Diarmuid O Keeffe ☘️

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

Boyd invokes Gödel, Heisenberg, and entropy to argue that inward-only refinement eventually increases mismatch and disorder. The system must go outside itself, create a broader concept, and test again.

[Jeff Dean @JeffDean]

People replace their phones every ~4 yrs. This means there are hundreds of millions of old phones discarded each year that are still perfectly usable as computing devices. @Google in collabration with @UCSD is exploring how to turn these old phones into cloud-computing “phone clusters”. Putting phones back in service in this way can directly reduce the environmental footprint of computing by avoiding the need for further raw material extraction, and taking advantage of the embodied carbon already incurred from manufacturing these devices, and modern phones actually are already quite powerful computers. Read more in the blog below ⬇️

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

Better models require simpler explanations.

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

The theory is the asset.

[Kat ⊷ the Poet Engineer @poetengineer__]

/humanity skills
│
├─ memory
│ ├─ myths.md
│ ├─ genealogies.md
│ ╰─ archives.md
│
├─ interfaces
│ ├─ alphabet.md
│ ├─ map.md
│ ├─ score.md
│ ├─ diagram.md
│ ╰─ code.md
│
╰─ runtime
├─ ritual.md
├─ library.md
├─ lab.md
╰─ workshop.md

[The City Mentor @TheCityMentor]

In postwar Germany, especially in Berlin, many buildings were stripped of their stucco and decoration

Today people realize that "Entstuckung" wasn't beneficial for either residents or the city itself

Fortunately some investors have started restoring their beautiful facades

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

HTML allows AI to design answers and unlock dynamic layouts, color-coded sections, tabs, collapsing accordions, and embedded visual indicators. It feeds data to the human brain visually and spatially, matching our cognitive processing power.

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

"Between stimulus and response there is a space. In that space is our power to choose our response. In our response lies our growth and our freedom." Viktor Frankl

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

Album Cover

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

The asymptotic nature of progress rather than perfection.
https://t.co/nYhFbUdad8

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

A strategy wins because its specific rule set effectively exploits its contemporary environment.

[tnkr @tnkrdotai]

This is what we've been building.

One repository for the full stack of a robot.

Documentation and Repository
Deployments
Collaboration and Contribution
Data

All versioned together.

The GitHub for robots is here.

[Diana @sdianahu]

8/8

a benchmark you saturate but can't connect to customer value is a vanity metric

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

The drop at 2:55 and the subsequent synth build is a work of art...
https://t.co/tYUwTpN7VK

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

This isn't backpropagation wandering blindly down a statistical hill; this is high-dimensional geometry executing a precise, surgical correction.

[Ahmad @TheAhmadOsman]

All roads lead to small and specialized models btw

[Mathematica @mathemetica]

Distance metrics help quantify how similar or different data points or sets are. This image shows visual examples of nine popular ones used across data science and machine learning.

It covers Euclidean, cosine, Hamming, Manhattan, Minkowski, Chebyshev, Jaccard, Haversine, and Sørensen-Dice, each with a simple diagram of how the calculation works.

These metrics sit at the core of clustering algorithms, nearest-neighbor methods, recommendation systems, natural language processing tasks, and geospatial analysis.

[Emerson S @Em_Nomadic]

Something wild is happening and nobody’s talking about it.

Individuals are open sourcing full-stack humanoids. 120 days. That’s the new normal.

HuggingFace went from 1,145 robotics datasets to 26,991 in one year. Hardware’s getting cheaper. Models are free. Code is open, but it’s all over the place.

You want to build something? Good luck finding all the pieces. It’s GitHub, HuggingFace, Docs scattered across five different sites, Discords, Reddit threads, Twitter threads nobody can find.

That’s the bottleneck now. Not access. Organization and whoever solves this owns the next wave.

That’s what we’re shipping. Coming soon. Big Updates… @tnkrdotai

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

@paulg In order to maintain the illusion of neutrality.

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

@johnschulman2

[Diarmuid O Keeffe ☘️ @FuzzieLabs]

The last mile of enterprise AI isn't a model problem, it's a runtime problem.

ICL — governed execution for AI agents: identity, approvals, rollback, immutable audit.

Turn every enterprise deployment into labelled training data that doesn't exist.

https://t.co/K0Bxu0wpFn