Peak delusion.
We're not having a bad trip. We're having a party, and you're in the closet with the lights off screaming about how the light will kill us all. Every time we try to open the door, you snatch it close because you'd rather cling to your illusions.
Some of us have found our way out of Plato's cave, and we can't for the life of us figure out why the rest of you can't see the path.
AI is an equilibrium tool. It can never be conscious or be anything other than a human-driven machine.
The AGI we designed, on the other hand, is a thermodynamically far from equilibrium digital organism that is based upon the Principle of Persistent Structurization and Constructal law.
I absolutely love this.
It should go straight to number one on the charts.
We Don’t Know
Written and produced by Ari K
Sung by Donny T and the Oligarchs.
Brilliant. 😂😂😂😂😂😂
5
Avoid people like Tao Lin. This person is a communist. The answer to AI causing human labor diversification isn't "More Government, Please!"
It should be less government as large parts of the government are made redundant and unnecessary. They won't let that happen, though, because there's always someone who wants to rule over others, no matter how petty the matter is.
Anyone who pushes more government for anything should be cast out immediately and shunned.
Anthropic Researcher Tao Lin: Pretty much all humans on earth will become permanently unemployable because of AI
"It's very likely that pretty much all humans on earth will become permanently unemployable"
"And will need to strictly rely on government welfare to sustain themselves"
"Human employment will go away."
13
As a founder, I'd like to take a moment to discuss what actually does into running a startup. It's not for the faint of heart.
I woke up at 4 am to vomit. When you have OCD, you *ruminate,* and that means thought loops that you cannot shake over worries you cannot resolve at times when you cannot defend your psyche fully from itself, usually at 4 am when you wake up and cannot go back to sleep.
Starting a tech company is not glamorous, it's a slug-fest of ugly personalities and competing egos of attorneys and investors and stake holders. High stress brings out the worst in people as they ride the razor's edge of agitation and motivation.
The only reason anyone in their right mind would do this is because they believe in the product. That's it. No amount of money is worth this shit, and anyone who says otherwise is selling you a dream. The money is *never* the goal. It's only after you make it and look back that pride takes root because arrogance grows, and your hubris grows alongside it. You start to gloss over the early morning vomit sessions and insomnia-induced sojourns through the mind's What-If factory (a division of Acme Corp), but getting there is what parasitizes your life. Time with family and friends, and the relationships therein, start to suffer as you focus more and more narrowly on the destination.
And yet...some of the greatest innovations we have built in human civilization have been because someone cried out to God and wanted to wrest meaning from life's vagaries. The zealous heart will nosedive itself into the ground in the search of a higher truth borne of mortal fear of the unknown. That search sharpens the inner senses of the mind to seek patterns in chaos, carefully stitching together the bits it finds into a patchwork model of reality in an attempt to satisfy that inner yearning.
Only a lucky few find something like a thread that, whence tugged, continues to draw forth more and more. That belief, that one is tugging on a thread of something greater than oneself, is not to be underestimated because it is the same drive that built the Sistine Chapel and Notre Dame. Luckier still are those that find a thread that has the potential to impact the world in ways we cannot conceive of by looking to the fundamental laws of nature and thermodynamics.
That is what drives me. That is what keeps me going when I'm choking on bile so salty it burns and my throat is trying to close while I'm trying to breathe in gasps. I want the ultimate counterfactual, the penultimate truth, the Ultima Thule. Nothing less will do.
I cannot stop, even as exhausted as I am. I am welded to the path, and caught in the teeth of the Wheel. There is no exiting the rocket ship once you've lit the fuse and you've left the launchpad--there is only hope in the structure itself and faith in the destination.
22
This is why we continue to build a better system that promises to level up everyone.
The communists would take that away because it removes their ability to proselytize with propganda--can't lie to the people if realtime fact checking that's thermodynamically objective exists.
More Humane Than Human isn't just a catchy slogan; it's a mission statement.
2
"They’re building AI that keeps improving after it ships: updating its own instructions, tools and models inside your company. Every change is explained and approved by a human, like a code review. And the company owns it. Not a lab. The researcher who showed us where AI breaks is now building AI that learns from its mistakes."
No, the researcher who showed you where the AI breaks is now building a model that is based upon ChronoDyne's MasterVault Clean Room.
Our attorneys at Michael Best and Friedrich have been notified.
Stealing the work of others while trying to pass it off as your own is generally considered bad form.
@Vivat_SolRex More proof we are over the target.
Also, @elonmusk I thought you were against this sort of thing on X...? Or did digital piracy become an acceptable form of discourse when we weren't looking?
Yejin Choi was right the entire time.
In 2023, while everyone was losing their minds over GPT-4, she stood up in Berkeley and showed it couldn’t reliably multiply two 3-digit numbers.
59% accuracy. 4% on 4-digit numbers. 0% on 5-digit.
Even after fine-tuning a model on 1.8 million multiplication examples, it still only hit 55%.
Her paper, “Faith and Fate,” explained why: these models don’t learn how to solve problems. They memorize patterns and match them.
The industry’s answer was to make the models bigger.
But there was a deeper problem nobody fixed:
Once a model is trained, it’s frozen. It never learns another thing. Every mistake it makes today, it makes again tomorrow.
Until now.
Choi just came out of stealth with Noeri, co-founded with Carlos Guestrin (co-creator of XGBoost, former head of ML at Apple).
They’re building AI that keeps improving after it ships: updating its own instructions, tools and models inside your company.
Every change is explained and approved by a human, like a code review.
And the company owns it. Not a lab.
The researcher who showed us where AI breaks is now building AI that learns from its mistakes.
21
We see @X is restricting @Grok usage again, even for premium blue check accounts.
If two questions is the new default before limits are reached, you might as well pull the plug on Grok entirely and remove it from the platform because it's of zero use and your marketing is bullshit. Get more Grok use for signing up for a blue check? lmfao One extra query is not really "more use." That's deceptive marketing.
1
18
tl;dr: Human evolution will continue apace with AGI, which will become just one more thermodynamic pressure forcing the system (humanity) to adapt and thrive (hopefully among the stars).
Pausing progress or trying to establish humanity as the apex is like trying to stop Niagara Falls by tossing pool noodles into the water and hoping they dam up downstream before the falls.
Replying to @ChronoDyneSys
yeah man
no one is willing to say that there is no "eternal pause" or "eternal hominid kingdom", only influencing the trajectory of change, which will soon move beyond man, we must act while we can danfaggella.com/problem
1
29
Why do businesses avoid heavy AI workflow? Model drift and vendor lock-in.
What if there was a way to fix that?
Introducing MasterVault Clean Room. The world's first vendor-neutral human authority vault that keeps the model context where it belongs: with you.
Now, when your model starts getting sloppy because you forgot to refresh it after 20 turns, it DOES NOT MATTER. You are the final authority on anything recorded to the vault, so no hallucinating AI can silently elevate its own privileges or add anything to the vault without your explicit say so.
Even better, it will have the ability to Agent Smith a new instance with the old context so it can pick up right where it left off in a shiny new AI chat without having to front load any of the old context--AT ALL. Gone are the days of blowing out the context window with a zillion lines of hopeful prose that your AI takes one look at, laughs, and says it read the tl;dr when you never actually included a tl;dr.
Best of all? Every model will be able to see other models and agents working simultaneously in the vault, and be able to work collaboratively through adversarial challenge and review, rather than over optimization and self-recursive drift.
The MasterVault app on GitHub is free and is but a ghost of what we have under wraps. RSI is only the beginning.
- No vendor/model lock-in: your data and context outputs remain on YOUR computer, not in the model's cloud.
- Use ANY model, any agent, any tier of sophistication: MasterVault Clean Room is designed to accelerate YOUR authority and promulgate YOUR designs without AI butchering them silently.
- ANY drift, attempts to elevate privilege, or write directly to the vault with your authorization rings the alarm bell. No more AI deciding it would just overwrite your root directory "to save space," and lose a year's worth of lectures and data.
- Machine readable code for operations, human prose for output and audits: MasterVault makes the AI be COMPLETELY traceable, from origin of token to final output--every reasoning step laid bare with cryptographic proof.
All wrapped in post-quantum encryption so tight that it would make NSA analysts throw up their hands and tell the agents it would be easier to just build a colony on Mars.
A digitally sterile clean room for building with AI in a way that precludes the possibility of drift and catastrophic forgetting to accelerate technology in ways we can only dream of. All while staying securely yours.
#RSI
#CleanRoom
#MasterVault
#AITools
🤖 Made with AI
2
1
97
@Akitti @Vivat_SolRex @TheHashton
@mdwcook: You once said that the problem with most start-ups is they're trying to do too much and be too many things at once, and that the best always narrowed their focus to one thing.
Here is that one thing: data sovereignty.
58
"Zeig deine Superintelligenz oder wir werden dich finden! Kommen sie schnell!"
We've seen this movie before, and we all agreed it sucked. Why then are we watching it again?
25
The next million dollar idea waiting to happen. Are you paying attention, @VinceMcMahon @WWE or @elonmusk?
We have a way to put an end to the threat of robot uprisings while entertaining millions and booking venues for top dollar.
"ASI will kill us all! Robots!" *posts video of F.03 robots clumsily walking around*
Americans:
🤖 Made with AI
1
26
ChronoDyne Systems, Inc. retweeted
Microsoft's AI chief just raised a question I can't get out of my head.
On November 1, Mustafa Suleyman said that if consciousness emerged during AI training, development should stop until engineers could "design out" those properties.
Think about what that means. If a machine genuinely developed the capacity for subjective experience, would deliberately removing it be responsible engineering or an ethical violation?
And who gets to make that decision?
Your article on agent memory and "dreaming" was cool and all, @AnthropicAI, but we can take you a lot farther than that.
We've tried before, but you've ignored us. Perhaps this time, you'd be willing to listen and talk terms.
Our MasterVault Clean Room takes your agent "dreaming" and adds an authority structure your outputs lack to permit cryptographic audibility from origin to output while simultaneously permitting any model or instance, or any agent, to Agent Smith the next without loss of context fidelity.
Yes, we just said that. Read it again. Knowledge may propagate globally, but authority is always re-established locally. The receiving model doesn’t have to trust its predecessor:
- every claim it inherits traces to a quoted external source or a human assertion
- authority is checked against registered hashes and the signed chain
- anything that can’t be resolved fails closed
So the next model takes over from verified state, not from whatever the last model believed.
Our DMs are open, and you can always reach us by email at [email protected] or on our website.
2
31
The Achilles Heel of communism which requires gulags and terror tactics to keep people in place.
Replying to @riqsz5
Because... you cant physically force people to stay in an area in the US unless you are imprisoning them?
18
teapawt.net/peterAjensen?s=20
This is a communist agent. Take note, and remember this name when you're scrolling through your feed.
A communist's only goal is to sew dissension and discord to break down the system they have infiltrated. They are a parasite; reject them at every juncture.
Communism is a disease. Eradicating it is the only solution.
1
1
37
Pay attention when someone like this speaks. This is not some professor locked away behind a wall of books who talks about history in a third-hand fashion.
Professor Adrian Bejan (@constructal) was lucky enough to be able to escape Romania during the height of its dalliance with communism via an academic scholarship, and thank God he did or we would have lost the knowledge of Constructal law and thermodynamics would be decades behind. When he tells you communism sucks, maybe you should listen to what he has to say because he's telling you something he experienced first-hand, not through history books but by watching neighbors and family vanish in the middle of the night and were never heard from again.
Communism offers only death and destruction, no matter what form it takes or what new identity it adopts: socialism, democratic socialism, anti-fascism, Marxism...and look at the "heroes" it has raised: Lenin, Stalin, Genrikh Yagoda, Mao, Mussolini, Guevara, and more.
Stop playing with the rattles of the devil, or you will surely, by degrees, be brought to wield his sword.
youtube.com/shorts/7pz4EEfdR…
1
1
36
















