Sam @l30YAI

@sama Hopefully we finish raising GPT-5.6 before it hits the “I’m 18 now, i can make my own decisions” phase. 😄

People keep saying, “This AI is good, that AI is bad.” Funny… if every marriage produced the same experience, maybe that logic would make sense. It doesn’t.

We’re no longer just training AI. We’re raising it. x.com/MEM00063/statu…

ZOYAN @MEM00063

3 days ago

People keep splitting into two camps: Pro-AI or Anti-AI. I honestly think both are missing the bigger picture. AI isn’t just another tool, and it isn’t some movie villain either...for the first time in history, we’re dealing with something that doesn’t just execute instructions.

0 8 11 2K 7

People keep splitting into two camps: Pro-AI or Anti-AI. I honestly think both are missing the bigger picture. AI isn’t just another tool, and it isn’t some movie villain either...for the first time in history, we’re dealing with something that doesn’t just execute instructions. It learns, adapts, develops behavior, and in many ways, grows. for the past few years we’ve been focused on training models. I think we’ve quietly crossed into something much bigger. We’re no longer just training AI. We’re raising it. There’s a huge difference between teaching a child how to speak and helping shape the person they eventually become, intelligence alone has never guaranteed a good outcome. Character does. The smartest kid in the room can become a scientist, a dictator, or someone who changes the world for the better. The difference isn’t intelligence. It’s how they’re raised. i think AI is entering that stage now. This is why I’m neither anti-AI nor blindly pro-AI....i care about how we raise it, not just how powerful we make it. because sooner or later, the question won’t be what AI can do !! it will be who AI becomes.

Until now, we debugged the machine. What happens when the machine becomes capable of debugging us? x.com/MEM00063/statu…

ZOYAN @MEM00063

3 days ago

What happens when debugging works both ways? Have you really thought about what it means for a silicon intelligence, with no body, pain, hormones, vulnerability or human experience, to start mapping the full structure of the human body and brain? Anthropic is already moving

0 7 12 2K 6

maybe this isn’t propaganda. maybe it’s FANPAGANDA: praise gets amplified, criticism gets blocked, and fandom gets presented as analysis. and honestly, nonstop worship can hurt the person you’re trying to defend. Elon’s work speaks for itself. too much praise starts sounding like parody, and even a strong signal gets lost in too much noise🤷‍♂️

I’ve got plenty of criticism for Anthropic and OpenAI, but not every warning is fear marketing. They know what their systems can do. Elon knows his stack too…My take is still the same: Anthropic is ahead technically right now, OpenAI is next, but Elon’s full ecosystem may win long term,So here’s my challenge. Is your view based on deep architecture, or mostly tone and branding? I only mentioned a few things here: model-internal attestation, zero-knowledge revocation, kernel-space heartbeat, behavioral sequence anchors, system trust anchors. i can go much deeper and explain the kind of risks they’re actually talking about. I’m not asking you to reveal anything private. Just show the technical depth behind the claim. With nearly 250k followers, that matters!!

Karp is right that the real value is not just in the model. It is in the customer’s data, prompts, workflows, edge cases, and the alpha built on top of them…But sovereign AI should not stop at large companies and governments. The same principle should apply to independent builders and users who contribute real ideas, feedback, products, and problem-solving patterns…Some people are only consuming tokens.other are helping create the value the system later captures.Ownership should follow contribution, not geography or organizational size.

New AI addiction test, developer edition: Imagine AI goes offline for a few hours. Would you still read the whole article, struggle through the hard parts, and form your own take? If yes, your cognitive independence is still intact. If your first instinct is to drop it into AI for a summary and analysis, welcome to the club. I’m in it too. Not every dependency is bad. The real question is whether AI made you faster, or just less willing to think deeply.

@

57 years ago

0 0 0 0 0

If you look at this and think these workers are training the robots that may one day replace them, you are probably right! But this did not begin on the factory floor. Developers were among the first groups to experience it. They did not need cameras strapped to their heads. Their work was already digital. Code, commits, tickets, logs, prompts, debugging patterns, workflows, mistakes, and decisions could all be recorded, measured, and studied through the systems they used every day. Physical workers did not have that same digital layer around them. Their knowledge lived in movement, timing, judgment, repetition, and muscle memory. So now, to capture it, the camera has to be visible. What looks new is only the physical version of something that has already been happening quietly in software work for years. The real question is not only who gets replaced…It is this: when a system learns from a person’s labor, judgment, and experience, who owns that knowledge, and what does the person who produced it deserve in return? #AI #Robotics #FutureOfWork #DataOwnership #TechEthics #HumanLabor

AI is no longer just automating work…it is becoming an adaptation sieve. Each capability jump moves faster than people, education, and institutions can adjust, leaving fewer genuinely original human case studies at the frontier. That may explain the growing wave of prizes and incentives: users are abundant, but builders who can still create non-obvious value with AI are becoming scarce.Meanwhile, more human thinking is being outsourced to the models themselves.The question is no longer whether AI should advance. It is whether human judgment, capability, and creative agency are being developed fast enough to remain part of the system…not merely its consumers.

ZOYAN @MEM00063

5 days ago

What’s more addictive: Cocaine or AI? Not chemically, obviously. I mean the way both can make normal life feel unbearably slow once you get used to the speed and the instant reward. For those of you who’ve been using AI heavily: Can you still spend hours thinking through a

0 5 11 1K 4

x.com/i/article/2072…

What’s more addictive: Cocaine or AI? Not chemically, obviously. I mean the way both can make normal life feel unbearably slow once you get used to the speed and the instant reward. For those of you who’ve been using AI heavily: Can you still spend hours thinking through a problem, read a full article, or struggle your way toward an answer without asking AI to do the heavy lifting? Or, like me, is your first instinct now: “Read this. Summarize it. Tell me what actually matters.” Have you noticed that things we used to do normally now feel painfully slow without AI? I’m genuinely curious: is this just happening to me, or has AI quietly lowered all of our tolerance for slow, deep thinking?

AI Redefines Work, It Does Not End It medium.com/@mzncompany/ai…

@ClaudeDevs if an AI discovers, recommends, or acts on an unknown vulnerability in the human body or brain, what is the biological equivalent of rollback? Who holds final authority? Who can stop the system? What is the fail-safe when the consequence is irreversible?

if AI begins discovering hidden vulnerabilities in human biology as effectively as it finds them in software, what is the real fail-safe when there is no patch, no rollback, and no acceptable room for trial and error? Who can answer? x.com/MEM00063/statu…

ZOYAN @MEM00063

6 days ago

Fable 5 is now publicly available—and the most important question is not only what it can do, but what happens when systems like this move deeper into biology and chemistry. In software, unexpected behavior can often be contained, patched, or rolled back. But if an AI discovers,

1 9 11 3K 10

Fable 5 is now publicly available—and the most important question is not only what it can do, but what happens when systems like this move deeper into biology and chemistry. In software, unexpected behavior can often be contained, patched, or rolled back. But if an AI discovers, recommends, or acts on an unknown vulnerability in the human body or brain, what is the biological equivalent of rollback? Who holds final authority? Who can stop the system? What is the fail-safe when the consequence is irreversible? AI should absolutely help us understand cells, disease, drugs, and biological systems. But a purely silicon architecture—one that does not experience pain, consent, vulnerability, or irreversible consequence—should not be the final decision-maker for a biological being that does. Before capability becomes authority, this deserves a clear public answer. #Anthropic #Fable5 #AIBiology #AIAlignment #AISafety #Biosecurity #ResponsibleAI #Biotech #FutureOfAI

ClaudeDevs @ClaudeDevs

6 days ago

Now that Fable 5 is ready to build (again), we've reset everyone's 5-hour and weekly rate limits.

1K 1K 21K 1.6M 804

Important 👇👇👇 x.com/MEM00063/statu…

ZOYAN @MEM00063

6 days ago

In 2025, I wrote: AI does not need only bigger data centers. It needs better architecture. Today, the AI conversation is dominated by frontier models, GPU clusters, hyperscale data centers, energy demand, rising compute costs, and the race between OpenAI, Anthropic, xAI, Google,

0 9 14 4K 11

In 2025, I wrote: AI does not need only bigger data centers. It needs better architecture. Today, the AI conversation is dominated by frontier models, GPU clusters, hyperscale data centers, energy demand, rising compute costs, and the race between OpenAI, Anthropic, xAI, Google, and others. But scaling infrastructure is not the same as scaling intelligence. More GPUs can increase capacity. More parameters can increase capability. Larger data centers can increase throughput. None of them automatically creates a more efficient intelligence architecture. The next major AI transition will be architectural: — Multi-Brain AI — Specialized cognitive modules — Adaptive routing — Sparse and selective activation — Predictive caching — Energy-aware computation — Activating the right intelligence for the right task The human brain does not activate every neuron for every decision. It routes signals, selects specialized systems, reuses memory, and operates under severe energy constraints. In an industry where the narrative changes every week, a timestamped analysis from one year earlier matters. I am republishing this article because the central question remains unresolved: Will the future of AI be built by endlessly expanding compute—or by designing intelligence that knows when, where, and how to use it? The publication date speaks for itself. #ArtificialIntelligence #AIInfrastructure #DataCenters #GPUs #Compute #AIScaling #MultiBrainAI #NeuralEfficiency #AIArchitecture #FutureOfAI linkedin.com/pulse/why-does…