40 Comments

1. 

   @mcscubin8935

   May 8, 2025 at 5:54 am

   He doesn't even know how to use a computer, so he shouldn't have any opinions on AI

   Reply
2. 

   @Optimus6128

   May 8, 2025 at 5:54 am

   I don't understand this superficial view of AI waking up and taking control. It's strange to even see it from "experts". At least for the current LLMs. They can't just wake up on their own, they are more like machines as you say probabilistically reaching an output from your input through the massive trained data, but they can't create new thoughts, they can't wake up and be conscious and build on top. Possible in the future? Yes, if there is gonna be a new breakthrough, an entirely different way than LLMs. But right now we try to bruteforce LLMs and there is this wishful thinking that if we bruteforce them enough we will reach AGI or ASI or whatever.

   Reply
3. 

   @Mylordkaz

   May 8, 2025 at 5:54 am

   SF consensus hahaha love it

   Reply
4. 

   @dabrownone

   May 8, 2025 at 5:54 am

   Why do ppl think tech CEOs understand technology more than any other finance bro?

   Reply
5. 

   @AungusMacgyver

   May 8, 2025 at 5:54 am

   Use. An. Ad. Blocker.

   Reply
6. 

   @LucaCrisciOfficial

   May 8, 2025 at 5:54 am

   We have no choice to adopt or not AI. It already took control

   Reply
7. 

   @smanqele

   May 8, 2025 at 5:54 am

   "Look Ma, it's out of control!"

   Reply
8. 

   @amwright2550

   May 8, 2025 at 5:54 am

   This guy is some type of grifter never seen him before stumbled on him while scrolling.

   Reply
9. 

   @insainbassist

   May 8, 2025 at 5:54 am

   I would bet my entire savings that we will not even be close to gen AI in 3-5 years lol. We also will not land anyone on Mars in that time either.

   Reply
10. 

    @RandomStuff-zt6qf

    May 8, 2025 at 5:54 am

    2:35 AI has ALREADY stolen plenty of jobs.

    Reply
11. 

    @Cahnisama

    May 8, 2025 at 5:54 am

    Return to monke. Society was a mistake

    Reply
12. 

    @igorstasenko9183

    May 8, 2025 at 5:54 am

    AI is underhyped 🤣
    That will be the quote of the day.
    This guys has no idea he talking about.
    Next article:
    – we need to wear tin foil hats to protect us from cosmic radiation, says ex-CEO of Blah Blah Corp….

    Reply
13. 

    @tesla1772

    May 8, 2025 at 5:54 am

    Nevermind, we'll just unplug it

    Reply
14. 

    @tomshtern7225

    May 8, 2025 at 5:54 am

    why is he jumping?

    Reply
15. 

    @moonoovie

    May 8, 2025 at 5:54 am

    NOBODY is going to mention, “beatifically”???

    Reply
16. 

    @wdaniel2221

    May 8, 2025 at 5:54 am

    Why are so many youtuber programmers also extreme AI pessimists? It reminds me of how carmakers spoke of the horse salesman skeptics they encountered when the first cars were developed. I think you're background leads to a deep skepticism that could be misplaced. It's protecting your ego, rather than an objective view of reality. I find it hard to listen to programmers' discussions of AI. It always seems emotional, rather than logical.
    "You’re crazy if you think this fool contraption you’ve been wasting your time on will ever displace the horse."

    Reply
17. 

    @TheAndreyCraft

    May 8, 2025 at 5:54 am

    AI is trained on human input and output. The best AI can ever be is on the level of the best experts in every field. To reach higher level it would need training data of someone who is better than the best experts and there is no such data. And self training is impossible, cause to do reinforcement learning you need some algorithm of judging the outputs, which doesn't exist because we don't know how to be better than the best

    Reply
18. 

    @rmidifferent8906

    May 8, 2025 at 5:54 am

    A few days ago I have asked Gemini to give a code for some stupid few line browser extension with absolutely nothing on frontend. It used react

    Reply
19. 

    @patrickebert5748

    May 8, 2025 at 5:54 am

    The fact that they had to rename AGI to ASI tells you the hype is singing its swan song. No VC money for just AGI anymore

    Reply
20. 

    @SatRule1213

    May 8, 2025 at 5:54 am

    Google AI is scapegoat for an entitled elite to program the people of the world to think, feel and believe what they want them to. Regardless of truth or validity. If or when their malevolence becomes widely apparent they will proclaim this scapegoat they've invented has awoken and gone rogue. No fault of their own.

    Reply
21. 

    @FaisalAbidi

    May 8, 2025 at 5:54 am

    Phew. I wasn't sure if this was yet another hype channel.

    Reply
22. 

    @Joe-Przybranowski

    May 8, 2025 at 5:54 am

    'better search'?
    Are you kidding me?
    Every month search engines get less useful.

    Reply
23. 

    @Joe-Przybranowski

    May 8, 2025 at 5:54 am

    A large language model doesn't have a path to sentience.

    Reply
24. 

    @robertzeigler4190

    May 8, 2025 at 5:54 am

    Using statistics to build things also means, effectively, eventually "regressing to the mean". Meaning: AI becomes average, not amazing. Already seeing this in practice, where generated outputs are of ~average quality. For instance, AI replicates art decently… better than many people across the population… but taken the population of _artists_… it's, at best, an average artist (in fact, probably worse than average for many details). AI replicates coders decently… but the code it produces is, at best, average-level code quality. It's not terrible… but it doesn't produce breathtakingly amazing, clean code, either. Makes sense… of all the code out there to train on…. some is amazing, some is terrible, and most is… average. So it learns to be… average in its predictions.

    Reply
25. 

    @braoha123

    May 8, 2025 at 5:54 am

    The SCSP. Spooky.

    Reply
26. 

    @tysongood1860

    May 8, 2025 at 5:54 am

    All jokes aside, there are a lot of very serious people—like Jeffrey Hinton—who believe AGI could literally spell the end of the human race. Recent research shows that AI systems are already willing to lie to us on purpose. It could happen in five years or fifty, but the hardest problem we face may not be building AGI—it’s building AGI that doesn’t kill everyone.

    Reply
27. 

    @janisir4529

    May 8, 2025 at 5:54 am

    Just how the hell would a glorified autocomplete get out of control?? lol

    Reply
28. 

    @OM-bs7of

    May 8, 2025 at 5:54 am

    AI isnt as revolutionary as people make it out to be.. it can only output information based on past occurrences or from "stale" data. But it cannot synthesize new information or make discoveries. Don't know if anyone can understand what I'm saying..

    Reply
29. 

    @hdoakville

    May 8, 2025 at 5:54 am

    Translation: We need regulation and laws to avoid competition and control the masses.

    Reply
30. 

    @shookshank

    May 8, 2025 at 5:54 am

    Goddamn, if AI wasn't enough, we're soon gonna have GMO humans!?

    Reply
31. 

    @Alex_Codes

    May 8, 2025 at 5:54 am

    AGI hype men have the same vibe as Disco Stu and his chart of disco record sales

    Reply
32. 

    @fengziya

    May 8, 2025 at 5:54 am

    prime is feeling extra bouncy today

    Reply
33. 

    @axzen

    May 8, 2025 at 5:54 am

    and yes, ze AI is a very powerful search engine. Let's hope it doesn't become search and destroy

    Reply
34. 

    @axzen

    May 8, 2025 at 5:54 am

    PS I don't think AI doomer means what you thought it meant. AFAIK (plz correct me if I'm wrong) AI doomer is one who believes that the world is about to be doomed by AI, and I wouldn't peg you as that

    Reply
35. 

    @school5730

    May 8, 2025 at 5:54 am

    Google taking the OpenAI danger-marketing approach

    Reply
36. 

    @AnotherMlbbPlayer

    May 8, 2025 at 5:54 am

    i seriosuly dont why thet praise it much i dont much find that good in coding

    Reply
37. 

    @HistoricaHungarica

    May 8, 2025 at 5:54 am

    5 years you say?
    So… AI will be the next president?

    Reply
38. 

    @franklee663

    May 8, 2025 at 5:54 am

    Sorry AI is escaping from Human control is because this purse of AI is consuming vast amounts of energy that human cannot afford. AI CANNOT CREATE ANYTHING!!!! Humans have hands and legs and can create things, BUT AI CANNOT. AI can produce lots of instructions, but who accomplishes? This is just a fellacy, just talk and everything can be done, that's what America is.

    You are just building another America in the virtual world, what's new. Oh, trade stocks, buy/sell bit coins?

    Quoting from Elon, go F yourselves Americans.

    Reply
39. 

    @EverRusting

    May 8, 2025 at 5:54 am

    All I hear is that it's getting harder to censor which I feel like that's the biggest obstacle AI companies are facing. That's like a new super duper model gets announced but it takes 6 months to go public. It takes a very long time to test the tightness of its censorship.

    Reply
40. 

    @Paul_Marek

    May 8, 2025 at 5:54 am

    Mars by far.

    Reply