Shiv Goswami

Founder, @Terraflock | Astrophysics & Neuroscience

Life is a blank canvas so create a meaningful art

[Shiv Goswami @ArtistShivG]

Entropis Technical Documentation — Secured

4 documents. 8 SHA-256 hashes.
Complete technical record of a synthetic brain that demonstrates memory, association, and prediction — without training.

32/32 benchmark markers validated.
Hardware invariant (NVIDIA CUDA + Apple Metal).

SHA-256 Hashes:
08: 003AB7F3BE9E5BC42C098C91461A7EFFEAA489328CAD9E04FFF3B753702E9A8B
09: D6AC868D42D8436D0C5B34105D97DFB1F238A8184A6E1DE5E1B8BA943A3D7F66
10: A9A2E36E4EAD74FFCD4D495299BEB853DC20F5D0E0F58B4717F8A587D9F56887
11: AC2427E8C2642FBA04D34D51F83BC489AFAAB416E44D545686004206E0223476

Full record: https://t.co/cDRK1rv8uz

[Shiv Goswami @ArtistShivG]

A theory shouldn't just fit existing data, it should predict unseen observations.

I fit my cascade model to one DESI dataset and tested it on withheld data.

Blind test (3 withheld DESI Y1 points):
• Cascade: χ² = 4.44
• ΛCDM: χ² = 5.54
• CPL: χ² = 22.81
CPL later attains a lower χ² on DR2 only after crossing into the phantom regime (w<−1w<-1w<−1). In the cascade framework, a no-phantom theorem follows from the underlying discrete structural axioms.
DESI DR3 predictions are already archived before release. When DR3 arrives in 2027, the data will decide whether this framework generalizes or fails.

Preprint: https://t.co/nLttIQVHLx
Website: https://t.co/suKXRIXjUz

[Shiv Goswami @ArtistShivG]

Cross-release test: trained on DESI Year 1 (2024), predicted DESI DR2
(2025) blind.

Cascade: 4.45
LCDM: 4.39
CPL: 2.17 -- only via a phantom crossing at z = 0.345, a Null Energy
Condition violation with no known physical mechanism.

A lower chi-squared by absorbing noise through unphysical dark energy
behavior is not a win.

[Shiv Goswami @ArtistShivG]

I made a sealed prediction for DESI DR3 (expected 2027) before the data exists.

A discrete cascade model derives the dark energy equation of state from
first principles -- no cosmological postulates. Calibrated on DESI DR2
BAO (March 2025). Six D_H/r_d values locked. Committed to GitHub on
21 June 2026.

If DR3 confirms w > -1 at all redshifts and the phantom dark energy
preference weakens, the no-phantom theorem is vindicated.

[Shiv Goswami @ArtistShivG]

The no-phantom theorem is not an assumption.

It is proved from energy monotonicity in a discrete stochastic cascade
model on finite graphs -- the same mechanism that makes phantom dark
energy (w < -1) structurally impossible.

Blind retrodiction on DESI Year 1 BAO:
Cascade chi-squared: 4.44
CPL chi-squared: 22.81 (twice the free parameters)

CPL dipped into phantom territory during calibration and collapsed on
withheld data.

[Shiv Goswami @ArtistShivG]

Three preprints, permanent DOIs:

Part I (foundations): https://t.co/0P5cNem7Vv
Part II (necessity proofs): https://t.co/eMzq2XJ7TN
Dark energy equation of state + DR3 prediction: https://t.co/rOGWtwRlaL

Code and sealed prediction: https://t.co/euCuFYdzNx
Research: https://t.co/TIES6aApBw

The test comes in 2027.

[Shiv Goswami @ArtistShivG]

Exploration.

It’s the core of human civilization.

We optimize. We develop.
But what truly moves us forward?

Exploration.

Without it, we stay trapped in a frame, refining what exists, unaware of what’s beyond.

Exploration creates new frames.
Across domains. Across time.

There are no boundaries.
There never were.

Today, we return to deep space.
The real leap is not comfort.

It’s the courage to step into the unknown.

#Exploration #Artemis #Space

[Shiv Goswami @ArtistShivG]

Most software is written as instructions.

This isn’t.

Demonstration 2 of Ntrop.

A small Newtonian system written in plain English.
Compiled directly to native x86-64.

No traditional control flow.
Just constraints.

Ntrop is a distilled execution layer of the Entropis system.

Built and conceived by Terraflock.
https://t.co/3LcOXYOgrv
https://t.co/3WJUsSj3yW
https://t.co/VmK9MGu6IO

#Compilers #DeepTech #ComputerScience

[Shiv Goswami @ArtistShivG]

We made English a compiled language. No LLM. No interpreter. The constraints resolve. The machine code emits.

This is what it looks like:
Reactor temperature = 150.
0 <= Reactor temperature <= 1000.
Reactor temperature > 800 -> Reactor becomes scram.
That's not pseudocode. It compiles to x86-64 assembly.

This demo: a nuclear reactor safety system. 5 entities, 11 competing forces, 4 state transitions. Fission heats the core. Coolant fights back. The constraint resolver decides. Not if-statements. Math.

Ntrop distills components proven at scale in Entropis, our core work in the field of efficient synthetic intelligence.
For 70 years, humans wrote procedural code for machines. We asked: what if you just described what must be true?
Built by TerraFlock.
https://t.co/3LcOXYOgrv
https://t.co/3WJUsSj3yW
@terraflock
@Entropis_

What the demo output shows:
The simulation runs 1000 timesteps. Each step prints 11 properties across 5 entities. Here's what you'll see:
startup (steps 0 to 175) Reactor temperature climbs from 150. Pressure builds. Fuel depletes. All systems nominal. The constraint resolver enforces every boundary every step.
nominal (steps 176 to 470) Temperature crosses 300. The state machine transitions automatically. No code triggered it, the threshold constraint resolved. Turbine spins up. Vibration accumulates.
warning (steps 471 to 764) Temperature crosses 550. Emergency cooling activates. Coolant flow starts increasing by 15/sec, but only because the context constraint "While the Reactor is warning" became true. Containment integrity starts dropping. Radiation climbs.
scram (steps 765 to 1000) Temperature crosses 800. Control rods slam in at 25/sec insertion rate. Coolant maxes out at 200 (clamped by constraint). Containment integrity hits 0. Radiation saturates at 100. Pressure hits ceiling at 300. The reactor is dying, but every value stays within its declared bounds.
The key proof: every transition, every emergency response, every cascading failure was described in English and enforced by the constraint resolver. Zero control flow. Zero if-statements. The math decided everything.

#ai
#india

[Shiv Goswami @ArtistShivG]

Entropis Doc 12 | Jan 17, 2026
2B neurons • 81/81 • 3,334×

MD SHA-256:
8EE71E7B07FB3E4EC65B3ED078500F4DC160C177923679B1495D74CAF4A7F6DC

PDF SHA-256:
659533C6B3DEE9B2C8A86F10570076B96E58282F0EC8572F8C81D10C20F64C23

https://t.co/tMbgP5lTfE

[Shiv Goswami @ArtistShivG]

@fransvanberckel @filos @r0ktech You didn't get the joke. Did you?

[Shiv Goswami @ArtistShivG]

@karthikponna19 Add an error function so the build fails

[Shiv Goswami @ArtistShivG]

@rasmalai Rust

[Shiv Goswami @ArtistShivG]

@elonmusk He knows his audience is gooner,

[Shiv Goswami @ArtistShivG]

@Rainmaker1973 Prince of Persia warrior within

[Shiv Goswami @ArtistShivG]

@Rainmaker1973 GTA 3

[Shiv Goswami @ArtistShivG]

Artwork Title: Beyond The Horizon

In the stillness of the self, the mirrored abyss reflects not just the outer world, but the boundless potential of the spirit to transcend its own shadows.

#InnerWisdom #artwork #philosophy