Winning MONEY with AI - Understanding Algorithms & Predicting March Madness with Neil deGrasse Tyson
Vložit
- čas přidán 19. 06. 2024
- Can artificial intelligence predict the future? On this episode, Neil deGrasse Tyson and co-hosts Chuck Nice and Gary O’Reilly explore algorithms, computing, and how to win Warren Buffett’s March Madness money and beyond with AI expert, Matt Ginsberg.
How do you create the perfect bracket? We discuss how to use algorithms and game theory to build brackets. What are your odds of getting it right? We’ve all heard of “the algorithm” but Matt helps us break down what algorithms are. How do they work? What can they do? Can they be predictive?
Could you write a code that can learn to write code itself? We discuss the odds of getting in trouble in Vegas and what you have to do to be good at sports betting. What sort of data is required to enter into your algorithm? You’ll also learn about the difference between data and knowledge. Can you code for invisible forces?
Is the human brain just a really complex algorithm? Dive into philosophies about intelligence and consciousness. Is there stuff happening in our head that is not computable? We explore Roger Penrose and the theory that the brain and consciousness might have a quantum mechanical element. Why did we give the Mars rovers a “serendipity mode”? Should big tech be so focused on data collection? Are we pointing enough AI at the world’s problems? All that, plus, we discover why using AI for exercises like tournament brackets and chess games can prove to be worthwhile.
Thanks to our Patrons Robert Bork, Nick Fugal, James Trager, Brian S, Nightfall, Chris Hernandez, Mithat Sezgin, Luke Fertal, Rhett Hogan, and Patrick Creech for supporting us this week.
NOTE: StarTalk+ Patrons can watch or listen to this entire episode commercial-free.
Get the NEW Cosmic Queries book (5/5 ⭐s on Amazon!): amzn.to/3dYIEQF
Support us on Patreon: / startalkradio
FOLLOW or SUBSCRIBE to StarTalk:
Twitter: / startalkradio
Facebook: / startalk
Instagram: / startalk
About StarTalk:
Science meets pop culture on StarTalk! Astrophysicist & Hayden Planetarium director Neil deGrasse Tyson, his comic co-hosts, guest celebrities & scientists discuss astronomy, physics, and everything else about life in the universe. Keep Looking Up!
#StarTalk #NeildeGrasseTyson
0:00 - Introduction
5:43 - Using AI To Build Brackets
7:49 - What Are Brackets?
10:48 - The Odds Of A Perfect Bracket
11:26 - How To Do 'Really Well' On Your Bracket?
13:21 - What Is An Algorithm?
20:59 - Can AI Help You Win In Vegas?
25:20 - What Parameters Would A Predictive AI Use?
29:58 - Where To Find Matt Ginsberg Online
30:29 - The Implicit Data Of A Victory
34:08 - The Difference Between Humans And AI
39:36 - Should We Use AI To Find More Data?
43:44 - Is Matt Entering Into March Madness?
45:01 - Why Matt Ginsberg Isn't Concerned With Money
46:32 - How Can AI Solve Our Problems?
49:13 - Is AI Improving Our Understanding?
50:33 - Can AI Predict Our Doomsday?
51:15 - Closing Notes - Věda a technologie
Who's your winning team this year?
Lazarus
The Steelers
Arsenal FC
daaaa bears
I really appreciate this channel and all videos you have. Believe or not I know it's not your purpose but you have helped me with my depression problem since I find extremely fascinating knowing about the universe and all the rest of your content that it makes me notice how silly it is to worry about my daily problems... thank you guys.
Thank you fort this heartfelt comment-it reaffirms what we do here. We're so happy to hear we're helping! :)
Such a nice conversation in these troubled times
Thank you so much! It's just so great to read things like this. You've cheered me up, too! :)
I'm experiencing depression right now and have had my anti-depressants increased. Startalk helps me tremendously also. I now have a purpose in discovering more about the universe and our place in it! Thanks so much!
Yes!!!! I feel the same way. Startalk and Neil Degrasse Tyson have opened my eyes to so many beautiful things. I am 46 years old. I grew up in a fundamentalist Christian family. I discarded that framework intellectually and knew it was all BS when I got into my 20s. But it isn’t until recently I realised the narrow religious ideology my dad taught me was the foundation to all my beliefs. And those narrow beliefs kept me from appreciating and being awe-inspired at our world! I was misdiagnosed with “depression” for years; it wasn’t depression it was being disconnected from my humanity and the universe.
This was one of my favorite episodes. love you guys
So glad to hear it!
I used to watch star talk all the time as a young adult, and now I’m coming back to it as a full fledged 28 year old parent.
I think this podcast, among others, has basically helped me to outgrow “pop culture” science. beginner level learning. But I don’t know much about sports, so I’m learning a lot again!
I just want to thank the StarTalk team and Neal and Chuck especially, for helping me to develop a love of learning and a sense of humor.
I always used to say, Neil degrasse Tyson is my favorite black astrophysicist, but more broadly he’s totally my favorite science communicator!
Can't wait to learn something new today and have a laugh of course!
Hope you did!
Really enjoyed this episode. I liked Matt who kept a stern and calm voice when they poked fun at his background image and asked why he was not rich, whether they would break your knees in Vegas, how he is sticking with classical problems, how poker is different from a slot machine, why he thinks human wetware behaves differently than machine brains, how the future is human+machine, and his confidence that data can predict outliers in basketball. I also liked that he said people will think a better algorithm is "fair" but having extra data is unfair. I also liked that he did not dismiss Penroses Quantum consciousness hypothesis. Great show! Made me think very hard. I liked his insight that AI has to win really big in sports to be noticed and marginal improvements will fly under the radar. Great work Matt. I think with Penrose winning a Nobel, Quantum consciousness will be an accepted brain theory and restore humanities free will from determinism. I think I am a compatibilist nowadays. I also saw a demo of AI writing code. Good one. It is called GitHub CoPilot. And hey Matt is a family man to boot. And Neil accepted serendipity. Cool.
I wrote a code once that automates writing code with my voice. I created a language and had to gerry rig a bunch of stuff but it worked and was fun (and a lot faster than typing code by the fastest typist). 😂
Oh! And I had to train my Google keyboard to recognize certain commands and word patterns since i didn't have money to access their vouce systems.
10:56 I believe this is known as the gambler's fallacy. There are individual properties and defects within what is being "flipped" that will cause it to fall more disproportionately on a certain "side", however.
Some dice for example may have higher probability of landing on certain numbers due to factory imperfections throughout the manufacturing process. Dice made in a plastic mold may have regions of higher density.
That's insane man.. I would love to hear more about it!!
isn't this just voice recognition, its quite different from prompting an AI and let it writes its own code
That's really cool! Bravo 👏
Did you use Dragon Voice? Did you use google voice? How about neural network learning models like OpenAI? Did you get the beta version of speech-to-code? How accurate is the speech recognition and what techniques did you use for noise reduction to coherency to correct code writing? Did you need a lot of shortcuts from voice to snippet of code? Did the code output work in a sense of "Create a search bar, create a white background, create a logo name, give colors xyz" And it produced Google webpage?
Great guest again. Lots of confidently presented info and, your editor gets kudos. Glad you stepped up from that Caesar's bunker you used to be in Neil.
Maybe the best Star Talk ever. Everyone involved firing on all cylinders! So nice to feel smarter than I was an hour ago.
...👍
Always excited to see you Neil‼️😌
Thanks for being my personal astrophysicist! Nice team contribution today
Hi. I took w course in Machine learning by professor N.G fron Coursera.
NG is the OG, what was your predictable outcome of your grade?
Congratulations
wow thank you to everyone on the meeting I learned so much, very good insight and levity from all :)
Been a programmer used to work with AI and Machine Learning this episode it's just awesome.
This episode was amazing. Good questions and great answers. After this, AI is not nearly as scary as people make it out to be. Thank you all an have a great weekend.
I mean I guess if you purposely refuse to not see 2 steps ahead or critical think the issue now.
AI is as scary as people make it out to be.
If Elon Musk says it's something that we should worry about A.I as much as nuclear war, then i'm worried about A.I as much as i'm worried about nuclear war.
@@brunomartinello1114 ;
@@brunomartinello1114 , c a , , , , , , , ,a. A. , , , a ,,
Don’t know when this was recorded, Chuck’s final remark however shows remarkable insight, the response was pretty scary.
They have a doomsday clock. Should hook up these AI to them. Autonomous Weapons with AI are round the corner as per experts.
Great topic! Excellent guest! Would love to see him again!
Watching Chuck on Brain Games! Way to go, Chuck!
2^63 or 1 in 9.2 quintillion. I googled it. A perfect bracket. I like those odds. No problem.
The bracket must be a US term. I have never heard anyone use that in Australia.
To complement dr Ginsberg's answer on what algorithms are, one of the preconditions of an algorithm is that any and all recipients, under ideal conditions, must understand and be able to execute the algorithm properly, without external help. In that sense, the code that enables predictive AI is an algorithm, but so is a set of instructions on a public phone or a bus stop (i.e. they are a set of simple steps, that anyone who speaks the language they are written in should be able to follow in order to use the service)
i want to hear more from your guest Matt he sparks a lot of thinking.
A dynamically changing set of interacting formulas that keep changing according to each variable within each formula or set of formulas
Thank you for carving this path, Neil. I am right behind you!
Another flawless educational episode as anticipated. #GaryO #LordNice #Dr.Tyson #Dr.GINSBERGROCKS
Brooklyn Claims Dr. Tyson & Lord Nice. Brooklyn NY loves 💘 StarTalk Sports addition ❤ #NCAA
Dr. Tyson! Mississippi thanks you for the shout out ❤ I'm surprised you even knew about Oxford, MS 😂
Algorithm = The evaluation steps taken to reach a relevant decision which will contribute to the correct answer to a specific question. You can start with a set of data, then determine every possible parameter about each of the elements in that data set. Then for a given parameter, measure or evaluate it for each element. If you can establish that the value of this parameter actually contributes (in statistical terms, 'correlates'), to reaching a valid answer to the question, give it a weight and add it to the voting of all the algorithms to reach the correct answer.
Thank you Neil!!! I didn't know what brackets were, because I'm not a native speaker
Thanks. AI is a powerful tool today 😁
I don't understand how Neil can't understand that there are people who don't get into sports at all who still love science and yes we watch this program
True. I only play Chess and Tangrams (Polygrams) nowadays.
Try watching sports as a documentary on human behavior in a game setting.
I watch Star Talk all the time and have been for years. This has been one of the best episodes I’ve seen. Thank you for talking about this
About professional coders nowadays writing code that writes code, I was doing that with my CAD app since the late 1990's. Back then my app didn't export dimensions well. I was well versed in my app's code and I learned enough code of the other app to write code in my app's language that created a file of code written in the other app's language that recreated the dimensions in the other app's native format. Worked like a charm.
Chuck if you do nothing why do I keep coming back!? You put a smile on my face, and so often I root for you on this show when you stand up and call out the evils in the world that so many cannot see or do not understand. I love the whole startalk crew, thanks to all but a special thanks to you for doing what you do, Lord Chuck Badass. Keep preachin it brother.
And even though I'm approaching my 80th decade I had NO idea what a bracket was. Which is why I keep coming back. I learn suff.
You're almost 800?! In all seriousness, we're glad to hear it because you're never too old to stop learning!
I had never heard the term either. I don’t live in the US, so I’m assuming this might be an American term.
In the very simplest of terms, It is a risk calculation that takes into account as many variables as possible.
Humans can do it innately.
PlayStation Network used cloud computing for the Folding at home AI during the PS 3 days.
I love a guest that happily rolls with the lads
He even burned Chuck early on
S Tier
this was a great stark talk episode
8:00 what is a bracket? Thanks
@Gary Let’s start a campaign to rename American “Football” as Handball, and call Soccer, Football in America.
Great episode
Thanks!!
"You always know that you don't have enough information."
Love it!
Great discussion - well done. Next up Sir Penrose!
Wow
"Probably" oh geez thanks for another banger of an episode yall. I am optimistic for the rising tide of sustainable tech and environmental/social benefits that it will bring.
41:13 this was deep. heavy on the mind.
Chuck Nice....Props Bro!
That problem with the raw data, is that you have infer what relationship or ratios are the important one, e.g. passing yard vs wins. You want to find that ratio that no one else see. That’s when you have a winning algorithm in sports betting.
Good Show, Guys!
Best startalk I've seen
In sports, one thing that would make for better data is a verifiable way to quantify athlete’s health, well being, and other attributes that may have an objective effect on outcome.
I’m surprised none of the podcast regulars didn’t tie this topic back to their multiple episodes on data-driven player/team development. However, that kind of data would be the unethical insider-training data that Matt said he wanted to avoid, preferring developing a better algorithm to consume the public data.
Google obviously understood the concept of “fiction” since they hired Matt despite how they’re portrayed in his excellent novel _Factor Man_.
30+ years ago, before AI, before Machine Learning, there was a class of software named Expert Systems. I arranged for my employer’s ES software to be evaluated for airport automation. In the 2 million miles I flew after that, I don’t think it was up to the task (or more likely, the computational resources cost too much). American Airlines still made us wait on the tarmac for “our” arrival gate, when we could see multiple empty gates. The ripple effect of a single gate change was costly.
Wow... I was actually the first person to start watching this video! I clicked on it less than one minute since posting and it listed no views. Now, regarding Warren Buffett holding on to his money and not giving it away... WRONG! Buffett has already donated Billions to charity greatly decreasing his wealthiest people listing. He has also arranged to give all of his money to charity when he died. His family will receive little if anything.
What an amazing episode👍 It reminded me so much of the book 📚 'Phenomenon' (Amazon) by Neil Fulcher it takes the reader on a real journey involving many of the concepts in this episode 😄 Thanks chaps 🌈🎱
Who programs the card shufflers?
Put simply, an algorithm is a step by step guide.
If you had to teach a robot how to walk, you'd be very detailed:
1. Pick up your right foot
2. Move it forward
3. Put your right foot down.
4. Shift your weight onto your right foot
5. Pick up your left foot.
6. Bring it forward
7. Put your left foot down
8. Shift your weight onto your left foot
You'd loop this until the robot needs to stop or change directions. This is a literal step by step explanation of an algorithm
what it if finds a step going down a ladder, or up a ladder, or a boulder? or a bear chumps on the robot's foot? you need more steps, not so SIMPLE eh?
@@Veronique487 of course. Just simplify each step. That's the beauty of programming. How does this seemingly complex thing get broken down?
25:03
Me when there’s a new StarTalk episode.
Made me chuckle
That’s the chessboard double every square in rice odds. Makes lotto odds sweet.
I like pennies instead of rice.$$$
You have a great point. Im gonna remember that next time I explain the odds of something to someone. Thanks.
off topic. in VACUUM, CAN h2o be liquid or solid
can h2o be a significant part of a comet's mass
in near Vacuum, is it POSSIBLE to fly, with rotors.
Dr Tyson! lord chuck! Clutch$?!@ in the data?IS what we call morale!!!!!!
Blue shirt kid should of been a guest of this video
The algorithms shown me life would be more difficult from there on out: now, that is specifically the case. Specifically relating to the winning average of life substining average itself.
I'm with Emily. I like the team with the bright colors too. lol
Colors may be relevant. I hear that in soccer teams playing in red shirts are more likely to win. Well, that's what the data says anyway 🙂. Now it's a question about correlation vs causation, and whether the difference is relevant.
Thanks!
Thanks for the SuperThanks!
Terminator’s red/white code overlay in the movie was Apple kernel code.
14:35 “please clap” micro moment.
I’d love to hear another one of these using the lottery instead of sports betting
I take this view every day of my life.
18:10 another way your social media data can be easily plugged in to an algorithm… if there’s not already a data cloud following NCAA players personal social media to predict odds, (A)I would be surprised .
Last week on Star Talk: the world is literally ending. This week, how to win the lottery!
Variety is the spice of life!
@@StarTalk If I won the lottery, Id attempt to save the planet hahaha
@@sketchtheparadigmyork1217 You'll need to win a few lotteries to be able to TRY that, unfortunelly.
At work had a football pool. 7-8 games being played that weekend. I know nothing about teams. I went down the list picked win/lose/win/lose etc.. I won out of pure luck. $70.00👍
Best I can do is make pixels move about, track ints, make noise and blink in and out of existence.
And Chuck of course(:
comercial breaks is on point. haha ad pops up right after you say it haha
Chuck should do more dr Phil. That was hilarious
A.I . seems to have changed what a algorithm is capable of, and even it's definition. I thought it was just a series of mostly repetitive instructions (many times a sub routine, "accessed" from the main computer program) to perform a task, like sorting/stacking data (not a "physical" sort/stack, but a place in the computer's registers , which the computer keeps track of). The algorithm concept may predate computers, as we think of computers (of today.) I get the feeling my "dinosaur tech
knowledge" is way outdated.
I have idea what brackets or in sports or March madness.
This clearly illustrates who has IA or controls AI are the ones that benefit over the person down the street that does not. The problem is how do we prevent everyone from becoming the person down the street vs the few with control of AI.
20:05 yes , they do, it’s called the spread or line.
Last! Nailed it!!!
The thing you need to do is pick the upsets which is still very hard, and the human factor always gets you, then there’s just luck so yeah, not easy
This REALLY feels like the wrong first question for Dr. Tyson but, coming from a Norwegian;
What is a bracket?
20:19 -- hahah Gary's face
Algorithms are like baking the way programs are like baking. Programs are a recipe, a list of ingredients and directions. Algorithms are like baking technique, like group all of the wet and dry together and add wet to dry (the muffin method); and running a blade across the top of each measuring cup for more accuracy. You can't use a muffin recipe for making cupcakes, but you could use the muffin algorithms for cupcakes.
If I'm completely out to lunch on this analogy please correct me. I'll learn something and we won't be putting incorrect info out there.
Algorithms can accept dynamic inputs (data changes). Don't try that with baking! Past that, I would agree that algorithms behave like a recipe.
The idea that the algorithm is a regimented process that operates based on data that is allowed to change, makes it fundamentally different and more valuable than a recipe.
Great episode lmao
Madden is a example of this
22:00 chuck is right. they will break your legs hahahahahahah
Pittsburgh does have super unpredictable weather. Especially in the past few years, I just don’t trust the weather reports more than a day out.
It’s dem valleys
a lot of legs getting broken in this ep. should name it A.I. Mafia hahaha
Why don't we have a analog version of tensorflow it wouldn't be that hard to do tensorflow is kinda simple idea but in digital the computer needs to calculate the same thing multiple times in different areas of the frame before it can respond but with a analog computer it can all be done in the amount of time light takes to go from one side to the other side the only problem is that computers are cheaper than making a chip from scratch but it can be done way faster it's currently too slow for like a space rocket but a analog tensorflow can function faster than humans can so it can pilot a space craft faster than a human so maybe we need to so we can go faster in space and not hit anything like a obstacle avoidance system that responds at the speed of light it would already be turning time the human can register what we sall
It's a rhythm by Al Gore. An Al Gore Rhythm
Is there an escape velocity for the moon? With photonic and gravitational anomalies and probabilities in the next 30 years?
23:54 a long-winded way of saying that casinos rig results so you lose after your bet is placed. similar to the stock market. robinhood always sells whatever you buy.
A computer is a clock with benefits
Damn Neil looking jacked up
Equations equate, algorithms algorate.
Classical or quantum?
We need an algorithm that predicts how much more natural resources russia will have if we don’t do anything about climate change. (Because of melting ice that would otherwise make locations to remote to drill or mine)
I WISH THERE WAS A SITE OR PORTAL THAT PROFESSORS LIKE THESE GUYS CAN EDUCATE STUDENTS ALL OVER THE WORLD WHOM DONT HAVE THE FINANCIAL ABILITY TO BUT HAVE THE CURIOUSITY TO LEARN TO BENEFIT THEIR FAMILY OR COUNTRY.
Why are you shouting at us? #allcaps
Can AI produce code that is more complex than the same AI's code itself?
Why do they keep interrupting?
My only concern with this collecting data is...what is going to happen in the future? What of these hackers collecting so many citizen's (around the world) data? Because I have a bad feeling about this.
WOOO