That podcast is so incredibly well done. Even if we pay no attention to the tonality of the voices or the inflection or the speech patterns (a very NPR-like presentation), the writing of the "script" is top notch. I would have believed it to be human produced and presented. Maybe I'm gullible. We should find this either very worrying and fear a future in which humanity becomes a useless organic component of a bleak and sterile world or we should be hopeful of what positive developments AI might bring to our existence and to our understanding of the universe we inhabit. I am not sure what I am feeling right now. Maybe AI can help me figure it out...
@@AudioUnleashed You are much more educated than I in such things, therefore, I will defer to your assessment. Unless of course you meant cannabis, in which case I might opine.
It's a tough audiophile equipment entertainment market. I can't imagine that fantasy people having discussions about real facts will ever displace real people having discussions about fantasy. :)
That's the problem, it SOUNDS believable, and will fool a lot of people into thinking they're being reliably informed. But all it is is an imitation of intelligence, not true artificial intelligence - as usual with these 'A.I.' large language models, it gets things wrong, misrepresents the science, and omits salient information (e.g. here that the vast majority prefer the Harman target). But then, so do many highly popular actual human podcasters/TH-camrs (on which this LLM was likely trained)...in particular when it comes to this paper by Dr Sean Olive I might add.
@@Patagoona, good points, although as we discuss in depth in episode 37, we both kinda think it does a better job than most audio press of summarizing the information. But yes, the biggest concern is how many people are going to be fooled by… well, perhaps not NotebookLM, but definitely other iterations of Gemini 1.5!
Hey guys, I’m a recent supporter of your podcast. On the surface both AI hosts seems to be scarily good, but to my ear just a little too perfect ...I’m not a fan of that “fake” inhale breathing. I can see this tool being used and abused to flood the market with all kind of things, which will just drive disengagement in the long run. Take care, Brian
Dennis here! Your last point was exactly my last point in the full episode. I worry there’s going to be more A.I. than non-A.I. podcasts within a year’s time.
That podcast is so incredibly well done. Even if we pay no attention to the tonality of the voices or the inflection or the speech patterns (a very NPR-like presentation), the writing of the "script" is top notch. I would have believed it to be human produced and presented. Maybe I'm gullible.
We should find this either very worrying and fear a future in which humanity becomes a useless organic component of a bleak and sterile world or we should be hopeful of what positive developments AI might bring to our existence and to our understanding of the universe we inhabit.
I am not sure what I am feeling right now. Maybe AI can help me figure it out...
We need Camus now more than ever.
@@AudioUnleashed You are much more educated than I in such things, therefore, I will defer to your assessment. Unless of course you meant cannabis, in which case I might opine.
@@formerhobbit5599 🤣 Why not both?!
@@AudioUnleashed Make it so!
Pretty amazing.
It's a tough audiophile equipment entertainment market. I can't imagine that fantasy people having discussions about real facts will ever displace real people having discussions about fantasy. :)
@@status101-danielho6 🤣 Good points!
Remaster it in fake stereo!
On video ✌🏼
@@eddietours3728 😬
Wow that podcast sounds so real. Bit bland but very believable. Insert some more humour and that's it...
That's the problem, it SOUNDS believable, and will fool a lot of people into thinking they're being reliably informed. But all it is is an imitation of intelligence, not true artificial intelligence - as usual with these 'A.I.' large language models, it gets things wrong, misrepresents the science, and omits salient information (e.g. here that the vast majority prefer the Harman target). But then, so do many highly popular actual human podcasters/TH-camrs (on which this LLM was likely trained)...in particular when it comes to this paper by Dr Sean Olive I might add.
@@Patagoona, good points, although as we discuss in depth in episode 37, we both kinda think it does a better job than most audio press of summarizing the information.
But yes, the biggest concern is how many people are going to be fooled by… well, perhaps not NotebookLM, but definitely other iterations of Gemini 1.5!
Hey guys, I’m a recent supporter of your podcast.
On the surface both AI hosts seems to be scarily good, but to my ear just a little too perfect ...I’m not a fan of that “fake” inhale breathing.
I can see this tool being used and abused to flood the market with all kind of things, which will just drive disengagement in the long run.
Take care,
Brian
Dennis here! Your last point was exactly my last point in the full episode. I worry there’s going to be more A.I. than non-A.I. podcasts within a year’s time.
That is pretty amazing and believable. As others have said, a bit bland and not to my taste, but very well done.
@@jimf5160 imagine how much more believable it’s gonna be in a year!
@@AudioUnleashed i shudder to think...good thing I hold Nvidia stock...