AI/ML research presents an interesting case study of a field that has to contend with the limitations you're outlining here.
There's been a longstanding naivety in the field about how straightforward it is to formalize information (as understood by humans) into data (as readable by machines).
One of the reasons that this is so hard is that when trying to convert real-world, common sense, human intuitive information into tabulated data via sensors so that a computer can process it, the information loss is absolutely massive.
This is because when humans process 'data' in their minds, they're in fact doing this on a massive, opaque, complex, changing substructure of assumptions, experiences, intuitions, inductions, deductions, biases, ideologies etc. and they're pulling in all kinds of contextual information and memories that computers have no access to. This allows humans to use 'common sense' to choose a single, coherent interpretation of a situation out of a theoretically limitless set of possible interpretations.
All computers can ultimately do is statistically manipulate these incredibly reductive datasets, and then researchers can crudely tack this onto a pre-set research or operational context.
It turns out that 'thinking' is much more art than science.
I have held for a long time that we are nowhere near AI singularity like the rats fear. this is bc the first and most important rule that undergirds all living beings is “optimize the spread of your genes/memes across space and forward through time” (eg “survive”). Without this, a hierarchy of values is impossible to instantiate and without a hierarchy of values thinking and decision making are impossible to instantiate.
TLDR If you do not start an AI upon the premise that its first objective is to survive, it will never be able to truly think. And no one who is pro-human would ever give an AI this directive.
We of course may only be one lone misanthropic genius away from human annihilation one day, but we are decades if not centuries from that opportunity.
I just want to thank you for your articles on here and on Actualization Hub. It's full of interesting stuff I wanted to look into, but never had the time. Continue on!
This is one of the most idiotic things i've ever read in my life - yeah, apparently now the ONLY reason one might be interested in civil rights, democracy, and what have you is because God told you so (that's called authority fallacy btw), not because you believe these things create a better society, because you care for your fellow human beings, not because you're pragmatic and think these work the best in a system, not because of a personal belief or philosophy, nope: It's the equivalent of evoking God.
Also, congrats on using the "communism when iPhone" argument unironically. That's both a strawman and failing the basic assumptiom (that you need capitalism for iPhones) at once.
I looked into this and I dont see a way to just automatically enable audio for the posts. I only see the ability to manually record and upload a text to speech recording. If there is just like a box I can click to turn this on can you find the documentation and send to me?
BTW there are plenty of voice to text readers you might be interested in. I use the one built into Pocket to have other peoples content to me and I use Natural Reader Text to Speech Chrome extension to read my own content to myself for correcting typos. Might be worth looking into.
AI/ML research presents an interesting case study of a field that has to contend with the limitations you're outlining here.
There's been a longstanding naivety in the field about how straightforward it is to formalize information (as understood by humans) into data (as readable by machines).
One of the reasons that this is so hard is that when trying to convert real-world, common sense, human intuitive information into tabulated data via sensors so that a computer can process it, the information loss is absolutely massive.
This is because when humans process 'data' in their minds, they're in fact doing this on a massive, opaque, complex, changing substructure of assumptions, experiences, intuitions, inductions, deductions, biases, ideologies etc. and they're pulling in all kinds of contextual information and memories that computers have no access to. This allows humans to use 'common sense' to choose a single, coherent interpretation of a situation out of a theoretically limitless set of possible interpretations.
All computers can ultimately do is statistically manipulate these incredibly reductive datasets, and then researchers can crudely tack this onto a pre-set research or operational context.
It turns out that 'thinking' is much more art than science.
Yes, great point and observation. Kinda related:
I have held for a long time that we are nowhere near AI singularity like the rats fear. this is bc the first and most important rule that undergirds all living beings is “optimize the spread of your genes/memes across space and forward through time” (eg “survive”). Without this, a hierarchy of values is impossible to instantiate and without a hierarchy of values thinking and decision making are impossible to instantiate.
TLDR If you do not start an AI upon the premise that its first objective is to survive, it will never be able to truly think. And no one who is pro-human would ever give an AI this directive.
We of course may only be one lone misanthropic genius away from human annihilation one day, but we are decades if not centuries from that opportunity.
Nit:
> ignore what people, trust only what they do
Overall: loved it.
I just want to thank you for your articles on here and on Actualization Hub. It's full of interesting stuff I wanted to look into, but never had the time. Continue on!
thank you king! AH will definitely take priority for a bit, but Minor Dissent will continue on.
This is one of the most idiotic things i've ever read in my life - yeah, apparently now the ONLY reason one might be interested in civil rights, democracy, and what have you is because God told you so (that's called authority fallacy btw), not because you believe these things create a better society, because you care for your fellow human beings, not because you're pragmatic and think these work the best in a system, not because of a personal belief or philosophy, nope: It's the equivalent of evoking God.
Also, congrats on using the "communism when iPhone" argument unironically. That's both a strawman and failing the basic assumptiom (that you need capitalism for iPhones) at once.
May you please add tts audio that substack provides, to all your posts.
hey oldmanflappynuts,
I looked into this and I dont see a way to just automatically enable audio for the posts. I only see the ability to manually record and upload a text to speech recording. If there is just like a box I can click to turn this on can you find the documentation and send to me?
BTW there are plenty of voice to text readers you might be interested in. I use the one built into Pocket to have other peoples content to me and I use Natural Reader Text to Speech Chrome extension to read my own content to myself for correcting typos. Might be worth looking into.
we like autistic representation (complete with textbook synopses) in our christian allegory apocrypha
the main disappointment with the original harry potter books is we never got to join hermione in arithmancy class
what did he mean by this????
(but seriously, I am retarded and don't understand what you're saying)