The Skeptics Guide #930 - May 6 2023 (Ad Free)
Added 2023-05-06 15:35:05 +0000 UTCInterview with Brian Brushwood; What's The Word: Catacoustics; News Items: AI Mind Reading, 10,000 steps per day, 30 Years of the Web, When Will Aliens Contact Us; Who's That Noisy; Your Questions and E-mails: Carnivore Diet; Science or Fiction
Comments
The mind reading tech plus elons neural-link tech combined would be a first step toward human telepathy, thought activated smart phones and cars. The ultimate bio-signature since each brain is unique that trying to copy one brain thought patterns would be so difficult. But on the flip side, governments would definitely be spying on us and our thoughts all the time.
Brian Grimes
2023-05-10 13:45:03 +0000 UTCRe: 10K steps: I recently started listening to the Just One Thing with Michael Mosley podcast, it's about just one thing you can to today to improve our health. An early episode was about walking briskly daily (particularly in the morning); 30 minutes a day, can be broken up. https://podcastaddict.com/just-one-thing-with-michael-mosley/episode/121223528
Asymetra
2023-05-09 22:33:04 +0000 UTCAs a person who regularly has to take "lie" detector tests, I can assure you, I have full confidence that the gov't will use this "mind-reading" technology at the very first opportunity to "detect" lies, if not sooner. The gov't is very much in love with being conned by pseudoscience. Any test that results in garbage will be viewed as "evidence" of attempting to subvert the results and proof of lying.
Asymetra
2023-05-09 22:18:38 +0000 UTCI remember about 10-15 years ago the development of "mind-reading" technology that was able to create rough B&W images of what the subject was visualizing.
Asymetra
2023-05-09 21:26:49 +0000 UTCWhen I bring up skepticism, I say, "practicing skeptic, I follow the evidence."
Asymetra
2023-05-09 21:24:02 +0000 UTCThat would be an interesting experiment. I doubt it would work. I think a good analogy is noise reduction on audio. I "train" the software on a sample of one audio track, then do noise reduction. If I accidentally do noise reduction on a different track without sampling that track, it distorts the audio in that track. I don't know how to sample multiple tracks, but if I sample a combined track, again it creates distortion in the audio. I suspect that if you trained the AI on multiple people the result would be the average of all the individuals, but that average would still be gibberish to any one individual's brain activity. Perhaps it might learn to identify different people within the training data. But could that data be used to read a new individual not trained on? I would highly doubt it. Your assumption is that there is readable commonality among all or most people in terms of their brain activity, and at that level of detail this may simply not be true.
The Skeptics' Guide To The Universe
2023-05-09 11:08:31 +0000 UTCHey guys, feels like I have to share my nostalgic experience on Cara's join! I started to listen to the podcast at ~500th episode - pretty much at the edge of her join, so for me it has always been you five. But then with all your references to the past I had to get the gist of dear Perry - so I started to listen from episode #1. Then I think it was #426 where I (OMG!!!) met Cara with you for the first time in her interview, Please do forgive me for committing this bias, but in hindsight it was so obvious:-) Anyhow, wanted to comment on the 'AI Mind Reading' item - where I think you are committing a fallacy. From the Item I understand that the algorithm was trained on one person and then failed when tested on another person - which I don't think is too surprising. But then you assumed the differences between people are such that it may not likely be possible to use it for reading the mind of a person it was not trained on. I think this only means that the experiment was not designed to do this inference. I think we need to see what happen when you train GPT4+ with data from hundreds of people - then expose it to a mind it has never encountered before. Is there a skeptic reason to assume that with high confidence such experiment would be negative? With my highest appreciation, love you and all that you do, you are my heroes :-)
Maoz Maylat
2023-05-06 20:57:43 +0000 UTCThe word "skepticism", much like the word "socialism" is too entrenched to give it up, so we need to keep defining it to people - over and over again, probably for all of time. Personally, I like to call it "evidence based thinking".
Ted Apelt
2023-05-06 18:38:32 +0000 UTC
