Start, restart, or continue your college journey with Study Hall! Take a college course that starts on RUclips. Go to link.gostudyhall.com/sst to learn more.
Next up on CSI: Las Vegas, the intrepid team has to use anal pattern prints to find the killer!!! It's positively coprolitic! (The adjectival form of coprolite there is an educated guess....)
12:26 There was an entire episode of Star Trek: TNG that highlighted the fact that Data couldn't really do humor except retelling jokes that other people had told before. So they nailed it. 🤣
I have a clarification for the noisy plants! I'm pretty certain that this paper is describing the same ultrasonic vibrations from an earlier study that have previously been taken out of context for the myriad of "plants scream out in pain" pop sci articles that people love to send vegans as a gotcha (after being sent these articles enough times I eventually decided to go and read the paper's abstract for myself). It turns out, according to those original researchers, what the microphones are picking up is actually noise from microcavitation resulting from air bubbles getting trapped in plants' xylem. Which makes sense, if a plant is drought stressed or losing water rapidly to an open cut, some air will end up getting sucked up, like when you're drinking something with a straw and reach the bottom of the drink, and these are tissues that are *not* meant to transport air so the movement is pretty rough, and so it produces an inaudible noise. So not screaming, just a mundane physical process
As a data scientist with a background in psychology I am SO excited for this!! And can't wait to share with my colleagues who are interested in AI but don't have a background in data. 🥰
i could tell immediately that the third paper was real because i didn’t understand a word, and if there’s one thing i’ve learned about cs papers it’s that they are not meant to be read by people who aren’t in cs
I genuinely was overjoyed to see my question being chosen. That was a very detailed response. And watching Hank's "What comes after a finger? A finger!" was delightful. Thank y'all very much! (Sam, you're brilliant. I pronounce it Poe Knee Yo Tee, but I admit it's not the easiest handle)
@42:45 Is what Ceri's describing those CGI videos I remember watching years ago of those squishy cubes "randomly" falling into place in a pile to form a specific image? They really begin with the cubes arranged in the finished image and reverse them to make it look like they fell that way by chance
i'm a grad student who pays their bills by grading undergraduate papers. I don't take off points if i suspect someone has used a chatbot (although the chatbot still needs to meet the brief of the assignment) because it's very hard to actually prove but after a while you start to notice patterns. The AI paper in the Truth or Fail very quickly registered to me as being generated. basically, AI is bad at writing academic papers but it is bad in an unusual way that is different from the ways scientists are usually bad at writing. scientists tend to write extremely information dense sentences with confusing grammatical structure whereas AI will write very flowery sentences that sound great but convey almost no actual information beyond vague marketing babble. That being said. prompting the AI with more information can help so i'm sure i've read things co authored by chat gpt without realizing it. if anyone out there is thinking of using it for writing just remember the less information you give the bot the more it will fall back on empty garbage.
Haven't watched/listened yet but had to let you know when I saw that title I smushed Jordan and Harrod together to make Jarrod which was a frightening thought to have him on SciShow Tangents!
My craziest "my phone is listening to me" moment to date happened just 2 days ago. At 3PM, I was at physical therapy, where my PT recommended that I try ice therapy for an injured knee. Her actual recommendation was to use a styrofoam cup ("bad for the environment" she admitted as I frowned scornfully) but she described what to do for some time. Immediately after PT, I had to go attend a Christmas party. I came home around 9PM, realized that I needed to do some last minute Christmas shopping, and of all things for Amazon to recommend to me, "ice therapy cup" was on my homepage's recommended products. HOW??? I was too busy at the party to even look into this, and I had never even heard of this type of therapy until just 6 hours earlier. I've not purchase any sort of PT equipment on Amazon for over 6 months, so it's not like this would be in my Recommends. It "just so happened" to show up 6 hours after my PT and I talked about it for a good 5 minutes. So of course I took a screenshot and shared it on Facebook to watch my Boomer relatives panic over "duh gubbmint" listening in, haha!
Because her name sounds so similar to Siri, anytime Ceri's name is said the Siri picks it up (and her name is often mispronounced as Siri instead of seh-ri)
I wonder, since we've gotten as far as we have with "hand made" A.I... Have we made the next step into "industrialization" of A.I.? Like an A.I that specializes in data-set 'noise' minimization paired with an A.I. that's designed to build data-sets ....et cetera. Like an A.I. suite that's intended to other create end user specific A.I?
I wanted to make a cute AI drawing of a cat with and eyepatch (cos a friend of mine her cat had to had an eye removed). But it reallyyy does not understand eyepatches. 🤣🤣
Just think of the computational power it would require to constantly process all audio received from your phone/alexa/whatever, the level of profit necessary to justify that expense, and the margins on the consumer products theyre supposedly using this data to sell you. It makes 0 sense. Now a targeted attack against a specific individual, better yet targeted to specific times with a suite of relevant “wake” words. That sounds entirely possible if not likely to be happening already
100% wrong about the mics not hearing and targetting. I was talking to a friend about my X being black, and the next time I went on Grindr, a whole bunch of profiles from Black Men showed up near me, there area I live in, that's not a possible reality. So the app populated what it think I wanted to see, based on what it heard earlier in the day.
Hank reading the raccoon optimization algorithm. Me trying to stay tuned in. A standard was set and this is not a glowing space racoon a genius may have had sex with or been abducted by so I am not interested.
Start, restart, or continue your college journey with Study Hall! Take a college course that starts on RUclips. Go to link.gostudyhall.com/sst to learn more.
waiting for butts
Thanks for having me y'all!
thank you for showing up!❤❤
You were an awesome addition to the episode. Thank you for guesting!
Thank you for being an awesome #womaninstem and giving me yet another awesome person for my daughters to look into! 🎉
Loved your fact and insights! 🏆
Next up on CSI: Las Vegas, the intrepid team has to use anal pattern prints to find the killer!!!
It's positively coprolitic!
(The adjectival form of coprolite there is an educated guess....)
“I would answer any question to be emotionally stable” you are so real for this
12:26 There was an entire episode of Star Trek: TNG that highlighted the fact that Data couldn't really do humor except retelling jokes that other people had told before. So they nailed it. 🤣
I have a clarification for the noisy plants! I'm pretty certain that this paper is describing the same ultrasonic vibrations from an earlier study that have previously been taken out of context for the myriad of "plants scream out in pain" pop sci articles that people love to send vegans as a gotcha (after being sent these articles enough times I eventually decided to go and read the paper's abstract for myself). It turns out, according to those original researchers, what the microphones are picking up is actually noise from microcavitation resulting from air bubbles getting trapped in plants' xylem. Which makes sense, if a plant is drought stressed or losing water rapidly to an open cut, some air will end up getting sucked up, like when you're drinking something with a straw and reach the bottom of the drink, and these are tissues that are *not* meant to transport air so the movement is pretty rough, and so it produces an inaudible noise. So not screaming, just a mundane physical process
I love when you have interesting guests on the show!
Sam "lemme ask the stupid question" *proceeds to ask a really complex and interesting question* 😂
Yoooooo!!!! Another amazing episode coming up. Sadly I can’t listen to it before going to bed tonight, but, there’s always tomorrow. Have fun guys!!
"Instead of Mars we made Facebook." That's going to stick with me for a bit...
Same here
As a data scientist with a background in psychology I am SO excited for this!! And can't wait to share with my colleagues who are interested in AI but don't have a background in data. 🥰
i could tell immediately that the third paper was real because i didn’t understand a word, and if there’s one thing i’ve learned about cs papers it’s that they are not meant to be read by people who aren’t in cs
The best “butt one more thing” yet!! 🔆😂😂
I'm happy to know that my butthole is as unique as my fingerprints.
What a perfect poem for the scishow tangents poetry book coming soon!!!!*
*By coming soon I mean... please?
Ceri, you could never disappoint us!
as an AI they would just have to say "listen im not angry im just disappointed." honestly
New Tangents!!!!!!!!!!!!!!!!!
I genuinely was overjoyed to see my question being chosen. That was a very detailed response. And watching Hank's "What comes after a finger? A finger!" was delightful. Thank y'all very much!
(Sam, you're brilliant. I pronounce it Poe Knee Yo Tee, but I admit it's not the easiest handle)
Tangents is my comfort podcast 🥰
Mine too❤
@@denisedavis676 ❤️
This is one of the best episodes this year! It is hilarious. I love you guys so much.
@42:45 Is what Ceri's describing those CGI videos I remember watching years ago of those squishy cubes "randomly" falling into place in a pile to form a specific image? They really begin with the cubes arranged in the finished image and reverse them to make it look like they fell that way by chance
This was a yet another great episode! Non sequitur: Recommendations for the best non-fiction pelican documentaries and books, please?
I love this show!!! ❤
Great show ! Thanks for making it! 💕
Ceri, making me feel seen!
i'm a grad student who pays their bills by grading undergraduate papers. I don't take off points if i suspect someone has used a chatbot (although the chatbot still needs to meet the brief of the assignment) because it's very hard to actually prove but after a while you start to notice patterns. The AI paper in the Truth or Fail very quickly registered to me as being generated. basically, AI is bad at writing academic papers but it is bad in an unusual way that is different from the ways scientists are usually bad at writing. scientists tend to write extremely information dense sentences with confusing grammatical structure whereas AI will write very flowery sentences that sound great but convey almost no actual information beyond vague marketing babble.
That being said. prompting the AI with more information can help so i'm sure i've read things co authored by chat gpt without realizing it. if anyone out there is thinking of using it for writing just remember the less information you give the bot the more it will fall back on empty garbage.
Haven't watched/listened yet but had to let you know when I saw that title I smushed Jordan and Harrod together to make Jarrod which was a frightening thought to have him on SciShow Tangents!
+
I felt so smart that my reasoning on the raccoon paper matched Ceri's exactly 😂😂😂
Ever heard corn grow? Plants can make sounds.
My craziest "my phone is listening to me" moment to date happened just 2 days ago. At 3PM, I was at physical therapy, where my PT recommended that I try ice therapy for an injured knee. Her actual recommendation was to use a styrofoam cup ("bad for the environment" she admitted as I frowned scornfully) but she described what to do for some time. Immediately after PT, I had to go attend a Christmas party. I came home around 9PM, realized that I needed to do some last minute Christmas shopping, and of all things for Amazon to recommend to me, "ice therapy cup" was on my homepage's recommended products.
HOW??? I was too busy at the party to even look into this, and I had never even heard of this type of therapy until just 6 hours earlier. I've not purchase any sort of PT equipment on Amazon for over 6 months, so it's not like this would be in my Recommends. It "just so happened" to show up 6 hours after my PT and I talked about it for a good 5 minutes.
So of course I took a screenshot and shared it on Facebook to watch my Boomer relatives panic over "duh gubbmint" listening in, haha!
"My mortal nemesis, Siri". I need to hear this story!
Because her name sounds so similar to Siri, anytime Ceri's name is said the Siri picks it up (and her name is often mispronounced as Siri instead of seh-ri)
I understand that reference!
I wonder, since we've gotten as far as we have with "hand made" A.I... Have we made the next step into "industrialization" of A.I.?
Like an A.I that specializes in data-set 'noise' minimization paired with an A.I. that's designed to build data-sets ....et cetera. Like an A.I. suite that's intended to other create end user specific A.I?
If you've ever tried to draw hands yourself, you'd realize that AI is actually only marginally worse
The dude just happened to have a Playboy magazine on his desk at work? Wild!
Amazing bonus butt fact!
When I'm half paying attention and I hear hank say "I want a hot chick and a sandwich" haha 😄
Did Ceri just say "Chocolate cow dairy farm" at 32:20?
I wanted to make a cute AI drawing of a cat with and eyepatch (cos a friend of mine her cat had to had an eye removed). But it reallyyy does not understand eyepatches. 🤣🤣
Did I fully lose consciousness or did Sam not have a fact? He always gets the best facts! I feel robbed of a third fact!
🦝👀 guys, we’re gonna get in trouble
Cool
what is the context of the fish 🤣
I thought the second paper or chatGPT was really obvious. It was just so very buzz word heavy.
you should have included the ai rat paper, it isnt offensive at all and is actually kinda funny and includes "words" that may not be actual words
I'd like a Hot Chick 'n Sandwich too, Hank!
Just think of the computational power it would require to constantly process all audio received from your phone/alexa/whatever, the level of profit necessary to justify that expense, and the margins on the consumer products theyre supposedly using this data to sell you. It makes 0 sense. Now a targeted attack against a specific individual, better yet targeted to specific times with a suite of relevant “wake” words. That sounds entirely possible if not likely to be happening already
Shall we play a game?
100% wrong about the mics not hearing and targetting. I was talking to a friend about my X being black, and the next time I went on Grindr, a whole bunch of profiles from Black Men showed up near me, there area I live in, that's not a possible reality. So the app populated what it think I wanted to see, based on what it heard earlier in the day.
President of the Raccoon Federation
Hank reading the raccoon optimization algorithm. Me trying to stay tuned in. A standard was set and this is not a glowing space racoon a genius may have had sex with or been abducted by so I am not interested.
poop id 🤣💩