?
Worth reflecting on IMO as this will likely highlight where our identities lie.
When defining concepts like intelligence and identity, is it helpful to distinguish between defining these for individuals vs humanity as a whole?
I don't think this is a distinction without a difference, but I'm not sure it's a qualitatively different distinction.
@datasociety@SofiQuaglia@SofiQuaglia - curious how you think about the broader implementation of this. If Google Translate has policies on its usage, how would they alert users to this? Especially since so many of us (myself included) often ignore the TOS/guidelines.
It would seem like developing human level AI requires training models on scenarios where humans fundamentally disagree on the right or correct outcome.
How will we respond when it comes a nonobjective conclusion we disagree with?
As communication grows and makes us increasingly intertwined as a species, this dense interlinking would seem to manifest some form of greater identity.
Do we have language for talking about our shared identity as a human species (ie humanity's identity)?
Reasoning is an important milestone for AI. But what is the difference between reasoning and memorizing an absurd amount of information?
It seems like one of the reasons we developed reasoning is almost as a coping mechanism for having a very limited working memory.
Is "human level" AI even the right way to think about it?
Is it even possible to effectively evaluate this?
Human-level in which domain?
And over how many domains until we feel seriously challenged by it?
If we want to make technology more human, we need clarity on what we mean by more "human".
This is easier with areas that have objective answers, but what about areas that are subjective or are layered with multiple meanings and emotion?
@mwjcoughlin@AmandaAskell Ah - good catch. Yeah, definitely unaware.
When you have them 'paint a picture', do you think they realize what they're illuminating for you?