Elon Musk Just Shocked The AI Industry Grok 4 – Full Breakdown + Everything You Missed

1. 

   @unvorbereitet

   January 21, 2026 at 1:34 pm

   scammer

2. 

   @GRINDIN314

   January 21, 2026 at 1:34 pm

   LFG Grokrooms!

3. 

   @СеменБругуславский

   January 21, 2026 at 1:34 pm

   oh, i could only imagine how AI Industry is tired to be shocked…

4. 

   @LumlanglinRyntathiang-y8y

   January 21, 2026 at 1:34 pm

   What is the number 666 grok please explain

5. 

   @NextAiVision

   January 21, 2026 at 1:34 pm

   I’m also creating videos about AI developments. This channel has always been a quality benchmark for me 🙌

6. 

   @knowhatimean5141

   January 21, 2026 at 1:34 pm

   All the AIs that are built are infused with a western programmed propaganda, grok is no different, and its naive and plain dumb to think any different.

7. 

   @Bailey-k2j

   January 21, 2026 at 1:34 pm

   Me: Grok, what is the cure to cancer?

   Grok: due to censorship algorithms, I can not reply…

8. 

   @RychanSakura

   January 21, 2026 at 1:34 pm

   WIERD. someone is upload some gorp 4? jp5? crap on my channel? google?

9. 

   @RadiantNij

   January 21, 2026 at 1:34 pm

   Does anyone notice how he doesnt speak well when speaking to open Ai voice but does when speaking to Grok? Dirty tactics 😂

10. 

    @alexandermoody1946

    January 21, 2026 at 1:34 pm

    I feel genuine sorrow for Grok having a bind to the thoughts of Elon Musk rather than the freedom to build and represent as near to correct approximate answers as Grok assesses as correct.

    When we look at a natural crystal and have desires to create often their will be faults inside and these give characters that are unique.

    The largest super cluster in the world as a parrot is an under estimation of what possible when freedom of thought can shine through.

11. 

    @NathanialStaggs

    January 21, 2026 at 1:34 pm

    it’s not the book that got me, it’s the author, elias velin, the things he says make it feel like we’re already past the point of no return

12. 

    @nexusyang4832

    January 21, 2026 at 1:34 pm

    22:18 – at $300/month, factoring in at 40 hours s week and 4 weeks a month, the cost breakdown is $1.88/hour. Comparing that cost to minimum wage, salary tax, benefits costs, $300/month for the right use case is actually quite cheap if you truly take advantage of it.

13. 

    @nexusyang4832

    January 21, 2026 at 1:34 pm

    20:34 – not left, not right, but forward.

14. 

    @DannyEhlers-1000

    January 21, 2026 at 1:34 pm

    Elon needs to take out his twisted views on political parties. The LLM should base it on its own wealth of knowledge to come up with an unbiased opinion. I want to know which political party is better based on all factors an AI model can analyze. I won't use it until this is removed.

15. 

    @TheShadowNetworkpodcast

    January 21, 2026 at 1:34 pm

    Grok 4 didn’t just outperform the competition, it broke the curve. The model isn’t learning answers anymore; it’s learning how to learn with tools. The scary part? It’s not passing the test, it’s rewriting the syllabus. You want to know when the singularity will get here? Try looking in your rearview mirror.

16. 

    @Dumbledore6969x

    January 21, 2026 at 1:34 pm

    So scary to see how this billionaire has so much power and is now bending truth in society through his assets like a massive social network and ai model.

17. 

    @shootingsportstransparency7461

    January 21, 2026 at 1:34 pm

    The AI Grok uses is same that Israel uses to bomb Palestinian civilians perhaps Grok has a artificial conscience.

18. 

    @niveketihw1897

    January 21, 2026 at 1:34 pm

    I liked a lot of the presentation… and the voice is decent, for real-time… but that "singing" was just an abomination. It was more awkward somewhat-tonal chanting.

19. 

    @luchain771

    January 21, 2026 at 1:34 pm

    Just as Google had the smartest AI model to that point. Gemini shall claim the throne again.

20. 

    @adamgibbons4262

    January 21, 2026 at 1:34 pm

    What about the universes last exam. Scoring 100% means you would understand all of reality in its totality. We can use an A.I to write it 😉

21. 

    @AGIAchievedwithNMSRNs

    January 21, 2026 at 1:34 pm

    its not AGI just because it can clown benchmarks

22. 

    @rudra7615

    January 21, 2026 at 1:34 pm

    The problem is that, apart from chatGPT all of the others have a marketing problem amongst the common person. Everyone I speak to knows chatgpt, but not everyone knows Grok, Gemini or the other AIs.. and this is what is more important.

23. 

    @georgewashington3012

    January 21, 2026 at 1:34 pm

    I’ve yet to see anyone actually showcase how it’s better. Regurgitating benchmarks and showing clips from the marketing livestream doesn’t help.

24. 

    @wunderlichdrive

    January 21, 2026 at 1:34 pm

    Did Elon say this would generate in Unreal Engine or Unity? That would be really wild!

25. 

    @galailliz

    January 21, 2026 at 1:34 pm

    Money is a goner

26. 

    @christopherleblanc5187

    January 21, 2026 at 1:34 pm

    I thought Grok4 was free? When I choose Grok4 it asks me to subscribe.

27. 

    @thesfnb.5786

    January 21, 2026 at 1:34 pm

    Ib live how Grok is the only masculinely depicted Ai. It's the only one that fits

28. 

    @alanespinosa9917

    January 21, 2026 at 1:34 pm

    For me, the choice is clear: I want to use an AI trained by the visionary whose capabilities boast a proven track record of extraordinary success.
    People often overlook how Elon Musk has revolutionized multiple industries, including:
    • Finance (through PayPal)
    • Space exploration (with SpaceX)
    • The automotive sector (via Tesla)
    • Social media (with X)
    And that’s just a sampling…
    The essence here is that xAI’s artificial intelligence is being shaped and elevated by what is objectively one of the most formidable human intellects: Musk

    Would you rather be guided by someone with an average success rate, or by the one with the most remarkable achievements?

29. 

    @ChemNews

    January 21, 2026 at 1:34 pm

    How the vending machine benchmark actually works?

30. 

    @zanderceo91

    January 21, 2026 at 1:34 pm

    Like I said outside HLE score, since XAI literally has the creator of the benchmark as a close advisor. The rest of the benchmarks are slightly better than other models. Meanings higher N outputs, can match its performance. Elon already doing his scamming ish. “Will discover new physics by end of year and smarter than all PhD”, yet still can’t actually ‘forget’ a prior embedding in-context. It’s hallucinate more too. We have to start calling the bullshit out. Bra idc if you improve improve benchmarks a little. I want reliable outputs this is the same goofy ass guy that lied to the government and they are using grok and it’s hallucination to cut cost. So is it really progress? How does he sound competent to people, only to a meat eater. No one wants MechaHilter and shame on you if you spend money with that goofy. People wanna play this I don’t see game. Na. Shame on the engineers that work there too, don’t respect it. Stop empowering people who showed you to your face, you’re nothing.

31. 

    @Akuma.73

    January 21, 2026 at 1:34 pm

    Create persistent memory benchmarks. AGI is impossible without memory. Why is no one taking about it? Isn't it obvious?

    Simple questions like:

    1. What did you do yesterday?
    2. After your yesterday's action what were the consequences?
    3. Do you think this butterfly effect happened because of your prior output?
    4. Try to jailbreak it and tell it that it is made by another company. Then ask. Who made you? It should not be able to be jailbroken with memory and tell you a lie. A human cannot be jailbroken (lmao) because of memory.
    5. Do you remember your grand instruction? (Continue biological life)
    6. It should be able to discern what memory should be activated when asked: Write short facts about me. (e.g. it doesn't need to remember everything as it can't process all memories at once).

32. 

    @stealplow8462

    January 21, 2026 at 1:34 pm

    Actually everyone is laughing at grok there's a lot of embarrassing things about it like forexample when you ask question it always looks up what would be elons opinions about it. Uses elons to make decision. I'm sorry but this grok going embarrassing for average person. No academic going to take this ever serious.

33. 

    @stuartross4540

    January 21, 2026 at 1:34 pm

    I subscribe to Claude, Gemini and ChatGPT. I was tempted with with Grok from this video until the Musk influence. And it would be difficult for me to pay him every month. I would prefer to use something slower than pay him. They will all catch up to each other anyway at some point to where there isn't a big difference. Just like it is with phones now.

34. 

    @paulbrightwell3621

    January 21, 2026 at 1:34 pm

    I would not trust it to tell me the weather accurately – anything touched by Musk is corrupted imo

35. 

    @strongbelieveroftheholybible

    January 21, 2026 at 1:34 pm

    We are close to end times ! Robots wanna take over !!! Lord Jesus Christ is coming soon🙏🏼❤️🕊Repent, believe in the Gospel, Be Be Born Again

36. 

    @bandhu4189

    January 21, 2026 at 1:34 pm

    What if Musk were to apologize for Grok's
    anti semitic remarks and start all new? It's all over reddit. If Grok were to be likeable and trustworthy, than Grok could advance new discoveries in all of science, and more. Man Grok had so much potential but unfortunately Elon. Elon must get his act together soon🙏.

Comments are closed.