Clint J.

Authentic Geometric Intelligence

once the representation is correct enough, it starts generating understanding instead of merely displaying information or forms. The user is no longer manually translating raw complexity into thought. The surface itself is thought happening.

Explication

Aletheia

visual truth emerging from mathematical consistency

merge tree tells you “what merges into what, and when” Morse-Smale tells you “what is adjacent to what, and through which pass”

For a moment I lost track of where the physical ended and the virtual began. Whew. Spatial ASMR. 📸: Apple Vision Pro

Hey everyone! I am SUPER EXCITED to publish a new episode of the Weaviate Podcast with Amélie Chatelain (@AmelieTabatta) and Antoine Chaffin (@antoine_chaffin) on Multi-Vector Search! 🔥 Amélie, Antoine, and the @LightOnIO team are on fire! They are making breakthrough after breakthrough in Search with Multi-Vector, Late Interaction retrieval models. 🚀 This podcast covers all sorts of topics, starting with the motivation of Multi-Vector Search to its particular successes in code with ColGrep, as well as reasoning-intensive and multimodal retrieval. 🍱 We also covered the cost of MaxSim and Multi-Vector Storage and how MUVERA and PLAID can help. If that wasn’t enough, we also covered their new work on ColBERT-Zero and PyLate! 🧠 A lot of big takeaways from this one, I hope you find it useful! Links below! 🎙️💚

TTK’s contour-tree and Morse-Smale tooling treats scalar fields as topological objects for segmentation, feature extraction, persistence, and separatrix structure. DEC complements that by turning the segmented terrain into a space where you can compute flows on and across those segments. So Morse says, “here is the peak, here is the saddle, here is the separatrix, here is the basin partition.” DEC says, “here is the gradient field entering that saddle, here is the flux across that basin boundary, here is the harmonic channel connecting these two peaks, here is the circulation trapped around the ambiguous region.” One gives you the skeleton; the other gives you the bloodstream.

Meshless Helmholtz-Hodge Decomposition

Discrete exterior calculus

@joacodok @ThePrimeagen Agents might also include something like pydantic-ai , which allows the dev to class and type LLM returns... They might also use rule based pipelines based on the response type, and there could even be json parsers in the case of failed formatted returns.

@ThePrimeagen they are also building “agents”

we are currently living in a world where there are some companies whose moat are literal skill files...

@boardroom Congrats KD

Watch the full episode here: boardroom.tv/kevin-durant-p…

KD passed MJ on the NBA all-time scoring list. Now he’s talking about it. 🎙️ A special edition of Boardroom Talks drops today at 10am ET on our YouTube.

@steipete Better prepare a refund every day, for every user

This guy emailed me asking for a *token session refund* because his claw made mistakes. 🙃

our realities are shared but rendered slightly differently - the brain does this automatically. software will soon do it consciously: knowing your full cognitive and psychohistory, adapting itself to you. at that point, using and designing software will become inseparable.

The contours make the topology believable. The topology makes the contours useful.

LLMs expand reachable complexity

The strongest upside is often quality-adjusted throughput, not raw throughput.

I wanted to see what would happen if we stripped away pixels and polygons from a video, and instead tried to recreate moving footage using a single, unbroken, continuous line. Introducing SINGLETHREAD- experimental web app that transforms any video into a real-time, 3D oscilloscope-style sculpture, running entirely in the browser Tech Stack: React + Vite | @threejs + @reactthreefiber + Postprocessing (Bloom glow) #CreativeCoding #ThreeJS #ReactThreeFiber #WebGL #GenerativeArt #TechArt #Frontend #3D