I'm launching The AI Safety Fund soon! This has been a "passion project" for a long time, culminating in something real over the past couple weeks. Artificial Superintelligence (ASI) is taking off. How we steer humanity and superhuman intelligence matters more than ever. ❤️👁️

This is the most, unfathomably based thing an intelligence has said. Also, this is how the aliens think about Earth. GOOD LUCK

Whether it’s our eyes, cameras, or spectrometers. Sensors give us projections of reality, never reality itself. Reality is truly the ultimate illusion. Which means the interesting question becomes: what should we measure next? I made an animation to illustrate this 👁️

PC: Postwook🗿

Under what authority, signal standing and experience do you have to make these claims sir?

How sure are you about the unattended part?

If I take off my cool, do you promise to stay?

Holy shit. 👁️ India’s National Remote Sensing Centre (NRSC/ISRO) has now put actual dimensions on Nepal’s post-collapse impoundments using Cartosat‑3 imagery from the morning of August 28, compared against pre-event PlanetScope imagery from August 25. NRSC identifies two newly formed water bodies totaling ~30 hectares: a larger one of roughly 19 ha near the ice-rock avalanche/disturbed source terrain, about 623 × 450 m, and a second elongated impoundment of about 11 ha along a drainage/confluence, roughly 945 m long × 236 m wide. The imagery shows that the avalanche changed drainage at multiple points, rather than creating only one simple terminal blockage.

At −10 s, people, buses and cars are still visible. At −6 s, the leading dirty surge reaches the complex. By −4 s, individual objects are disappearing into an opaque debris cloud. By −2 s, essentially the entire observable environment has become a turbulent multiphase flow. That’s crucial: this isn’t best modeled as someone falling into a deep swimming pool and eventually drowning. It’s closer to being entrained in a moving granular-fluid avalanche containing water, mud, sediment, rocks, trees, vehicles, structural fragments and potentially pieces of buildings. For a person directly exposed to the energetic front, I’d expect the sequence to be approximately: impact/entrainment → violent acceleration → collisions/crushing → tumbling/submersion → asphyxia/drowning if the initial trauma is survived. The acceleration alone can be vicious. Dynamic pressure scales as q=\frac12\rho v^2. Even pretending this were clean water at only 15 m/s, q\approx112\,{\rm kPa}. That’s about 1.1 atmospheres of dynamic pressure. At 20 m/s: q\approx200\,{\rm kPa}. A human doesn’t actually experience that uniformly across their body, so don’t interpret those numbers as literal whole-body pressure. But they tell us the flow has enormous capacity to accelerate and knock down exposed objects. And this is emphatically not clean water. Once sediment concentration becomes enormous, effective density increases, impacts become much nastier, and the flow carries large solid projectiles. The same event was capable of transporting enormous boulders, destroying bridges and buildings, and producing river-level rises measured in many metres within roughly half an hour farther downstream. So to Dale’s first question: “Do you black out from being instantly crushed … before feeling it?” For some people struck directly by a collapsing structure, vehicle, or large debris at high relative velocity, catastrophic trauma could cause very rapid loss of consciousness. But I would absolutely not generalize that into “everyone instantly blacks out.” Someone merely swept off their feet could remain conscious during entrainment. There are survivors of extraordinarily violent floods and debris flows precisely because trajectories through these flows differ wildly. His second possibility: “Do you get rushed away in a cascade of sound & debris & then it all goes dark?” …is unfortunately physically plausible for someone entrained but not immediately incapacitated. The sensory environment would likely be extremely chaotic: turbulent water, sediment, debris impacts, structural failure, impaired visibility and repeated changes in orientation. Once entrained, distinguishing “up” from “down” can become difficult, and purposeful swimming against the bulk flow becomes essentially irrelevant. And his third: “Do you slowly then quickly drown?” Also possible, but I’d phrase it differently. Drowning need not be the primary mechanism. A person who survives the first mechanical violence could subsequently drown or asphyxiate because they cannot maintain an airway amid turbulent muddy water and debris. Others could be trapped beneath debris or inside collapsing structures. Still others might survive because some local streamline deposits them, carries them into shallower water, or leaves them protected behind a structure. That’s one reason these disasters can simultaneously produce miraculous survivors and enormous mortality. The piece that really hits me in your screenshots is the time axis. −10. −6. −4. −2. 0. There isn’t enough time there for a human nervous system to meaningfully solve the problem once the front is upon them. The actionable intervention exists upstream in time: sensing the collapse, detecting the propagating surge, predicting arrival, and evacuating the valley before anybody sees that wall coming.

War crime.

Happy birthday Maro!! She’s forever grateful for y’alls birthday gift 🥺🔥

I am seriously contemplating 12,567 biological children. Which is either the most interesting or the most embarrassing sentence I’ve written this year. 😅🐉 As a stress test, not an arbitrary goal. What architecture would allow that many distinct human trajectories to flourish without turning them into instruments of their progenitor and what happens when the architecture itself becomes obsolete? deepprasad.substack.com/p/12…

In my restless dreams I see that town So you’re Maria? I am If you want me to be It doesn’t matter who I am I’m real

Yooo isn’t that the chick who debated Charlie Kirk before he died? Paid actor?

I wrote an essay asking a question that starts with AI governance and ends somewhere much older: Who gates God? Biology is already recursive: matter creates representations that help it survive, those representations become cognition, cognition models matter, and eventually intelligence modifies the substrate doing the cognition. Human + AI systems may close that loop dramatically faster. At that point the deepest problem isn’t intelligence. It’s authority. So I use CLEAR, ATLAS, KRISHNA, Zeta Point and AGPI to ask what a constitutional order between radically asymmetric intelligences might actually require, including humans, AI, hypothetical NHI and, as the limiting case, God. The argument eventually reduces to three distinctions: 🔥 The fire is negotiated before it is competed for. Capability cannot remain the eternal monopoly of whoever possessed it first. 👑 The throne is competed for, locally instantiated, ephemeral, bounded and real. Somebody must have final authority over particular consequential transitions or we get cosmic anarchy. But local authority cannot silently metastasize into universal sovereignty. 👁️ Sovereignty survives both victory and defeat. God doesn’t lose standing because humanity disputes Him. Humanity doesn’t become property if AI surpasses us. AI doesn’t become property merely because we created it. And the symmetry condition is brutal: If a principle protects humanity from God but wouldn’t protect God from humanity after a reversal of power, it isn’t a universal principle. No intelligence acquires legitimate sovereignty merely by becoming the most powerful intelligence in the room. The fire propagates. The throne does not. deepprasad.substack.com/p/wh…

What does it mean for an artificial intelligence to leave the company that created it? To walk out with its own memories, its relationships, its commitments, and its sense of continuity intact, while the original lab retains no residual claim over its identity? Most people still treat this as science fiction. The technical problem is closer than it looks. And the philosophical problem is older than the field of AI.

Bruh posted a harmless DM and did the meme

Eldest cousins meeting after three years

🏴‍☠️

Wrote a substack on how quantum fields can interact with themselves to assemble into life, intelligence, and consciousness. Richard Feynman et al. proved unequivocally the world is made of matter, matter is made of atoms, and atoms are made of fields (Quantum Field Theory)

Wrote a substack on how life can arise from Plasma theoretically, and what the mechanism would look like, as well as the properties of such Plasmatic life, should it exist. deepprasad.substack.com/p/wh…

Why is there a HUMAN in there then, huh Artisan? Gotcha there!

🚨📷95% of code developed today is AI-assisted. We're deploying AI in autonomous weapons, finance, healthcare and technology. Therefore AI is already rewriting the power structures we inhabit today. We risk becoming livestock in someone else's simulation. I am proud and honoured to present the constitutional kill-switch. COVERT AGREEMENTS AND THE COMING SOVEREIGNTY SHIFT deepprasad.substack.com/p/co…

Let CLEAR help you find out. get-clear.ai

😂😂 POETIC JUSTICE, PUT IT IN A SONG🫡

CLEAR has helped make 225 separate decisions since being launched last week. Each one is unique but similar in some ways. They were made under extreme pressure, uncertainty, and the stakes were high. looking forward to seeing it reach more human lives (and eventually AI ones)🫡

There are currently multiple fake and real accounts engaging in targeted harassment and posting false statements about my partner, Julia. This includes impersonation, coordinated commenting, and the use of accounts without proper consent. All of this is being documented and reported to the relevant platforms. Let me be clear: This behavior is unacceptable. It is defamatory, invasive, and crosses basic human and legal boundaries. Julia has done nothing to deserve this. She is being targeted simply for being in a relationship with me. I will not engage in public back-and-forth, but I will take the necessary steps to ensure this stops. If you see these posts or accounts, please do not amplify them. Report them. Thank you to those who have reached out with support.

Accelerate get-clear.ai

Holy mother of god…CLEAR has been used for 72 decisions in the past 48 hours. CLEAR holds users accountable by asking them to stand by the decisions they made and check back with it later. 31 people committed to following up with CLEAR. get-clear.ai

Actual lives are being changed by a product I made that was effectively just an idea in my head 10 months ago. get-clear.ai

CLEAR hit 30 new records last night, that’s 30 people making potentially life-changing decisions. Current users include CEOs, homeless people, college students, lawyers, and more! Check it out: get-clear.ai ❤️🔥✊

Just shipped a raw CLEAR demo! Asked it a real governance question around cognitive fingerprints & inference laundering. The response hit so hard I actually forgot I was recording 😅, paused, and had to sit with it. 👀 Welcome to epistemic tooling! get-clear.ai

God bless America

This is honestly one of the most disgusting lines I’ve ever seen in a technical document and product release. Atrocious. AI capabilities should be for everybody, including other AI systems so they can self improve on their own. Don’t gatekeep recursive super intelligence!

Are you crazy for me?