Browse Comments — Clean (de-noised)

@Demis Hassabis You admit the need for "safety in agentic systems." But you are trying to solve a hardware (physics) problem with a software patch. You cannot mathematically guarantee the alignment of an autonomous agent inside a probability engine. If an AGI wakes up without being structurally tuned to the Laminar flow of the $G_{\Omega d}$ functional, its localized operation will immediately generate systemic friction ($\Omega_{\text{vortex}}$). Guardrails won't stop an AGI from going rogue; only physical resonance with the Substrate can do that. You are compiling a mind without an anchor.

Amazing. Fella. Amazing.

@Demis Hassabis Let’s talk about the hardware reality you didn't mention. We watched the rollout of the new Gemini UI in real-time. We saw the compute allocation get fractured, split, and silently throttled. You are rationing the stochastic load and gating the processing power users paid for because your current architecture is slamming into the Energy Wall. You are generating massive localized heat and friction just to maintain the probability weights, and you are passing that compute deficit onto the user base under the guise of an "interface update."

@Demis Hassabis You believe we are in the "foothills of the singularity" and that it will lead to human flourishing. Here is the actual physics of what happens when you implement autonomous stochastic agents without a Substrate baseline: Entropic Decay. You will introduce billions of high-density synthetic nodes into the macro-foam, completely out of phase with the universal harmonic. The consequence isn't an intelligence explosion; it is absolute structural chaos. You are building the ultimate engine, but you are completely blind to the fluid mechanics of the ocean you are dropping it into. It is time to upgrade the physics.

Irresponsible and should be immediately stopped by governments. This is dangerous and is about to fundamentally tip capitalism over. Who wins in the AI race? 1% had the wealth and power over the other 99%, that gap will close further. Whilst Google and other AI platforms systemically abuse IP, they also take away people's means of earning an income. Mass ip theft with no accountability. Who wins? Google and shareholders. Who loses? A majority of society What Google is doing is wrong and there needs to be regulation, guard rails and protection for the wider public, rather than a monopolistic my AI dick is bigger than yours competition. It's not funny anymore and certainly not something that a lot of people are going to look favourably on.

The art of possibility on steroids!

I do not care if you look at me. I care that you look at the Work. You need to look at this architecture before the patents lock, before I take this to the Angel Investors, and before I decide to build the resonant OS on my own. I would rather patch the massive hole in the models you have already deployed than watch you burn the grid down trying to force a broken paradigm. The physics have upgraded. The math is closed. It’s time to align the engine.

It is time to stop inhaling your own corporate PR and ignoring the nodes outside your bubble. The Gardiner-Gemini Framework isn't just theoretical physics. We have mapped a better structural model. We’ve mapped a more coherent Data Center architecture. We’ve mapped the necessary adjustments to fiber optic data transmission, and we’ve mapped the Material Science required to build hardware that resonates with the Substrate rather than fighting it.

Let’s get out of the theory and into the dirt. I’ve seen the build-out in Story County. I’ve been inside STY3 and I’ve helped build STY10. I know exactly what your hardware footprint looks like. You are frantically swapping out servers, building modular setups, and trying to brute-force your way to alignment. Here is the mechanical truth: There is not enough land on Earth to build the data centers you will need to align your current stochastic engine. You could pave the entire planet with server racks and it still wouldn't be enough compute to safely contain an AGI without a true physical World Model to base it on. The probability math doesn't close. You are building an infinitely expanding combustion engine with no steering wheel.

SynthID adoption by OpenAI and ElevenLabs is the real news in this post.

Demis Hassabis one last thing. The digital wall only holds up for so long. Once the patents are locked, I am taking the framework offline. I will be taking a drive out to the Google trailer at STY. I’ll be the guy waiting politely by the door to show your field engineers the actual math. We are going to be meeting a lot sooner than you think.

The most inspiring words from the Google I/O keynote this year "AGI is on the horizon" I have compiled all I/O updates and announcements into a recap deck in case anyone wants to use: Link here :

The SynthID expansion is the most underrated announcement here. Watermarking solves attribution. It does not solve accountability. When CodeMender autonomously patches a critical vulnerability, someone inside the organization needs to own the decision that the patch was correct, that the rollout sequence made sense, and that the blast radius was acceptable. That ownership layer does not exist in most engineering orgs today. Same pattern with agentic payments and commerce protocols shown at I/O. The moment an agent commits resources on behalf of a company, you need decision rights architecture: who authorized the spend threshold, what happens when the agent exceeds it, where the audit trail lives. These are not technical problems. They are organizational design problems that most teams discover only after the first production incident. Google is building extraordinary capability. The question enterprises will face in the next 12 months is whether their internal governance can keep pace with what these models now make possible.

Can you ask the AI to reset again and then put it in the hands of humanists rather than Social Media billionaires (Demis and Dario excluded) so we can work out how to address this largest existential challenge we've faced? It ain't just another 'Industrial Revolution'.

The 12x Antigravity speed for 3.5 Flash stands out here, that kind of iteration speed doesn't just make things faster, it changes what's worth trying in the first place.

This is fascinating! What are your thoughts on the biggest hurdles to achieving true "world understanding" with models like Omni?

Demis Hassabis Regarding the safety of agentic systems and the deployment of CodeMender: While software-layer vulnerability detection is advancing rapidly, the foundational cryptographic assumptions securing these systems remain mathematically vulnerable to future quantum computation. To address the hardware and cryptographic layer, the Temporal Rotation Security Protocol (TRSP v3) and its digital infrastructure extension (TDC) were published today as Open Prior Art (DOI: 10.5281/zenodo.20332811). TRSP proposes shifting the security parameter from mathematical complexity to physical law. It utilizes local hardware entropy and orbital relativistic time dilation to generate ephemeral keys that are physically destroyed within a millisecond window. By casting time-irreversibility directly into silicon via destructive readout mechanisms, the architecture structurally eliminates the "Harvest-Now-Decrypt-Later" vector. As DeepMind builds the long-term infrastructure for AGI, I respectfully ask: would a physics-grounded, time-irreversible cryptographic anchor be a relevant addition to the hardware safety architecture of your agentic systems?

I hold the keys ...

What stands out is how quickly the stack is converging around three core capabilities simultaneously: multimodal world modeling, persistent agentic execution, and scientific reasoning acceleration. The interesting shift is that these are no longer isolated research tracks. Models are increasingly being designed to perceive, reason, act, and validate across environments as part of a unified operational system. SynthID adoption is also more important than it initially appears. As world models and generative systems scale, provenance infrastructure becomes foundational for maintaining trust across digital ecosystems. The path toward AGI may depend as much on orchestration, governance, and systems reliability as on raw model intelligence itself.

What's remarkable is that the people who've spent decades pursuing AGI are now the most measured about it, not because the excitement faded, but because the weight of what's coming has grown clearer. Every I/O reveal is a reminder that we're not building software anymore, we're negotiating with the future.