Is this article worth reading? Because stuff like this: sets off all the alarm bells...
Is this article worth reading? Because stuff like this:
Karl Friston’s free energy principle might be the most all-encompassing idea since Charles Darwin’s theory of natural selection. But to understand it, you need to peer inside the mind of Friston himself
I don't know if this will work out for making GAI. If evolved intelligence is fundamentally 'free energy' based and not a series of hacks and cludges to run heuristics targeting the evolutionary...
I don't know if this will work out for making GAI.
If evolved intelligence is fundamentally 'free energy' based and not a series of hacks and cludges to run heuristics targeting the evolutionary environment I'll be mighty surprised though.
A rather long and interesting article that's more about the man and the story than it is the concept. Would have been nice to see more of the practical application, but while reading through the...
A rather long and interesting article that's more about the man and the story than it is the concept. Would have been nice to see more of the practical application, but while reading through the page I couldn't help but repeatedly think "yeah, this makes perfect sense, at least hypothetically."
I'm not sure it's truly a "novel" idea, however, as behaviorists have, for some time, watched people "test" the bounds of behavior. What can I get away with? What will maximize my gain? I see the theory of free energy as an extension of this - in multivariate analysis we have a lot of variables and we often can't know which will result in the best optimization, so the best goal is to repeatedly test them in different configurations until we arrive at the desired result. It's a bit more complicated than that, however, as we can have inputs that are never constant, and desirable outcomes that are never constant as well.
He's a fascinating individual and I hadn't heard some of this backstory before (or how it relates to his theory), so I found it a fun read.
From what I read, he seems to have practical knowledge of using the math to create technology, which to me gives him some credibility even though I have no chance of understanding it. That others...
From what I read, he seems to have practical knowledge of using the math to create technology, which to me gives him some credibility even though I have no chance of understanding it. That others are seeking him out on some of its applications at least means some others are taking his thoughts seriously.
Yes, the article might have been a little too playful, especially with references to psychohistory and Hitchhikers Guide to the Galaxy. But I found it interesting enough to look up some of the wikipedia articles on Bayesian Inference and Free Energy.
I'm definitely not trying to downplay his significance in the field - he's a giant. I'm just pointing out that the article acts like it's this "brand new" idea. It's really not that novel. Or at...
I'm definitely not trying to downplay his significance in the field - he's a giant. I'm just pointing out that the article acts like it's this "brand new" idea. It's really not that novel. Or at least I felt like the article was treating it like it was more novel than it was. I dunno, I guess it's an odd bone to pick and doesn't really matter that much.
Ha! Pick away! The more background on things like this the better. I had issues with the article too, but having never heard of the guy I thought it was interesting enough to put out there. Thanks...
Ha! Pick away! The more background on things like this the better. I had issues with the article too, but having never heard of the guy I thought it was interesting enough to put out there. Thanks for your feedback.
An interesting article on Karl Friston, a prolific neuroscientist who uses Bayesian mathematics as a basis for the Free Energy Principle. I don't pretend to understand Bayesian inference but the...
An interesting article on Karl Friston, a prolific neuroscientist who uses Bayesian mathematics as a basis for the Free Energy Principle. I don't pretend to understand Bayesian inference but the article makes the case for Friston having formulated a very broad and useful approach to AI and many other areas.
‘Free energy’ sounds a lot like the ideas of Jurgen Schmiduber and his ideas on maximizing the compressibility of models. I can’t really tell though because the author of this article isn’t...
‘Free energy’ sounds a lot like the ideas of Jurgen Schmiduber and his ideas on maximizing the compressibility of models. I can’t really tell though because the author of this article isn’t technical or knowledgeable enough to clearly formalize the ideas they’re attempting to relay. If you can’t make the idea clear for your reader, find someone who can. Just because you decided to become a journalist it doesn’t give you a free pass to peddle mere flirtations with ideas as something worth reading.
Is this article worth reading? Because stuff like this:
sets off all the alarm bells...
Yes, he's pretty respected. But the article is light in the math. Quite a mind.
I don't know if this will work out for making GAI.
If evolved intelligence is fundamentally 'free energy' based and not a series of hacks and cludges to run heuristics targeting the evolutionary environment I'll be mighty surprised though.
A rather long and interesting article that's more about the man and the story than it is the concept. Would have been nice to see more of the practical application, but while reading through the page I couldn't help but repeatedly think "yeah, this makes perfect sense, at least hypothetically."
I'm not sure it's truly a "novel" idea, however, as behaviorists have, for some time, watched people "test" the bounds of behavior. What can I get away with? What will maximize my gain? I see the theory of free energy as an extension of this - in multivariate analysis we have a lot of variables and we often can't know which will result in the best optimization, so the best goal is to repeatedly test them in different configurations until we arrive at the desired result. It's a bit more complicated than that, however, as we can have inputs that are never constant, and desirable outcomes that are never constant as well.
He's a fascinating individual and I hadn't heard some of this backstory before (or how it relates to his theory), so I found it a fun read.
From what I read, he seems to have practical knowledge of using the math to create technology, which to me gives him some credibility even though I have no chance of understanding it. That others are seeking him out on some of its applications at least means some others are taking his thoughts seriously.
Yes, the article might have been a little too playful, especially with references to psychohistory and Hitchhikers Guide to the Galaxy. But I found it interesting enough to look up some of the wikipedia articles on Bayesian Inference and Free Energy.
I'm definitely not trying to downplay his significance in the field - he's a giant. I'm just pointing out that the article acts like it's this "brand new" idea. It's really not that novel. Or at least I felt like the article was treating it like it was more novel than it was. I dunno, I guess it's an odd bone to pick and doesn't really matter that much.
Ha! Pick away! The more background on things like this the better. I had issues with the article too, but having never heard of the guy I thought it was interesting enough to put out there. Thanks for your feedback.
An interesting article on Karl Friston, a prolific neuroscientist who uses Bayesian mathematics as a basis for the Free Energy Principle. I don't pretend to understand Bayesian inference but the article makes the case for Friston having formulated a very broad and useful approach to AI and many other areas.
‘Free energy’ sounds a lot like the ideas of Jurgen Schmiduber and his ideas on maximizing the compressibility of models. I can’t really tell though because the author of this article isn’t technical or knowledgeable enough to clearly formalize the ideas they’re attempting to relay. If you can’t make the idea clear for your reader, find someone who can. Just because you decided to become a journalist it doesn’t give you a free pass to peddle mere flirtations with ideas as something worth reading.