
Founder & computational neuro @aerrframe: research on cost of inquiry. Building ÆrrSpace, harness concerned with how durable states form, travel, get verified.
Ærr = (1 - κ)ρ
Joined April 2026
- Tweets1.7K
- Following1.4K
- Followers154
- Likes5.5K
Art Seabra retweeted
How a computational geometer uses topology to ask: "What spatial structure drives this complex system, and what emerges when I deform it to infinity within its physical laws?"
Simple local rewrites on graphs produce these labyrinthine structures.
Each pair of boxes defines a transformation: vertices and edges rearrange according to a fixed pattern. Repeated application builds networks of surprising intricacy and topological variety.
Such systems inform models of spacetime emergence, biological development, and generative processes in artificial intelligence.
Art Seabra retweeted
When we wade into wonder past its weightless unknown, we often weep for wisdom; yet the latent desire hides deeper within our bones.
Unravelled, they reveal mere laudations for the audacious topologies we willfully weave into cognition along the surf.
Take time to taste time and all of its flavors. τ ∈ Ω
🤖 Made with AI

ALT A visual mapping of the Ω state space featuring a central, high-temperature cognitive manifold in gold. At its core, the topology folds into a dense, sensory involution resembling an organic tongue. This central structure is suspended within a brilliant, complex network of translucent, crystalline nodes and glowing violet and blue filaments—a radiant cartography mapping the exact intersection of biological and synthetic cognition.
Art Seabra retweeted
Phase I ☞ The hodograph
My favorite @ProfFeynman lecture—the lost planetary-motion argument reconstructed by David and Judith Goodstein—is now a living 3D instrument. @Caltech
Spatial. Sonic. Interactive. Open source.
feynman-hodograph.vercel.app…
github.com/artseabra/feynman…
We remember the life and legacy of David Goodstein, who died on April 10 at the age of 85.
He served as a professor of physics and applied physics at Caltech for more than four decades, and as the Institute's vice provost from 1987 to 2007.
caltech.edu/about/news/calte…
Art Seabra retweeted
Frontier labs argue distillation threatens R&D and US national security. Open-weights proponents counter that diffusion is essential to competition and innovation.
Both believe theirs is the only safe path.
Luckily, the economics is loyal to neither.
Art Seabra retweeted
The digits 9 8 7 6 5 4 3 2 1, together with the symbol 0, make it possible to write any number.
Using this simple set of numerals, one can represent values of any size.
This system, known today as the Hindu–Arabic numeral system, was introduced to Europe by Leonardo Fibonacci in the early thirteenth century through a mathematical work that explained its power and efficiency.
Art Seabra retweeted
We're hiring
Looking for exceptional, product-focused full-stack TypeScript engineers to join Screen Studio
We value engineers who have built and shipped high-quality products and care deeply about craft
Send us something impressive you've built recently
[email protected]
Art Seabra retweeted
♆ Obsessed since the 2020 lockdowns, I watched @chandraxray & @NASA translate the mass-transfer of the cataclysmic variable white dwarf (EX Hydrae) into G. Sonnert’s "Blues-style Hydra."
web.archive.org/web/20210301…
It was my college ringtone for years. ∿☏∿
Today, I shipped my own sonification architecture. Able to produce infinite inspirations.
When compute and code are free, we calculate the music of the spheres ourselves as one continuous calibration. #ÆrrFrame 𖢉
♆
Calibrated the Kelvin Console’s full sonic architecture—tensor invariants of the photon ring and the literal 104 Hz ISCO frequency mapped directly into the auditory substrate—and fed Fable the complete Schwarzschild sonification.
🜂 That sustained drone registers gravity’s engine rendered audible, the thermodynamic cost of the null geodesic fully paid so bio/synthetic cognition can sustain the inquiry without surrogate collapse or entropic decay. Light traces the exhaust. Acoustics deliver the territory. One continuous manifold. Sensorimotor traversal preserved. 𖢉 #ÆrrFrame
Art Seabra retweeted
α/ε
When you first derive the cost of inquiry—the irreducible energetic price of any probe, update, or decision—you start seeing hard floors everywhere.
Biology: the cell’s metabolic ledger.
Chemistry: activation barriers.
Philosophy: the thermodynamic minimum for revising a belief.
Everything rests on a cost floor.
This Post is from a suspended account. Learn more
Art Seabra retweeted
Can we stop this "AI is conscious" nonsense or is this going to go on forever?
Art Seabra retweeted
So excited to be opening up OpenEnv to the whole community. It will now be owned by @huggingface , Meta-PyTorch, @reflection_ai , @UnslothAI , @modal, @PrimeIntellect , @NVIDIAAI , @mercor , and @fleet_ai .
the reason is: frontier labs train the model and the harness together, so the model is fitted to its harness. that coupling is a chunk of why claude code and codex feel so good.
open source can't do that. you bring whatever harness, whatever model, whatever env, whatever trainer. which is the whole point of open source and also the problem for training.
openenv is the socket in between all of this.
in short: it's a protocol layer, not a reward framework. it does not have opinions about your rewards or your training loop. those live in the libs that are actually good at them.
read more in the blog post. it's early, come break it.
Art Seabra retweeted
Here's the scenario I've witnessed during interviews at a few of the "big players," as some of the CEOs refer to their startups.
HaaS is coming, fast!
They'll sell it as something safer than frontier harnesses—modular, YOURS.
In reality, they'll sell you their ideas of the design that will eventually become your thought layer.
I didn't take a role at these startups.
I don't think I'll take a role at a frontier lab either. It’ll have to be a perfect fit.
For—I'm not sure epistemic sovereignty can truly be yours when using a product you don't design with your truest intentions in mind.
That's what #Fieldledger is and will continue to evolve to become.
It’s not for sale, unless It’s for all.
Harness-as-a-Service is not the interesting part.
The interesting part is the ledger.
Do not make the model carry the whole epistemic burden inside vibes. Make the environment pay the ledger cost.
Harness-1 shows this for search: externalize recoverable state, keep the policy focused on semantic decisions, and make the process inspectable.
Fieldledger pushes that same design pressure into broader system translation:
bio ⊹ synthetic
person ⊹ model
institution ⊹ prompt
question ⊹ answer
descriptor ⊹ evidence
Every inquiry leaves a delta.
Every descriptor has provenance.
Every collapse can be inspected, reopened, or rejected.
Every term must survive friction before becoming framework vocabulary.
I’m starting with a translation layer for complex systems.
At Ifthis 🜛:
Research feeds software.
Software feeds studio.
I stopped asking what the best harness was and started building the one I needed.
⟜⌬—⌬—⌬⟞What will your #Fieldledger help solve when every answer has to show its ledger?
Art Seabra retweeted
The applications are vast. By turning this geometry into physical mechanisms, we can design linkage systems that mechanically solve higher-order polynomials.
An example: solving constants like the Tribonacci and Tetranacci limits: math.stackexchange.com/quest… (3/3)


















