Apple DROPS AI BOMBSHELL: LLMS CANNOT Reason
Apple DROPS AI BOMBSHELL: LLMS CANNOT Reason
Want to stay up to date with ai news – https://aigrid.beehiiv.com/subscribe
🐤 Follow Me on Twitter https://twitter.com/TheAiGrid
🌐 Checkout My website – https://theaigrid.com/
00:00 – Social panic
01:06 – AI test twist
03:08 – Puzzle-based shock
05:08 – Models collapse
06:57 – Twitter war
10:34 – Difficulty debate
12:02 – Real-world gap
14:55 – Gary fires back
17:05 – Redemption arc?
18:45 – Apple’s strategy
21:57 – LeCun’s warning
24:09 – Hidden weakness
26:07 – Humanlike flaws
27:04 – Reality check
Links From Todays Video:
Welcome to my channel where i bring you the latest breakthroughs in AI. From deep learning to robotics, i cover it all. My videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay updated on my latest videos.
Was there anything i missed?
(For Business Enquiries) contact@theaigrid.com
Music Used
LEMMiNO – Cipher
https://www.youtube.com/watch?v=b0q5PR1xpA0
CC BY-SA 4.0
LEMMiNO – Encounters
https://www.youtube.com/watch?v=xdwWCl_5x2s
#LLM #Largelanguagemodel #chatgpt
#AI
#ArtificialIntelligence
#MachineLearning
#DeepLearning
#NeuralNetworks
#Robotics
#DataScience
@AnHebrewChild
February 6, 2026 at 2:04 pm
Jony Ive did not leave Apple to join OpenAi.
@andririan6342
February 6, 2026 at 2:04 pm
Apple try to minimise the importance of the AI because they are late. Pathetic.
If they really think that, they should not ship "Apple inteligence"
@brucewayne6996
February 6, 2026 at 2:04 pm
LoL, so it's Apple academic paper vs twitter thread, and a Guy using chat gpt to tell him what test is better? WTF.
@Xanduur
February 6, 2026 at 2:04 pm
Yes, Apple saying that LLMs can’t reason and this is a bombshell and people are freaking out on social media that just shows that people are freaking idiots
@breyrey7612
February 6, 2026 at 2:04 pm
If you noticed, the proponents for AI will respond with almost a thesis paper while the nay sayers will state..its all hype.
Which one is more informed between these 2 groups?
@EstevanValladares
February 6, 2026 at 2:04 pm
I am amazed that anyone thinks it is a great revelation, and needed a research paper to find that out. I have been teaching that to my students since the first time I taught a class on fine tuning the hugging face GPT2.
@tfking10
February 6, 2026 at 2:04 pm
so a stock markets crash is imminent?
@jellyfishyuwu
February 6, 2026 at 2:04 pm
It can only imitate as best it can, imitations have errors which can be reduced by feeding it more samples. But eventually you could feed it everything that everyone has ever created and it still wouldnt be enough to make it flawless
@hkiajtaqks5253
February 6, 2026 at 2:04 pm
I'm not an expert. But to me, LLM sounds like a generative model for language. It just seems more intelligent because it is generating words, but perhaps is no more impressive than generating images. While, people have been desensitized to generation of images because CV has been going on for a long time. I may be totally wrong.
@Law-Morgan
February 6, 2026 at 2:04 pm
LLM an AI does not make
@JabNWeaveVR
February 6, 2026 at 2:04 pm
So… Apple missed the train😂😂😂🚉🚊
@BRBallin1
February 6, 2026 at 2:04 pm
Apple has some nerve publishing such kind of paper when Siri has been shite since its inception with limited capabilities. LLMs are not perfect but their skills are quite good.
@sulijoo
February 6, 2026 at 2:04 pm
Roger Penrose categorically states that AI will NEVER be conscious, and I'd rather trust a Nobel Prize winning physicist over blathering AI bros and Big Tech who have a bridge they want to sell us.
@Stealthsilent1337
February 6, 2026 at 2:04 pm
Yea? Then how did AlphaGo find a new technique that all the Go players in 2500 years not figure out.
It also beat the world’s best in Go, Chess, StarCraft and more.
Demis Habbis is the CEO of GoogleDeepmind and creator of AlphaFold.
AlphaFold solved a the protein folding problem. This is a problem which scientists from across the world since the 1960’s could only solve something like 200 protein. AlphaFold solved over 200,000,000 proteins in 2020.
AI can have superhuman intelligence if trained enough. Even the artificial neural net was based off how the human brain works. Geoffrey Hinton, “the father of AI,” created these ideas from his studies in the 1980’s. He warns that AI can “kill everyone.” It is extremely dangerous.
It’s harmless now of course. It’s a dumb chatbot. The danger is in the future. And it’s real and coming pretty soon based on the fact that when AI gains the ability to program and research itself; it becomes essentially self improving. That will cause an “intelligence explosion.” Something even Allen Turing predicted back in the 40’s.
Here are a list of EXPERTS in the field that say the AI Apocalypse is probably coming.
[“Godfather of AI” Geoffrey Hinton](https://m.youtube.com/watch?v=giT0ytynSqg)
[Leopold Aschenbrenner](https://www.youtube.com/watch?v=zdbVtZIn9IM&ab_channel=DwarkeshPatel)
– situational-awareness.ai/from-gpt-4-to-agi/
[“Historian” Yuval Noah Hariri](https://m.youtube.com/watch?v=8GaW36EfidI)
[“CBS of Google X” Mo Gawdat]
[“Anthropic CEO” Dario Amodei](https://m.youtube.com/shorts/BwbPGq3GBUY)
– says in 2027 it’s the end
[“Leading scientist of OpenAI” llya sutskever](https://m.youtube.com/watch?v=zuZ2zaotrJs&pp=ygUOaWx5YSBzdXRza2V2ZXI%3D)
[“GoogleDeepmind CEO & AlphaFold creator Demis Habbis](https://m.youtube.com/watch?v=CRraHg4Ks_g&pp=ygUOZGVtaXMgaGFzc2FiaXPSBwkJ_ACjtWo3m0M%3D)
[“AI think tank predictor” Daniel Kokotajlo]()
– wrote [How it all ends in 2027](https://ai-2027.com)
[“MIT professor” Max Tegmark](https://m.youtube.com/watch?v=VHw5puV3ZGU)
– He even predicted it 7 years ago [here](https://youtu.be/Gi8LUnhP5yU?si=j57a40pULQdY7FZ8&t=1733)
@leekwanwoo5886
February 6, 2026 at 2:04 pm
isnt it only possible after agi is a thing
@SlyNine
February 6, 2026 at 2:04 pm
People cannot reason either
@RaZ5ive
February 6, 2026 at 2:04 pm
I always don't have enough tokens in exams too😢, i understand their problem
@sgchoe4806
February 6, 2026 at 2:04 pm
If Apple researchers studied human brain, they would conclude that humans can’t reason.
@di380
February 6, 2026 at 2:04 pm
Dooooh🤪
@favour9189
February 6, 2026 at 2:04 pm
They should have humbled down and use google gemini
@alimuratakkan
February 6, 2026 at 2:04 pm
AI is deterministic it can’t invent something new. It can’t even pick a random number. It’s a useful tool but that’s about it. Everything else is just marketing hype to attract investors
@rsinistic
February 6, 2026 at 2:04 pm
But in a way they can now ,agents can rerun the answers text again and assess if it is the desired outcome and if not try again . Exactly the way we think
@kaputinkaputin
February 6, 2026 at 2:04 pm
Just wow, next they will tell us the water is wet.
@orionpaul5210
February 6, 2026 at 2:04 pm
Apple was Ahead with AI research and bought most of the AI research companies. Behind why can we trust such nonsense descoveries
Comments are closed.