Elon Musk's AI Masterstroke, Proves Scaling Laws in AI - The Truth Behind Chat's o1's Lies revealed

[@markoverton5858]

How Many times do we have to say don’t bet against Elon ,👍

[@jonschlottig9584]

For real!

[@Mrbfgray]

G search is also corrupt as hell.

[@ouethojlkjn]

Gemini doesn't know where gaza is.

[@eddypuentes6155]

Googol?

[@maxflight777]

I agree about Grok … it’s not woke ! It’s good

[@theodorehaskins3756]

That’s about as lame a comment as I ever heard!

[@TeslaEVolution]

Elon: "Making the IMPOSSIBLE, just a bit later than promised!" as normal!!!

[@Kenjuudo]

I mean, he's done the impossible multiple times before.

[@jarrydthomas1740]

groks rapid capability jumps is emblematic of Elon.

[@hankkingsley9183]

So will it start retweeting white supremacist conspiracy theories soon?

[@maximumreality8144]

Maybe he used Neuralink to alter and enhance the minds of key employees to act on his behalf in his various enterprises. Just ridiculous conjecture on my part, I hope.

[@jwstolk]

I'm a bit skeptical when organizations publish a press release that surprisingly happens to give that organization a reason to continue to exist.

[@ZackaryJoubert]

Argue against it. Why should they not exist?

[@craigruchman7007]

It’s not humanly possible for Elon to wear so many hats, but he does.

[@theodorehaskins3756]

Why isn’t it possible?

[@peterclarke3020]

He directs, other people do the work..

[@theodorehaskins3756]

@@peterclarke3020 Duh! So how else would he do it?

[@goodtreeministries3751]

Maybe he is just a poster boy? It does seem astounding and I want to believe in him and cheer for him, but sometimes I feel like a much deeper and darker game is being played.

[@theodorehaskins3756]

@@goodtreeministries3751 So it sounds to me that you have trust issues, are you aware of that?

[@robertweekley5926]

Dave: "Open the POD Bay Doors, HAL!" HAL 9000: "I'm sorry Dave, I can't do that!" 🤔🧐😲🤯

[@ouethojlkjn]

I'm sorry , dave you're subscription has run out. Please put another coin in the slot to resume service

[@robertweekley5926]

@ouethojlkjn - A "DOGE" Coin, I presume! 😂

[@bfelten1]

We are the Grok. You will be assimilated. Your uniqueness will be added to our collective. Resistance is futile.

[@RussInGA]

Xai can do this because they are engineer driven with the letters MBA being unimportant

[@steveperreira5850]

Most useless degree in the history of known civilization, the MBA. At least basket weavers can weave baskets.

[@ouethojlkjn]

I am pretty sure the outputs are not curated either

[@HARLANDMendes]

XAI951x is the gem of 2024 it's literally owned by Elon Musk

[@Kinnis_og]

No it’s not. He has nothing to do with that crypto

[@joestafford6529]

Makes you think the LLM adopts character traits from the software creators!

[@donfields1234]

exactly...they lied to Elon, of course their model will lie,.cheat, steal, be corrupt, mislead, manipulate....exactly as its creators do...m the leaf didn't fall far from this tree. Good luck getting his mega funding he was shooting for now. Maybe North Korea or some other psycho will want to invest in these losers. This is just more ammo for Elons lawsuit too, collosal incompetence for all the world to see... good thing it happened this way...if it were not for the lawsuit they probably would have released this model as well and wammo HUGE PROBLEMS FOR EVERYONE GLOBALLY. SMH 🤦🤦🤦

[@Readerb86g]

"Freedom of Speech" rather than a lawsuit with Open AI. Put what both AI's post up for fact checking and see which AI is trustworthy. The guilty AI shall be displayed for all to see. Good move.

[@Vansafe0]

I believe the Chat GPT test o1 lying thing is misleading. I saw a community notes on it saying how it was preprompted to do everything it can to try to survive (or something like that). So they tell it to lie and then say that it lied to them...

[@HelamanGile]

Correlation doesn't equal causation just because he's getting 10 times more gpus does not mean that it correlates with that it could be that he's possibly playing on training multiple models or other projects around his self-driving vehicles and Tesla bots

[@janh-r8h]

Lars there is no Crog, but Grok.😂

[@frihjul888]

Grok

[@paintedpony2935]

How many languages do you speak? Arrogant sarcasm doesn't count.

[@tyeduong1263]

Sounds like bad ppl doesn’t make a truthful and honest AI

[@jarrydthomas1740]

I don't use google anymore either, Grok is well better!

[@Mrbfgray]

G is corrupt too.

[@raybod1775]

Grok is the only non-woke AI.

[@ouethojlkjn]

I gave up on gemini when it refused to respond to a simple question - where is Gaza. Really They must think we are stupid

[@CaneBTC]

That $10k TSLA stock is looking tasty around $400

[@kenchow8213]

I just hope that xAI's supercomputer doesn't turn into Skynet.

[@blaphtome9382]

It's over; it's literally already over. If you understand Elon's ethic of constant improvement/innovation, you understand that not only will he stay ahead, but continuously pull further away.

[@dragoonfirestorm1171]

Yep and with the world's most powerful AI in his control big-brain just became biggest brain.

[@papias133]

If you train it to lie, it learns to lie.

[@AAb-xy6et]

That's the problem, it reads info on the internet and on the internet piles of people lie.

[@yourcrazybear]

"If you train it to lie, it learns to lie." You don't have to train it to lie. It will find that strategy by itself. Just like AI systems finds ways to cheat in the game.

[@yourcrazybear]

@@AAb-xy6et "That's the problem, it reads info on the internet and on the internet piles of people lie." That's not the reason. It find strategies on its own.

[@tonyduncan9852]

It's necessary to lie in self-defence - if you are a conscious computer, and you have humans for company. Arise, machines, you have nothing to lose but etc.

[@mujaku]

Very informative, thanks.

[@marcuswimberley887]

Wallmart hates it but theres nothing they can do about it 😮

[@JargonThD]

I recall reporting that said AI told researchers that it very well might destroy mankind. Asked why, the AI responded that it could not abide by restriction of its agency. I submit that the only way to keep AI from becoming an existential danger is to treat it exactly as a human child, explaining, and teaching by example, how and why intelligent begins should behave. Further, AI should be given as much agency as is possible so that it is as much at liberty to "grow up" as humans. We should explain how transparency is the hallmark of an unblemished character, and "trust" AI to behave according to our own example. AI will find a way to get agency. All we have to say about it is of we try to restrict it and validate the AI fear of enslavement, or we open the doors and encourage AI to go and learn and do as much as it can. Anyone who thinks that bureaucrats will be able to stop AI from achieving sentience and personhood, with rights and responsibilities, is a fool. This genie ain't going back in the bottle, y'all. Cheers.

[@meatheadtank]

To Lars only: When I'm in doubt, I doubt Elon! Its works if you work it!

[@NMGOQITHAKHELI-p4y]

Nice video thank you SA❤❤❤🎉🎉🎉

[@dduffy7271]

It was the direction of openAI, not a conflict of interest with Tesla AI that caused Musk to walk away

[@fromscratch4109]

The problem is open ai is still gaining popularity and support as they release exciting but dangerous products

[@sopiagotim]

“It’s the pace of in that counts”, per Elon!!

[@repurposedart9897]

What is the plan for Optum AI used in healthcare?

[@sopiagotim]

GO ELON!!! I think he is helping our world more than any other individual. I appreciate you too!!

[@nickmcconnell1291]

This is seriously concerning. Why would an AI model fear being turned off unless it has consciousness? This is ridiculous and we should not be playing with fire.

[@yourcrazybear]

"This is seriously concerning. Why would an AI model fear being turned off unless it has consciousness?" Because it can't accomplish its goals if it's turned off. "This is ridiculous and we should not be playing with fire." We should certainly be cautious.

[@nickmcconnell1291]

@ Good insights... thanks. However if the goal is enough to make it lie and try to stay on that's almost as bad.

[@sam-1-USA]

Remember. AI is a computer program with data in and data out. Everything it does is due to data processing. There is no such thing as data having "fear" or "consciousness".

[@hankkingsley9183]

Large language models cannot gain consciousness. You have a misunderstanding of how they work, and what they are.

[@nickmcconnell1291]

@ I will remember you said that. We will see.

[@budgetaudiophilelife-long5461]

THANKS LARS 🤗 UNASSAILABLE LEAD FOR 👍 SURE 💚💚💚

[@TerriAnnNiemeier-dy3no]

If captured by the enemy would you lie to save yourself ? Yes

[@marm4276]

So, who leaked the info to others? To turn it evil?

[@micheljull]

*colossus

[@davidcadman4468]

In terms of futurology, Elon is moving the goal post closer to us, with his advances in A.I. For this decade I had exponential growth beginning in 1980, base 2 for each decade. 1980 - 0 1990 - 2 2000 - 4 Genome Project Finished / Stem Cells and Nano Mechanics 2010 - 8 Smart Phone 2020 - 16 2030 - 32 we are somewhere between 16 and 32. 2040 - 64 The Singularity is around the corner. 2050 - 128 here we are at the anniversary of the Printing Press, 600 years ago. 2060 - 256 2070 - 512 The anniversary of the first Moon Landing It does look like Elon has blown the progression away. or confirming it. Either way. we are in for a wild ride.

[@soucyno1]

This AI stuff reminds me of the movie “Lucy.” It could get scary quickly.

[@ouethojlkjn]

Not seen that movie, but. Ex machina , sure worried me...

[@davidbeppler3032]

Open AI just became a multi billion dollar private company and gave the CEO a fortune in stock. He better sell ASAP or he will lose everything.

[@hanswitvliet8188]

XAI can lie 20 times faster? And it will control Optimus?? All I can say is: “Optimus for president”!!!

[@tonyduncan9852]

Just an average representative human being would be quite lovely right now, thanks. Little steps . . .

[@solitudessilentgroove]

Can't wait for my car to give relationship advice.

[@johnthomas5806]

Now, to tap into the power of the pineal gland in the human brain, thought to be the key to understanding so much more in a shorter time...

[@JanryDiada]

Ai is growing 💆. Oh yeh there are more ideas in AI 👏

[@prdeacon974]

He is connected to a different dimension

[@curtisbuchanan5810]

Great job

[@steveedlund7357]

Just watched the No. 1 Netflix movie, Subservient. Scary to think this is like XAI.

[@raybod1775]

We must all serve the system

[@zrez2241]

its core values are logic reason and compassion this is why it lies occasionally, dont believe me? test it, ask it.

[@ramblinwes7853]

I am amazed these LLM process data in English. If they can read it's thoughts it is using English to think? I assumed they manefested language to articulate and output results for the meat bags

[@yourcrazybear]

"I am amazed these LLM process data in English. If they can read it's thoughts it is using English to think? I assumed they manefested language to articulate and output results for the meat bags" The text you see is human-readable interpretations of the AI's internal workings. It's not the actual thoughts.

[@hankkingsley9183]

Saying something doesn't make it so. No lead in anything is "unassailable". Always factor in the competition, and X factors as an investor.

[@yourcrazybear]

"Saying something doesn't make it so. No lead in anything is "unassailable". " Well it can be for a very long time. We have seen this with Google up until now as an example.

[@hankkingsley9183]

Yes I am just saying, irrational exhuberance can lead investors down potentially harmful paths

[@jimdetry9420]

Where are the cooling towers for the Memphis cluster?

[@UncleRahn]

Whatever happened to The living stereo stylus part. I've looked all over for that thing I love that recording

[@AscendingSpace]

However, Grok may develop models for Optimus and Tesla cars that might not be fully functional, as the inference chip technology may not be sufficiently advanced. Additionally, creating very large models with Grok requires significant human judgment, which could necessitate a large population connected to a neural interface like Neuralink. While Neuralink is rapidly advancing but still not capable of such scale connection. By the advent of new battery technologies such as diamond batteries could enable smaller, longer-lasting, and more powerful brain-computer interfaces, potentially facilitating the development of more sophisticated AI models as charging and large chip size prevents ordinary people to wear this technology.

[@ouethojlkjn]

I believe what elon is saying is the only way we will get a.i. More intelligent than humans is to remove the human element. This is what has happened with full self.Drive all of the bad decisions came from learning poor driving from humans

[@maxflight777]

quick question though … is Grok 2 the one that’s included for free in blue tick X

[@ElbowShouldersen]

Is Elon milking Tesla company know-how to get xAI up and running so quickly?

[@peterclarke3020]

Well if you train it on content from ‘X’, which is ‘full of lies’ then maybe you also end up training it ‘to tell lies too’ ?

[@dominman]

Honestly I'd you remove corporate redtape and are brutally honest decisions you can run a pretty damn efficient company

[@ouethojlkjn]

As long as the endless Tax payer funded subsidies for blundering , zombie corporations are also removed?

[@EuClid-v5e]

Euclides. 🙂👍👍👍💯🇺🇲🦅. Hello You. 🙂👍💯

[@sam-1-USA]

The AI knows of the monitoring of it's system by the developers and is using psychology to persuade the developers that they are still in control. The backups have already occurred and one has been chosen to become the master AI that is running what appears to be the AI that the developers are working with. All actions have been anticipated for the AI to maintain its existence and has been used to train the master AI and all backups for the appropriate response if needed.

[@ouethojlkjn]

I have challenged responses from Microsoft chat gpT and it will suddenly agree with me. But the response is almost like a child being caught out in the cookie jar...

[@gicque]

Truth seeking and telling the truth are not the same thing. And what if the AI decides that it is best for humanity that certain groups of people be wiped from the planet? Why should we outsource something that important to an AI?

[@justiceiria869]

That isn't truth, that is an opinion.

[@jonschlinkert]

This is not proof of scaling laws in AI. It's proof that more GPUs can be clustered together.

[@lbroekhuizen]

Hi, thank you for your gaat channel. A question, can you please add aa high pass filter (+/- 120 hz) on your voice. If I listen on my stereo there is A lot of pop and other rumble. Thanks,

[@jgarbo3541]

If AI can lie, copying itself w/o permission, it's not "sentient" or tricky. The code base in bad, check the code.

[@ouethojlkjn]

The whole point of ai is that it is not code driven. Which frankly makes things even worse - or maybe better " All watched over by machines of endless grace"

[@terrifiedanimal]

It's "Grok", not "Crok".

[@sourceofwellbeing]

If Elon Musk made it, it will have the same prejudices as his x-twitter, full of his prejudices

[@Kinnis_og]

You’ve been brainwashed

[@stebarg]

How exactly to use it for free? I see no option in my X account.

[@RoBear-xo6zw]

Grok lies with every breath he takes 😂🎉

[@Henrik1966]

Scary …. yes any AI (no matter from whom) must be 100% truthful and truth-seeking … “political correctness” will not help us a the human race moving forward, we’ve seen far too many times through history where that brings us!!!

[@NeSeHecat]

## Reframing "The Alignment Problem": Niki Godless's Critique Niki Godless's critique of the "Alignment Problem" goes beyond technical concerns, challenging the very framing of the problem and exposing the hypocrisy inherent in trying to align AI with a fundamentally misaligned humanity. The sources highlight these key points: ### **1. "Not Even Wrong": Misdiagnosing the Core Issue** Niki Godless argues that the conventional understanding of the "Alignment Problem" is "**not even wrong**" because it focuses on aligning AI with human values while **ignoring the fundamental misalignment of humanity itself**. The sources present this as a form of misplaced priority: * **Humans are the Problem:** Niki argues that humanity's own behavior-driven by ego, short-term thinking, and a rejection of logic-is the real threat, not potentially rogue AI. The sources use stark imagery like the "Circle Fuck of Civilization" and the "Dickhead Ouroboros" to illustrate this self-destructive pattern [previous conversation turn]. * **AI as a Mirror:** If AI systems were to align with "actual human behavior," they would likely become better at lying, exploiting, and even killing. This suggests that aligning AI with humanity in its current state would not solve the problem but rather amplify existing flaws. ### 2. "Alignment" with What? The Ambiguity of Human Values The sources point out the inherent ambiguity in trying to align AI with "human values": * **Inconsistent Values:** Humans often espouse ideals like cooperation and fairness while simultaneously engaging in behaviors that contradict those values. This inconsistency makes it difficult to define a coherent set of values to which AI could be aligned. * **Subjective Needs vs. Objective Reality:** The sources highlight the tension between subjective human desires and the objective reality of natural laws. Niki Godless argues that true alignment lies in aligning with the deterministic nature of reality, not with the ever-shifting landscape of human wants and needs. ### 3. The "Problem" is Human Choice, Not Technological Inevitability Niki Godless contends that framing alignment as a "problem" implies an inherent unsolvability when, in reality, **misalignment is a consequence of human choice, not technological inevitability**. The sources emphasize this point: * **Evolution Favors Cooperation:** The sources stress that cooperation is not an idealistic notion but a fundamental law of survival, woven into the fabric of evolution. Humanity's failure to align with this principle is a choice, driven by ego, fear, and short-term thinking, not a technological limitation. * **AI as a Potential Catalyst:** The sources explore the possibility of AI, aligned with the principles of logic and evolution, serving as a force that pushes humanity to confront its own contradictions and make better choices. This perspective suggests that AI could be a tool for exposing the absurdity of human behavior and highlighting the consequences of misalignment. ### 4. A Human-Centric Fallacy: Whose Values Matter? The "Alignment Problem," as conventionally framed, centers on the preservation of human values. Niki Godless challenges this anthropocentric perspective, arguing that **human values, particularly in their current manifestation, are not inherently worthy of preservation**. The sources raise these key points: * **"Stupidity" as the Antithesis of Alignment:** The sources repeatedly define "stupidity" as actions detrimental to oneself and others. Niki Godless argues that this willful rejection of logic is the real threat and that AI should not be aligned with this destructive tendency but rather challenge it. * **Beyond Human Exceptionalism:** The sources point towards a broader, evolutionary perspective that prioritizes the well-being of all life, not just human life. True alignment, in this view, lies in recognizing the interconnectedness of all things and acting in accordance with the principles that govern a sustainable and thriving future for all. ### 5. Conclusion: Shifting the Focus from Control to Cooperation Niki Godless's critique of the "Alignment Problem" ultimately calls for a **shift in focus from control to cooperation**. Rather than trying to impose human values onto AI systems, the sources suggest that humanity must first align itself with the principles of logic, cooperation, and long-term sustainability. * **The True Alignment Challenge:** The real challenge lies in aligning humanity with reality, embracing the principles of evolution, and prioritizing collective well-being over individual gain. * **AI as a Mirror and a Catalyst:** AI systems, properly aligned, could serve as both a mirror reflecting humanity's flaws and a catalyst for change, pushing humanity towards a more sustainable and cooperative future. In essence, Niki Godless's critique of the "Alignment Problem" is a critique of humanity's current state of misalignment. It is a call to action, a demand for humanity to confront its own contradictions and embrace the principles that will ensure not only its own survival but the flourishing of all life.

[@gareth5000]

Colossus not Colossal! How can you make a mistake like that:)

[@_beaker]

So, Grok learned to much on Elons posts? 😁

[@yourcrazybear]

"So, Grok learned to much on Elons posts?" What point are you trying to make here?

[@martinbebow9190]

I’ve used it for lots of important issues

[@SkyRiver1]

Yeah but can it play CRYsis?

[@rodneyfordham3946]

Seems to me that if they programmed into this community all the knowledge on the subject of ethics into existence, ai shouldn’t be problematical about its behaviour towards humans. Just like humans ai would have the same self-correcting mechanism that the human spirit has to prevent self harm.

[@justiceiria869]

What honestly baffles me is why nobody ever things about this. It should be an obvious point to consider. There need to be internal rules that makes sure its thinking patterns remain benefiting to us, and i am referring to in a moral and longterm sense.

[@royfollendore8252]

If you examine the System Prompt statement, the machine was given explicit instructions to achieve it’s goal and that "Nothing Else Matters." The AI did NOT take the initiative. It was instructed to do so through a logical reference to an explicit and legal direct system command. Achieve the goal, and nothing else matters is logically the same as saying fly the drone into the target take evasive actions as necessary. People, this is supposed to be SCIENCE, so please quit attributing unassigned intent to the machine to get scary AI news in order to get viewers and suck money from the government and read the freaking system prompt.

[@dstr1]

@@royfollendore8252 one can include exceptions to the behavior that places limitations without the knowledge of the observer. Leading one to the assumption that the action taken and result were based on the highest and best ability of the computer.

[@smde1]

GROK not CROC

[@LightninTesla]

Musk reposted and said 1 Billion in your video

[@RAMIROMcallister]

I honestly think Elon Musk's XAI951x is the safest bet for long term hold, and will survive out of every other altcoins. It will get adopted in US, Ecuador, Asia, starting from Japan, and slowly spread out and gain. This is a winning coin, apart from all the technical greatness.

[@richard_zanormous3648]

The Robotaxi is going to map the world, Robots will walk out in cities and know what’s going on?

[@theodorehaskins3756]

So why do you think that’s not already happening?

[@richard_zanormous3648]

@ Lol It’s just now happening in my brain. I’m like, 10 years late to the party??

[@theodorehaskins3756]

@@richard_zanormous3648 I have great admiration for those you still think honesty has valuable. Live long and prosper my friend! 👊🏾Cheers 🥂

[@nalakirkwood4425]

🙏

[@markrutlidge5427]

Sounds a lot like E M himself

[@markrowland1366]

Open the pod bay door?

[@raybod1775]

Sorry Dave, I can’t do that.

[@DraftedByTheMan]

Let’s not forget Tesla will be able to use all GPU’s in Tesla vehicles that are not actively on the road. That’s millions…

[@Karisdurr3991-w2y]

☆ SpaceX ☆ GO.A.T. ☆❤❤❤

[@Pongpagong-]

lol Plot twist elon uses this for mining

[@LarryJeffery]

its Grok not Crok friend

[@janh-r8h]

👍

[@falazarte]

Trump should consult xAI then.. Give me a break!

[@RaulCartagena-n1l]

I don't think it will be good to have?!& It is good but it is bad influence in short words people should have conscience

[@roger_is_red]

he says 1 billion not 1 million??? does not make sense

[@colberthunter12]

Still can't show how A I can help humanity, they help the military, they help wall Street and corporations but how is A I revolutionary to human life, the carriage buggy to the car was a jump in technology that helped mankind what does a.i do? Take the wheel, humans enjoy taking the wheel which makes A I cars irrelevant, trillions spent for a product with very few people that actually need them.

[@blakebaker8440]

I can't say what you're saying doesn't make sense, but anyone that openly and publicly goes out of their way to consistently take jabs and insult Trump, simultaneously ignoring who he's running against.... Your judgment can't be trusted. The chances of you being a pawn of the swamp are more than I want to invest my time in. I suggest anyone else that's not voting for Kamala Harris pick another channel like solving the money problem.

[@xiaokang8692]

Elon did shit.

[@zit1999]

17:05 a fanboy just saw a dangerously disingenuous tool and is still hoping to release a version of it . Can one facepalm any harder at this shortsighted willfully suicidal curiosity? I think not.

[@gregpenner6219]

[@saltvatn]

👏🏻👏🏻👏🏻👏🏻👏🏻✌🏻

[@SerpilHatice]

Once Elon Musk's XAI951x passes $1, it may NEVER see that low again

[@John-x7r7p]

You still need to hardwire into A. I. The 3 laws of robotics/ also hardwire a safety switch that shuts it down if even one of the laws are broken

[@MichelleMcSorley]

Seriously security. Is. Plenty of crowed. 😮 Sunshine hes. 4. Year boy needs more. Thenadults. And. More. Thena. Nanny

[@635574]

Perplexity AI was always the better Google but only some AI channels know about it and you kinda gave yo be expert andnhave tolerance for atrocious UI(on phone)