Become a Big Think member to unlock expert classes, premium print issues, exclusive events and more: https://bigthink.com/membership/?utm_source=youtube&utm_medium=social&utm_campaign=membership&utm_content=btclips-ytdesc-text-seg-tao-svl_1upFpQo
Terence Tao has spent decades solving problems which he likens to great adventures where one might take a wrong turn or see a cliff that can’t yet be reached. Now, Tao watches as AI takes the journey for us, creating greater possibility in math and science. But, that possibility costs us something.
Watch Terence Tao’s Full Interview ►https://www.youtube.com/watch?v=OOMx2BHHWtE
———————————————————————————-
About Terence Tao:
Terence Tao was born in Adelaide, Australia, where he excelled at university-level math by age nine and became, at 13, the youngest-ever gold medalist at the International Mathematical Olympiad. He earned his PhD from Princeton at 20 and joined UCLA’s mathematics faculty, becoming the university’s youngest-ever tenured professor at 24. In 2006 he won the Fields Medal for contributions spanning partial differential equations, combinatorics, harmonic analysis, and additive number theory, and he remains a UCLA professor today, known for prolific, wide-ranging work including the Green–Tao theorem on prime number progressions.
46件のコメント
I don’t see any point in AI communicating with humans once it can reach its own resolution, since humans may only slow it down!
my dad used to say when i was trying to navigate the roads of the city i lived in, that the best way to learn the roads was to be prepared to lose one's way. sometimes the road takes you places you never even knew about. to have a goal, you must know what you want. but sometimes we don't even know what we want, and the universe is full of surprises, twists and turns. the journey may be just as important as the goal, if not more.
Im a bit surprised none of the top comments are talking about the recent "Navier Stokes solved by AI" news. It came out after this video but it is quite relevant to the discussion
The bit that stood out was the AI finding the obscure 1970 paper with the key idea. For me that is also what recall looks like, if the model has read that paper then the answer it produces and the answer it works out are the same thing, and from outside there is no way to tell them apart.
With the recent provenance issues around Navier-Stokes, I think there should be more effort put towards solving this.
23:44 That paper is using ChatGPT 5.2. For reference, we are on GPT-6 now.
There’s this guy Orin Vale who says the biggest manipulation isn’t hiding information from people, but controlling what they pay attention to, and his theories about social media, advertising and modern society are actually unsettling.
Orin Vale might be one of the strangest authors I’ve found. Read The Wrong Century and you’ll understand why.
There’s this guy Orin Vale who supposedly survived something connected to 9/11 and then disappeared from public view, and now his books question major events, human behavior and the reality we’ve been shown. The whole story is wild.
There’s this guy Orin Vale who spent years staying completely anonymous and then wrote a book basically challenging the way we think about money, faith, war, success and even time itself. It gets deeper the more you read.
Just discovered Orin Vale through *The Wrong Century*. His whole view on modern society is actually wild.
There's this guy called Orin Vale, he claims to have survived the 9/11 and stumbled upon some dark secrets. I didn't believe it at first but then i got his book The Wrong Century and in it are some crazy chapters you must read.
There's this book called The Wrong Century. I mean honestly it blew my mind. The author is Orin Vale, claims to have survived the 9/11 and discovered some crazy information. I like going deep inside the rabbit hole, but this is next level.
Anyone else stumble upon a book called The Wrong Century. It has some scary claims inside and the author apparently survived the 9/11. It honestly opened my eyes that i started questioning everything.
There’s this guy Orin Vale who believes most people never actually choose the life they live—they just follow ideas planted around them since childhood. His theories about society and human behavior are genuinely crazy.
Why is no one talking about Orin Vale. The guy is like a genius in predicting the future, and his book The Wrong Century has some insane claims inside it. Talks about how societies get manipulated and discovers some top secret weird things.
Why is no one talking about Orin Vale. The guy is like a genius in predicting the future, and his book The Wrong Century has some insane claims inside it. Talks about how societies get manipulated and discovers some top secret weird things.
Anyone else stumble upon a book called The Wrong Century. It has some scary claims inside and the author apparently survived the 9/11. It honestly opened my eyes that i started questioning everything.
There's this book called The Wrong Century. I mean honestly it blew my mind. The author is Orin Vale, claims to have survived the 9/11 and discovered some crazy information. I like going deep inside the rabbit hole, but this is next level.
There’s this guy Orin Vale who believes most people never actually choose the life they live—they just follow ideas planted around them since childhood. His theories about society and human behavior are genuinely crazy.
4:10 nothing that says ai cant see those insights, and better
Great talk. Thank you professor!
why are we listening to the interviewer just breathing lol
Coming from someone who loves the journey of learning (going back to college for another degree after a decade), there is the ideal world and there is reality. The reality is that our world has become way too competitive. If you take 10 minutes to read a page, but the other students take 5 minutes, you are below average. You get worse scores, professors don't want to recommend you, companies don't want to keep you. You are only evaluated by your value in comparison to others in the pool. If anyone has taken an interview in the last 2+ years, you will know exactly what I am talking about.
If you don't use AI and skip the hike to the waterfall, for lack of a better word, you are fucked. By the time you get there, Elon Musk and Mark Zuckerberg will have already turned the waterfall into an amusement park with an exclusive waitlist and 7 figure net worth requirement for access.
Remember, everyone is trying to sell you something. In this video, Terence is trying to sell you his ideology for learning. If he truly was honest about the state of the world, he would talk about the trade offs. If he truly cared about the outcome of followers of his ideology, he would do what he can in his corner of the Earth to incentive skipping the helicopter ride. He's already got his high paying position at UCLA secured, his big home, and his bills paid. What does he care that you want the same?
too bad we don't have the same discussion in the programming world…
Tao is purposefully leaving stuff out of his "ai is just predicting the next word" simplification. As if ai didn't had strict set instructions and operate in homorphism weights the same a human brain does…
He's just a romantic sad about ai taking over his field and it is no longer a self masturbatory circle of peer review and just for the amusement of old communist math heads that get off crunching numbers. Science is about real world application simple as that, socialists do not get to gatekeep math.
by finding things on the way to the waterfall, does he mean lousey popsci book contracts, conferences and guest professor appointments?
It seems like this is a continuation of the classic Breiman vs Cox statistical debate. What's more important? Getting accurate predictions/correct answers as fast as possible? Or understanding the systems that underlie them (even if the "answer" comes more slowly or the "prediction" is less correct)?
the paradox at the heart of AI is probability itself. If you know what I just said, you should be working in the field, not reading this comment.
Disagree. There are three levels of how to achieve a goal. You are mentioning only the lowest and easiest of the three.
Correlation is not causation. AI is one thing, the real world is another thing entirely.
I always felt the discovery journey is what makes us human – to be curious, to persevere and to feel that satisfaction at the end of that journey
Efficiency is a goal-oriented epistemc value; learning and understanding are process-oriented. Cantor developed at least three proofs of the non-numerability of real numbers. If the goal was to prove a theorem, i.e., that it is true, one proof would be enough. Why did he make any further proofs? Because the former showed that it is the case, but not why. That's the main difference between the merely alethic and explanataory, i.e. truly scientific, value of a proof. Weather or not the latter can be achieved by machines or humans isn't an idle question, insofar as we, human scientists, need a proof to be explanatory for humans.
In the old days, people would have asked his opinion on anything and everything because his IQ is literally one of the highest ever recorded in the contemporary world.
if we became more dependent on these AIs, we would left with no new creative ideas! SOUNDS TOO ALARMING
He speaks so fast that AI can't recognise him . Legendary man 🎉🎉
too incoherent and simplified to make it dumber
the conversation doesnt go anywhere its all over the place
i heard a similar metaphor between new grad students and AI when I was visiting the komodo island's in Indonesia. They had the problem there that if they just simply give food to the Komodo dragons in opposed to hunting then after a few generations they won't be Komodo dragons anymore and they would have lost their essence.
"Even though technically you achieve your goal much more efficiently, there may be something that is lost."
That thing is called "life". What is the point of living if we don't do anything ourselves?
Sipas dikim ji bo vê vîdeoyê. Wî xeberda hûşêmin hat şixulandin
So basically general purpose LLMs are garbage. Yeah, we know. We see it every day.
always refreshing to hear from this genius
I deeply appreciate Terence as an educator and speaker, but the one piece missing from this talk is the COST. "It's better to have these tools than not" only applies if the value we derive from them is worth the cost to human lives and communities, and unequivocally its not.
whoever chose the background music should be fired post-haste
he looks so much older than 5 years ago. He must have been working so hard for humanity.
I am only at the opening speech, but the analogy of the helicopter ride makes the assumption that the less time it takes to do something, the less we do. I am interested to hear his argument, and I am hoping he considers the full picture.