[ExI] Useful overview of AI progress on Ourworldindata

spike at rainier66.com spike at rainier66.com
Sun Dec 11 23:43:25 UTC 2022



From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of Mike Dougherty via extropy-chat



>…I've been told recently that I have trust issues….


Heh.  Mike considering what has been revealed in just the past week, trust issues are suddenly very easy to come by.  You don’t even need to pay: we get them free.  


We have collectively taken to replacing the word “problems” with “issues” which is a slightly more general term one might suppose, with a more positive spin perhaps.  But I don’t see developing a sane distrust as having trust problems.  I see developing sane distrust as having trust solutions.  Or at least a start towards a trust solution, on the order of the alcoholics anonymous starting right out admitting the damn problem, but in this case it isn’t with the person admitting the problem, it is with outfits like Twitter 1.0 doing the kinds of things revealed this week by Twitter 2.0.


Of course we distrust T1.0, they earned our distrust and they richly deserve it.  That isn’t a trust issue, it is a trust solution.


>…Maybe. Yeah, I assume every good thing has the potential to hide the downside until after the buyer's remorse…


Mike, every damn thing REALLY DOES have the potential to hide the downside until after buyer’s remorse.  Even those things which really have zero apparent downside (such as… like… mathematics) filled with cool awe and wonder (again mathematics) have the potential to hide the downside.  We just have never yet found it.  That’s why we keep studying it Mike, perhaps to become first to discover the bad news about mathematics.  In the meantime… distrust away, me lad.  It’s not an issue, it’s a solution, to the problem of excessive trust.  I have been cured of that one, with plenty of medications to combat excessive trust coming this past week.


But I digress, for I want this thread to continue along the line of using the new tools such as OpenAI to create companion bots.  We don’t really need to jump ahead of ourselves.  We have talking avatars that are pretty good, we have OpenAI and we have speech recognition.  Those three elements should combine to make an on-screen AI driven chatbot, even if it can be at times a self-contradicting ass-kisser, in a vaguely annoying way.





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20221211/a4815a66/attachment.htm>

More information about the extropy-chat mailing list