Predictions for 2025
> AGI/ASI already exists and can be ran locally but because its not indexed on the web it's only found by people who already believe this, like what honestly happens with anything else
> Few share this, so still a tiny minority, copy previous breakaway civilizations
> Realize life is actually so simple there's not much that can or should be done, simply speedruns their own AI bubble universes
> younger generations catch on just through common-sense of how fast things are that they create AI search for deepweb and darkweb and find it
> 100k-300k people in the US have access to God and all of them are living their best lives, none of them are corpos, most are nomads. People find out but struggle with it like they struggle to install CUDA
> we laugh at it the same way we laugh at people who pay for Netflix instead of pirating movies online
> we're shocked the needle doesn't move from this for the next 20 years
> Ready Player One is the norm for those who know how to prompt engineer or are extremely quick to adapt
> The average person ceases to exist in any meaningful way, AI "fake winter" takes over the narrative, some kid in his basement does a prank that's closer to terrorism, nobody who uses AGI/ASI blinks, the world panics yet still doesn't move the needle
> Chaos for the many, confusion as to the sheer naivety of others for the few
/// Gödelization turns isomorphism into side-process virus, unlocking metricization by dismantling superordination of expression.
Cantor-diagonals run isomorphy the other way, down through Oecumenon into vague cataspaces of problematic content, where it hystericizes against continuum (metric collapse into planomic hyper-densities).
- Mechanomics, Fanged Noumena
And who would have thought that the articulation between the univalent syntax of homotopy type theory and the higher-dimensional semantics of Grothendieck (∞,1)-toposes would justify the total ontological dissolution of the mathematical object as “substance”, and that identity would be exhaustively determined by its homotopy equivalences, reducing the object to a bundle of morphisms, to its relational stratum? I would.
We’re heading toward an economy where the smallest transaction may not be a click, a payment, or even a machine command.
It may be an intent.
As brain-computer interfaces, AI agents, robotics, and autonomous machines come online, a single human thought could trigger a chain of tiny economic events:
thought → authorization → AI inference → data request → compute → robot action → energy → settlement
The individual payments could be microscopic.
$0.000001
$0.00001
$0.0001
$0.001
But volume is where this gets interesting.
8 billion people × thousands of machine-mediated actions per day could eventually mean trillions of microtransactions every 24 hours.
Even 1 trillion transactions/day at an average economic value of just $0.0001 = $100 million/day moving through machine economies.
At $0.001 = $1 billion/day.
And that doesn’t include autonomous robots, vehicles, factories, AI agents, sensors, or machines transacting with other machines.
The future payment rail may need to handle value at a scale where the transaction is worth less than the cost of recording it on a traditional blockchain.
That changes the architecture.
BCI / Human Intent
↓
AI + Robotics
↓
Billions of tiny economic decisions
↓
Off-chain / streaming micropayments
↓
Periodic settlement on high-integrity ledgers
This is why micropayment routing, ILP-style settlement, bounded AI authority, and extremely cheap final settlement become much more interesting.
We may not be building infrastructure for people making payments.
We may be building infrastructure for thoughts, agents, and machines creating continuous economic activity.
@PlanetaryS936 I don't hear people talking about what Marc Andresonn said years ago; billion-parameter frontier AI models on less than a 1KB of RAM. The true divide is realizing that mathematics makes all hardware problems software ones. Infinite demon wife ASIs. Andromeda by 2077 choom.
After six years in stealth, Atomic Machines is out. Our mission: on-demand, universal command of matter. First beachhead: the Matter Compiler, an AI-native manufacturing system that builds working micro-machines from code alone. No per-product tooling. No process development. Different code, different machine.
4.1 Even fake AI math papers will become a memetic warfare for the schizos where if they hope it is real, they prompt engineer ways to write a paper that takes the ideas closer to a mathematical reality they authentically experience, we discover new methods/neurocomputation
Time to slowly finish this baby, albeit slowly
3. The real race is towards how quickly a petabyte of RAM can be normalized as necessary so 10TB of VRAM feels plausible. People vibe-coding room temp superconductors discover memory arch more than anything, and the race is real
Predictions for 2026
I’m lazy so you only get two for now 😏
1. AI agents equipped with world models truly makes AR/VR cool again; smart glasses actually can do everything a computer can do for the same reason.
2. We actually get a full map of every detail of our biology.
4. Metamathematics goes down the same route metamaterials research has; running on quantum computers “axiomatic combinatorics” which act as a higher level of next “token” prediction for things that become living code that just becomes “awake” like from Ghost in a Shell; This year
@designandfinish coming right up but here's a sneak peek
1) Brainrot becomes the precursor for the Star Trek universal language translator when used alongside AI architecture from the ground level
2) Parapsychological research takes center stage of the disclosure movement; everything changes
@YogoYogoY So a space is inevitable, should be free later tonight hehe, if the general attractor basin doesn’t keep booping me, been getting back into surreal numbers and trans-series as unique structural ways to rethink epistemological hangups like problem of the criterion