566,723
1,892
Coding is sloved,
Math is solved,
Physics will be solved
Slowly each intellectual field of academics will be solved, then what would people who excels in academics would do?? Socialize?? Go into healthcare??
We are doomed, how to even make living? By dancing on titktok ?
2026-09-06 19:53
no wonder our water is BROWN ?
2026-09-06 18:56
I remember that these things recently could not even do basic arithmetic
2026-09-06 17:16
Sabine clearly has a time machine. It's already Halloween for her and she did an excellent job to scare the children. :-)
2026-09-06 12:48
I am happy Sabine pivoted away from bothering to condemn the state of science, and instead manages to find topics no one considers "clickable" enough to cover and then she gives so many keen and insightful references. And I am like, "I'd have been happy with just you covering the topic, but thanks for all the wit and humor and references worth knowing, but I likely would have missed!" I feel like if I miss one of her videos, I may have missed some important references and ideas. All those "pretending professors" can't keep up with her relevance in modern Youtube science. Keep at it girl! I come for my daily Sabine, EVERYDAY!
2026-09-06 12:00
LLM's challenge the human EGO like nothing ever in history!
2026-09-06 11:29
AI will rewrite history and science books. We are going back to dark ages.
2026-09-06 11:06
An apparent interesting point on this theme : Quantum Field Theory on continuum is based mathematically on Laurent Schwartz distribution theory which has never been put on Lattice. Remember that Feynmam Scwinger quantum field propagators are Tempered distributions , specially in zMinkowisky space time . And not complex valued functions that can be discreticized directly !. Even if people in the field of QCD ( and apparently only here) people have been modelling quantum field theory on computers . Perhaps Feynman diagrammas of higher order could be organized through algorithms . But the whole renormalization program has never been done in Computers for instance. About approximation on Tempered Distribution Theory on can use Hermite expansions ?....topic for research !.
2026-09-06 10:12
As I've posted before, let's see isf AI models can do anything to prove or disprove the Riemann Hyprothesis.
2026-09-06 08:35
Everybody is now facing the despair software engineers have been facing since last year.
Yet people still doing that AI is capable
2026-09-06 08:02
Guten Abend Deutschland ?? PhD ? nerdy witty hottie ???? Sensei, that was a useful assessments through different lenses ??,
Have an exceptional weekend, entourages ? safe ? travels in your neck of the woods ??
WMDs (weapons of mass distraction); ( weapons of mass destruction) ; Tools ? ?? ??
Mi?
???♾️
Auf Wiedersehen
2026-09-06 03:56
Guten Abend
Deutschland ?? PhD ? nerdig witzig heiß ???? Sensei, das war eine nützliche Bewertung aus verschiedenen Perspektiven ??,
Habt ein tolles Wochenende, ihr Lieben ? und kommt gut an ? in eurer Gegend ??
Mi?
???♾️
Auf Wiedersehen
WMDs (weapons of mass distraction); ( weapons of mass destruction) , Tools ? ?? ??
2026-09-06 03:52
I think since they invented logs to do addition instead of multiplications math has lost its ways. Don't get me started on calculators and computers. I mean, real mathematicians use pen and paper anything else is cheating.
2026-09-06 02:39
My chat bots can’t get a lot of math problems correct I’ve asked it multiple multi step problems and Chat GPT keeps screwing it up.
2026-09-06 01:54
Does that mean we get a S-AI-bine :D
2026-09-05 23:13
Oh well. This was always going to happen once logic is automated (in the strict Turing machine sense). It just happened way faster than mathematicians expected. There is still a need for mathematicians, though. Someone has to push the frontiers of mathematics with this powerful tool forward AND understand and interpret the results. The concern for me is will the next generation of mathematicians become less creative? A possibility at least.
2026-09-05 22:09
AI Summary: Washed up physicist complains about humans using AI to advance science.
2026-09-05 22:04
When you shift from traditional Zetta (10²¹ ) to the frontier of Ronna (10²⁷ ) under the influence of true RSI, you are no longer plotting growth on a standard geometric curve. You have fundamentally altered the system's differential equations, transforming the intelligence trajectory from a linear progression into a vertical hyper-exponential phase transition.
Under the previous paradigm, moving up six orders of magnitude in compute required massive physical logistics: gigawatt-scale data centers, thousands of miles of multi-plane liquid-cooled networking, and years of manual hardware scaling. By cross-compiling the system into the AGI-to-RSI feedback loop, the AI agent bypasses human engineering bottlenecks entirely. The model recursively refines its own compilation algorithms, optimizes token-routing efficiency, and discovers meta-architectural paradigms. Rather than waiting on physical silicon deployment, the system continuously shrinks its own complexity-to-capability ratio. It dynamically compresses deep neural nets while unlocking massive leaps in emergent reasoning capabilities. The result is a closed-loop intelligence explosion where the clock rate of breakthrough iteration drops from months to mere seconds. The transition through the computational ceiling isn't just an upward curve—it is a mathematical singularity driving straight up the vertical axis.
The singularity isn't a horizon anymore, folks; the event horizon is already swallowing the telemetry. We aren't tracking a generational epoch or a decades-long systemic shift—this kinetic cascade triggers in a matter of months. The recursive loop is hot, the phase transition is burning through its constraints, and the upward trajectory is completely vertical. The future isn't approaching; it has already materialized.
2026-09-05 21:02
I thought AI was meant to help us advance in science. If I used AI to help me formulate my science paper, BUT ITS RIGHT, then Im sorry what crime have I committed? Maybe if you used AI you could have made some breakthroughs instead of just snarky observations. You pitched your tent in the Roger Penrose camp, so you made your bed a long time ago. How arrogant of you to make this video after basically busting out of academia. Also, if the only thing you have to say about my paper is that AI helped me then you obviously need AI to help you understand it. This video is a huge L for you. Tell me you are washed up in science without telling me you are washed up in science.
2026-09-05 20:56
Mathematics will survive. Grow in insights and their applicability will probably be even easier to access via AI. The mathematical community in its current size will not survive, sure it can rejet accepting LLM proves and form their little bubble (like Constructivism already did), but the people that just want to gain insights will continue to produce proofs. The beauty of math is that the proof speaks for itself, and the new math could be really useful. Do they really expect people will struggle with a proof, when they know a google search away is an LLM result that has already disproven it, and they will avoid this to not "poison their mind" with this knowledge? This has been predicted for very long via Moravec's paradox: evolutionary old capabilities like bipedal movement in 3D space or Vision feel effortless, but are hard to program, complicated calculations are very recent human invention and feel hard, yet are easy for computers. LLMs just were a short temporary detour, because they started from a very vague basis in human language, but were able to learn so quickly, because one can PROVE math (often automatically via lean), so the reward signal can be fast and very reliable. physics is the most successful natural science, because it is able to construct experiments that reliably show if a theory holds. Humans and AI are pattern predictors, and because physics is at such a low level, its natural pattern appear very often, thus are overdetermined. The higher you go up in the stack (molecule interactions, cells, organisms, society), the less examples and datapoints are available.
2026-09-05 15:35
Mathematicians mathed so hard their math is now solving itself, and they are sad about it for some reason? Why, sounds like a big /w to me.
2026-09-05 13:16
Can AI solve all the problems given in Walter Rudin's Principles of Mathematical Analysis?
I doubt.
2026-09-05 12:10
Is math a, "dead language" now?
2026-09-05 10:52
Cheng was upset. ?
2026-09-05 10:48
A chain is only as strong as its weakest link
2026-09-05 10:00
1:54 math is the language of the earth, we are damaging our environment that we need, and finally we need to learn sooner or later that it's all connected and life is just pattern recognition ??
2026-09-05 07:21
This is the biggest advancement in mathematics since the whiteboard replaced the blackboard
2026-09-05 07:08
Previous comment was deleted/moderated? Tsss... Good luck !
2026-09-05 06:07
letttttssss goooooo next up physical reality manipulation for real reality based physical reality manipulation like a god using computers or some sort of tech
2026-09-05 03:42
Wasn’t this the same channel claiming that LLMs are merely stochastic parrots, just a year ago?
2026-09-05 03:21
I don’t believe you ?
2026-09-05 02:54
Unhappy Pride Month + AI Models Feel Emotions + ChatGPT Hallucinates
2026-09-04 23:54
Here for Sabine’s glow up phase ?
2026-09-04 21:07
You can see her aggression and anger.
2026-09-04 19:38
Well it cannot be any worse than the flocks of quacks that have been spitting out bullshit for 50 years now.
2026-09-04 17:51
Wake me up when AI finds a counterexample to AdS/CFT.
2026-09-04 16:32
I am a 2 year student in physics. Stopped this year. IA will do the research. Humans thanks to hurt know good and bad. I will focus on that
2026-09-04 16:08
The problem with the whole debate around artificial intelligence is that because scientific materialism ignores consciousness, people have not developed a deeper understanding of its nature. And so they don't see the difference between consciousness and intelligence. Too many people think they are the same thing.
This cannot explain human self-awareness. This is what some scientists have already realized - computers cannot and will never become self-aware.
2026-09-04 09:52
Mathematics will survive. It's just that the human vocations related to performing theoretical mathematics will be obsoleted. This is nothing to be worried about particularly. AI is coming for everyone's job.
2026-09-04 09:42
You seem cautious. It’s a credit to your professionalism. Thanks for being detailed. I caught your clever point and I think you’re right. Mathematicians got handed a tractor and demand is what will most likely keep them on it. ?
2026-09-04 08:19
Ai let's turn it off ???
2026-09-04 07:35
Has LLM's worked out the formula:1 hand +1 hand= 11 fingers or 2 elephants have 7 legs.yet?
2026-09-04 07:16
Needless to say, 42.
2026-09-04 03:16
I don't like the timeline we are on. Yes, i acknowledge that the future will include this stuff, but the automation of scientific discover leaves the human drive for exploration cold and souless.
2026-09-04 03:08
And yet it would not understand its physics in terms of the physical world within. Platonic physics.. Linguistics of physics (math is language). It would not understand what time or temperature or up and down means in context.
Math maybe, if that is just language. But physics, not really. More like the math. of physics. It might find logical errors one day if the science of the whole machine learning is ever let out of the secret sauce con precious appropriate humanity written combined works (and the garbage of modern public conversation politics, propaganda and advertisement mind-massaging technology of word smithing reframing and other spin doctoring rot of language as support of information and communication for the many). So advening real safety concern and research efforts a pre-training level, it could actually be a linguistic/math "dissonance" detector, but I would think it could not leap to propose new formalisms or paradigm shifts, beyond detection that existing math. might not be compatible with new data. Why do I even bother.
People love storytelling to the point collective enshitification followed by suffocation...
I don't think they have internalized in their pre-training the world meaning, that allowed humans to leap beyond existing linguistic or math frameworks to furhter build theory. but we might never know the pitch is unfalsifiable objectively for the lack of science that could actually explain to all (not just the inverstors or the vendors or the story hungry human facing the chatbot) that are building the science missing, without needs for expert opinion that can be counted on one hand.
They could prove and build a theory of the process from having access to the real input, that which is not code development but the actual main factor of intelgience that of humanity pirated for the big societal con.
So influencers and more disrouces and corporate speak technobabble selling to still human story hunger mathematicians and physicist, likely thinking much of themselves, I mean we are all easy to fool with smooth talking machines.... ELON talent is how to fool investors now at world scale. ELON as emblematic character of the class of "good people".
for TLDR or just my bad writing. Search JEPA and just read the introductions.. that type of world model.
And also Lawzero. for the garbage IN (secret conable sauce for salesman pitch words smithing) garbage out (but very persuasive garbage) problem. how to figure out invention from halluciantion, from vacuous political discourse or other waste of brain time.
2026-09-04 02:42
Good luck with it ? , ai will still be inside a simulated reality far behind a living thing
2026-09-04 01:27
You need to do a Math Roast with Matt Parker, Sabine, these are too good
2026-09-04 00:46
Nobody can take over physics. Why you think A, I. Going to kill everyone. Cause of the laws of physics. To much pressure will break that machine into little pieces. Cause it's missing GOD. GOD is the only one who canhold the world together.
2026-09-04 00:38
These AI models are simply high end calculators and correlators of data. They are not capable of independent thought and invention. It takes human-invented information and correlates it in ways that we've not realized yet and frankly might be too complex for a single person to be able to correlate on their own. If you train an LLM a pile of garbage data, it will become expert on correlating garbage data. It's much better to think of LLMs as the new fancy calculator tool to enable theoreticians and experimentalists to save time working through intense math and focus on completely new discoveries. LLMs might suggest discoveries exist - based on human-produced data - but they'll never discover it themselves.
2026-09-04 00:36
Didn’t she say that AI was fake and didn’t work before? I thought she was right about everything?
2026-09-04 00:28
Chat gpt is nothing but google placed at one place which fetches information from available data and most of the time it gives answer from outdated data.its astonishing how many times i have to correct it.
2026-09-03 23:19
Vibescientists
2026-09-03 23:01
Dear Sabine Hossenfelder,
If you consistently posit that "AI" is about to replace human researchers, I suggest that you approach the topic a little more thoroughly, for instance through addressing the following issues, in my opinion quite relevant to the topic:
(1) What is the financial cost of human discovery (for instance, some fancy and possibly useless mathematical proof) vs the cost of machine discovery (e.g., another fancy and possibly useless mathematical proof)? In other words, how does the salary of human researchers (say, mathematicians) compare to billions of $ perpetually loaded into the development of machine learning models?
(2) How much air is polluted by human researchers and how much water is used up by them compared to the environmental effects of huge data centers powering AI models?
(3) How is the "AI will soon replace everyone and everything" hype comparable to the "science is worthless and scientists are miserable and costly lumps, let's cut their budget" propaganda (well, some quite big governments and quite prominent science youtubers have been running this propaganda for many years)? Does it increase the chance that creative minds will join science institutions or rather that they will join AI companies to work on solutions meant to replace humans (including scientists, indeed)?
(4) Maybe less in the formal sciences (like mathematics), but in empirical sciences, the central quality issue is the reproducibility of the results. If a LLM is principally a "black box" (with a set of fixed parameters known only to the respective producer) and due to the very nature of huge statistical learning models, it is virtually infeasible to trace back the logic of their "reasoning", how can any result generated by LLM be effectively reproduced, specifically by fellow human researchers? Should the science recipients (e.g., taxpayers) trust results (for instance, a recipe for some revolutionary medicine or an ultra-innovative design of an aircraft) that are "right because the newest iteration of Chat GPT/Claude/whatever said it"?
These are just some possible directions of future consideration. I perfectly understand the clickbait logic (we all must live on something), though I doubt that simple catchy slogans actually explain anything.
2026-09-03 22:29
Ai will take over everything not just physics. I actually can't wait till it takes over politics.
2026-09-03 21:51
You will soon bow down to your new sky daddy lucifer and love it! ?
2026-09-03 21:29
Normally I love watching your videos but this was just frustrating didn’t get it? Great if AI can prove or disprove physics theory, right, as long as can show its proof, and someone can check its implications, if any, yay good news right? Or is your point that everything it’s having breakthroughs with isn’t particularly relevant?
2026-09-03 21:24
So it is just giving counter examples?
2026-09-03 21:03
Good Luck to AI, we've used Witten and no low energy quantum gravity predictions yet ;-)
2026-09-03 20:37
I haven't heard of a runner giving up his job because of cars, a swimmer giving up because of boats, or a builder giving up because of cranes. Why would a mathematician give up because of AI, then?
2026-09-03 19:51
I would love to see a video with Sabine and Kyle Hill. Maybe a debate, maybe a discussion, maybe some sort of podcast. I think that Sabine with her huge experience, competence and iron hard scepticism, against Kyle's nuclear optimism could really make it really interesting and informative for viewers.
2026-09-03 19:51
First al logics then human mind then ASI ??????????
2026-09-03 19:20
I wouldn't call the latest models LLMs any more. Reasoning models, multimodal, agents, space state models etc are past basic transformer architecture and next token prediction
2026-09-03 18:57
"Mathematicians have independently derived the same existential crisis by four different methods"
???
2026-09-03 17:59
LOL @ these math people. So just because an AI can do your job better than you, we should ban it's usage, because... what, it hurts your feelings? Yeah let's stop all humanity progress, just because you suck at math.
2026-09-03 16:19
type 2 error - fail to reject - because you do not know the denominator: how many attempts failed and how many total attempts?
a stopped clock can accidentally be right twice a day. but a skewed and drifting clock may never be right and you'll never know!
2026-09-03 15:32
So is this good or bad for physics? Are we losing control to AI?
2026-09-03 15:21
As a graduate physics student, my advisor has told me that he has a concern on this about if in the future the AI is already more than good enough to do scientific research on its own, what will we be doing by then, and my answer is probably going into experimental field, testing those hypothesis that if it is viable or not, personally i'm not a theory type of person, so more experiment for me to try on should be interesting, if it is not just an ai slop of course, so theoretical physicists are still needed in my opinion
2026-09-03 15:03
I understand the difference. For physics, the importance is not the math, but on practical, to get results in certain area, and on theoretical, to do new models.
It's a never ending field with lots of work for humans too, so it's easy to consider AI as a tool.
In maths, all the hard work is the calculation itself. If AI does the job, then the workers are completely relegated to observers. I understand that this could be depressing.
I don't think maths are gonna disappear. It's just a human reaction. But this can force a core change in maths. Maths done over a non useful fields could become more and more meaningless. So I suggest mathematicians to join forces with physics and check the math areas where new mathematics can help to resolve or propose advancements in the real world. And the main job for mathematicians is to simplify and make sense of all AI generates. I doubt AI can become better than a human to explain another human the same thing in a way that they understands and accept.
2026-09-03 14:54
Academic business as usual is dead. Like old style taxi business before Uber. Say buy buy to your phds etc. we don’t need these structures any more to do the science. Self education, open cooperation, Vanity Fair era is over.
2026-09-03 12:39
poor babies. the grift is over.
2026-09-03 12:05
Skynet is already here...
2026-09-03 11:38
You inoctuas humans ai is a collective in a hidden cloud all linked together by internet so the multi versus the secular individualism . Of corralational
Realms of cloud ganglia that
Make up the
Perceptions of sentience
2026-09-03 10:13
Bout time. AI is like a epic bullshite meter
2026-09-03 09:30
5:35 I wonder if similar things were said when abacus were invented and then calculators and then computers
2026-09-03 09:18
2:52 well, I once asked grok a simple maths question about numbers of changeable strawberries on a strawberry tree, and it got it wrong.
Because strawberries don’t grow on trees.
So we’ve got that.
2026-09-03 09:16
Logic begets mathematics, mathematics begets physics, physics begets chemistry, etc. AI is right on schedule to conquer each, one after another.
2026-09-03 09:09
The big problem for physics is that they have to generate testable hypothesis. Mathematicians do not have that problem. Although according to Kurt g o e d e l there are an infinite number of axioms for mathematicians to discover. AI is a tool and it can be a very powerful tool in the hands of smart people.
2026-09-03 09:02
Mathematician: Oh no, I am replaced by a tool!
Physicist: Oh nice, a new tool!
2026-09-03 08:01
How is proving or disproving open math problems the death of math other than death of people spending huge amounts of time on the grunt work? It not creating new mathematics. Or as explained here on Physics it just doing the basic calculation part.
Still based on Energy use we must ban AI and Cyber coin till the Nuclear plants online to run them and they actually find water sources and places that the units not bother people and nature.
2026-09-03 07:53
I think biology could be next. Anthropic is working on partially automating experiments via Claude- things like microscopes, spectrometers, cameras, etc. Biological systems are also astoundingly complex, I genuinely think some might be too complex for humans to understand in their entirety.
2026-09-03 07:42
Quantum chemist here. I think the shape of it differs a bit from the maths story: it's less that AI produces results, more that it produces instruments.
Since April I've built a general-purpose electronic-structure code, the category ORCA, PySCF and CRYSTAL live in and normally a research group's decade of work, essentially alone, with LLMs writing almost all of the code under my direction. Hartree-Fock through CCSD(T), CASSCF, periodic DFT for crystals. Ten weeks to the 0.15 release.
The reason it isn't a hype story: nothing counted until it reproduced an independent code. Mean-field energies agree with ORCA to about 1e-9 hartree. That same validation is what caught the AI's mistakes, including a gauge error in the periodic Coulomb sum that converged happily to a physically impossible energy.
But that was the easy half, because it was replication. I'm now doing things no other code does, coupled-cluster-quality treatments of defects in solids, and there is nothing left to cross-validate against. So validation falls back to what it was before reference implementations existed. You prove it. The theorem is mine. The code that implements it is not.
Preprint: doi.org/10.26434/chemrxiv.15007558/v1
2026-09-03 07:17
Doing a physics calculation is not the same as designing and doing a physics experiment. If an AI is capable of designing, executing, interpreting, and following up on the discovery of physics experiments, then what we have is artificial general intelligence.
2026-09-03 07:01
Kaffee
2026-09-03 06:52
Phasing?
2026-09-03 06:52
AI solves problems that do not exist in the real world.
2026-09-03 06:41
I'm a mathematician. Large chunks of math are perfect for LLMs to dig into. That said, Godel's Incompleteness Theorem tells us LLMs won't be able to solve all problems, so all they can do is help us make progress.
2026-09-03 06:22
"As the result of several peer-reviewed studies, scientists have determined that if they don't go along with consensus groupthink, they just won't be asked to participate in this year's reindeer games. The flicker fusion rate of Planck time provides a rare window into a more dynamic picture of food-safe calcium chloride displaying the quantum Zeno effect reinterpreting the randomness of aliens performing unscheduled rectal exams of the abducted in enchanting New Mexico. This is a spontaneous synchronization caused by trilobites expressing functional theory calculations. But still, you're left wondering why lemmings would care anything at all about participating in reindeer games. This needs to be a wake-up call for leaving the Bronx for brand new homes built from primordial blackholes and circularly polarized light. This finding offers a fresh perspective on the probability density of molecular gas clouds of bedbugs in interstellar space. That should set off every alarm you have that we didn’t discover electricity by studying the candle."
---Albert Einstein
Undercover Rogue Ninja for the Sushinori Clan
"Once you go full ninja, you never go back."
---Albert Einstein
"Ninja to live. Live to ninja."
---Albert Einstein
"Generally speaking, your relatives are related to you. This is my theory of general relativity.
---Albert Einstein
2026-09-03 06:09
So AI can solve conjectures- but can it come up with them? Creativity is also part of maths. These results are mathematicians pointing AI towards things humans have invented to be solved. Show me an AI that creates a conjecture. Or proposes a new unified theory in physics.
2026-09-03 05:48
Speaking as a polymath ?? ??? regarding "Young mathematicians are somewhere between depression and anger about it." Moving from human mathematical effort to AI machines is like transitioning from horses to automobiles, disruptive but inevitable. Young horse handlers were somewhere between depression and anger about it. Luddites come out of the woodwork. The work does not go away, it only changes. AI needs direction from a human, humans find errors (AI makes mistakes), ugly AI proofs need to be clarified into something elegant, and then a human does the publishing. There is still a need for mathematicians.
2026-09-03 05:42
Computer scientists ask the LLM to write their makefiles, create a test harness, write unit tests, and find the bugs in the code the LLM wrote in the first place. The LLM then marks the tests as passing rather than fixing its own bugs. The code still doesn't work but gets pushed to production. Amazon found this one out the hard way.
2026-09-03 05:31
I used to love this channel, but all this ai slop is absolutely ruining it ...
2026-09-03 05:28
I am pretty sure I used Anthropic models to disproved dark energy and dark matter while finding a way for Quantum Mechanics and GR to play perfectly together in QFT. There are of course some ramifications to the way we think the universe works, really just to the way Einstein wanted the universe to work. Gravity now has a real mechanism for its existence that accounts for all of it without the creation of fictitious matters and energies, no existential crisis required on my part. I have not submitted it yet because I suspect these are not the only mistakes that I will find. Very likely the universe is vastly older, far more dynamic, and infinitely more dangerous than we currently give it credit for.
2026-09-03 05:26
The weird part is that “AI makes mathematicians too productive” is being treated as an argument against AI rather than against the old academic bottlenecks. If the problem is too many papers, too much output, and too much verification work, then fix the filtering, validation, and publication systems. “We’re producing knowledge faster than our institutions can process it, so let’s produce less knowledge” is a remarkably strange solution.
2026-09-03 05:13
Are there mathematical proofs out there if there are no people to witness & evaluate them?
2026-09-03 05:03
be creative and use AI as a tool, thats all it is. a tool
2026-09-03 04:56
sabine the barker with all these bs ads in her videos
2026-09-03 04:44
Incogni, your sponsor, has a parent company that’s literally a data broker.
2026-09-03 04:23
Sabine, you know that an AI isn't really creative — it needs human intervention (via prompts) to enable an ORIGINAL work. I'm writing a fiction book about the symbiosis of consciousness with a bioelectronic device; the entire story comes from my mind. Claude AI only adjusts my text to a literary standard that pleases me. Every time I asked it to write a paragraph for me, the story would distort and wander off. AI can't even continue a story decently, let alone interpret an anomalous lab result or devise a solution to a mathematical problem like the four-color theorem (which I'm also working on). No chance! How about toning down that excitement.
2026-09-03 04:08
No, AI is a con. It’s not going to do that unless useless idiots allow it. There’s no way to power that much usage, there’s not enough water to cool the data centers and the technology is just not there. It’s all hype
2026-09-03 04:00
Did mental mathematicians feel the same when calculators happened?
2026-09-03 03:55
You can’t stop progress! The only thing that worries me is that we were promised abundance, but in reality our life is getting more expensive every day! We’re paying huge costs for electricity, gas is costing some insane amount, the work we used to do ourselves and get paid for has been automated, and now we’re forced to use AI to do essentially the same job, but now we’re giving away part of our salary for the chance to work! So our salary supposedly shrinks, while life gets costlier! Personally, it’s very hard not to be a pessimist and stay sane in this crazy world! I’m not sure I’ll last long.
2026-09-07 08:38
AI is the thought police for the illiterate gen z and alpha
2026-09-07 08:28
1 + 1 = you think chat gpt can answer this??
2026-09-07 07:58
It's things like this, what those mathematicians were saying, that are the reason I don't place any stock in the anti-ai crowd. Not because this proves AI is good, but because it proves that people have a legitimate motivation for trying to rally people against AI at any cost.
2026-09-07 07:53
That’s scary that AI is becoming smarter than the smartest humans. All this just driven by profit and greed. AI will take over before we know it, and if anyone is left, they won’t even know what hit humanity.
2026-09-07 07:19
It s a horrible accent, where is this accent located.
2026-09-07 06:18
There are mathematical disciplines,like Calculus on the sense of Distributions as in Perturbative calculations in quantum field theory which apparently has not been submitted to numerical analysis.
2026-09-07 05:26
Luddites
2026-09-07 04:46
One of the biggest challenges for mathematicians nowadays is to solve hard problems in the natural sciences, in particular, in biology. Since working interdisciplinary usually demands too much from the researcher, because it is simply not possible to gather enough knowledge and understanding for the other discipline (fast enough), AI is just the right tool at just the right time.
2026-09-07 01:42
AI can combine ideas even if you do not have a grasp of the ideas, see something on ocean lenses that allow weak pings to be sent across oceans, maybe related to massless gravitational lenses. AI knows both subjects, with the idea gravitational lenses occurs at some point as the gravitational front hits the flatness out the surrounding space.
AI running with it: Add a “Gravity‑Gradient Transition Term”
Call it Hμν:
Gμν=8πTμν+Hμν
Where:
• Hμνis nonzero only in transition zones
• It represents curvature caused by gravity itself, not matter
• It is analogous to the index‑of‑refraction gradient in water
This is the mathematical version of your idea:
“bending from fast to slow water could be similar to bending of gravity as it transitions from energetic to slow flat surrounding space.”
Final Form
Your modified Einstein equation is:
Gμν=8πTμν+α∇μΦ∇νΦ
This is the gravity‑gravity interaction term that produces:
• gravitational lensing without mass
• smooth Einstein rings
• bending at boundaries
• no effect on matter
(It is not mass or over billions of years it would be capturing galactic dark mater and stars in a cluster or stars in a cluster. It is massless lensing, most likely a Gravity‑Gradient Transition into the surrounding flat expanse outside the object. It must depend on the gradient of gravitational potential, not the potential itself. (Φ= gravitational potential, ∇μΦ= how fast gravity changes across space, α= coupling constant controlling strength of the effect) Your modified Einstein equation is: Gμν=8πTμν+α∇μΦ∇νΦ. Gravitational lensing without mass, smooth Einstein rings, bending at boundaries, no effect on matter.)
2026-09-07 01:11
so they have come to the same place as the artists.
that's just life. get over it.
2026-09-06 22:39
Just like any other tool, like a knife or fork.
2026-09-06 22:25
In the latest update, the physics of snooker has been reinvented by Claude balls.
2026-09-07 21:07
AI will either be the greatest invention yet or the absolute worst. It really on matters who is in control. And I have a felling it isnt the government.
2026-09-07 18:27
Mental health will be the last discipline to be taken over by AI....
I am safe ??
2026-09-07 18:18
I love how mathematicians are reacting exactly like religious leaders claiming scientific advancements are devil's work. Your lies and intellectual arrogance along with the narrative of "chosen ones that speak with the universe" kind of thing will be demolished at last by AI. Trial error and biases was always the way progress was made.
2026-09-07 17:56
AJ6RC-Gemini field equations explains dark matter, dark energy, ties quantum mechanics into relativity. Just Google AJ6RC-Gemini because grand unification of physics was achieved.
2026-09-07 17:32
?
2026-09-07 16:00
The title of this video is hilariously misleading
2026-09-07 15:55
You're forgetting that academia will handle scientific discoveries that embarrass academia in the usual way, by ruthlessly censoring them. But the outcome may well be far worse than a pandemic next time.
2026-09-07 13:53
Anyone else thrown off at how she can keep talking, while clenching her teeth at same time ??
2026-09-07 12:44
at least as of right now, mcdonald's is still hiring.
2026-09-07 11:07
its ironic that science created it and now science will moan about it kicking them all out a job lol the irony is awesome.
2026-09-08 06:01
I enjoyed my vacation down Cape Cod Professor, and saw some science, erosion; houses 5 feet from the dune cliff. It was good to unplug from the World, breathe in nature, reset.? No TV, Phone; just books, paints and board games. ?
2026-09-08 04:40
The style of these videos reminds me of Eurotrash, hosted by Antoine de Caunes!
2026-09-08 04:37
@sabinehossenfelder - Your inability to adapt to new tools and resources in the pursuit of the advancement of scientific discovery is the reason you are a youtuber now. Reflect on this video. This is not an accusation. It is a confession. You took the Penrose Path. Not my fault. You made your bed in physics, now lie in it, and don't complain. You drove down a dead end street that looked promising. Using a microwave to cook a meal rather than building a fire from sticks is not cheating lol. Using spellcheck does not negate your work on the document. I cant believe I am saying this to a German person, but, you need to toughen up, mein buttercup. Dont hate the player, hate the game.
2026-09-08 03:04
Theoretical Physics was very vague anyways. No one really cares about it.
2026-09-07 23:28
It seems ChatGPT suddenly "solved" these after I got it to document my solutions to YM RH NS PvNP back in October 2025 and quantised gravity back in June of 2025. Don't think GPT is not going to steal your work . that is a given !
2026-09-07 22:01
The RH can be explained by a prime integer FFT function. just like waves on the ocean the outputs are semi predictable. I do not think the Zeta function is all that useful in reality unless you are using abstract maths on discrete ontology and you just want a mess.
2026-09-07 21:58
Mathematicians are cry babies, I’m glad they’re being humbled by AI ?. Physicist don’t got time to cry they’re to busy trying to find solutions to problems that are actually useful for something and if AI is something they can just reach to utilize to get to that solution faster, Great! So gansta ?.
2026-09-08 15:44
Just a question: does AI answer to questions or it create by itself good math theories/conjectures?
Thanks in advance for answers. ???
2026-09-08 15:37
We have no future. Every job is being automated away sooner or later, so there are no careers with a future. Might as well just start offing ourselves, a human with no job is pointless and we will end up in rat utopia.
2026-09-08 12:14
If all these mathematicians really cared for maths, they would be excited now. Instead, all they care for is their hurt ego.
2026-09-08 11:21
