jonchi

generating avatar generator || stealth

[jonchi @jonching_]

RSI soon ??

[Rishabh Karnawat @rishabhkio]

guys, this is a serious breakthrough. Please read.

Researchers at MIT CSAIL just showed that GPT-6 Astra can read and edit the raw weights of another neural network directly. Weights are the huge set of numbers that make up an AI model. To a person they look kinda scary, so nobody edits them by hand. If you want a model to improve, you normally train it by showing it lots of examples and adjusting those numbers little by little. Nearly all of modern AI is built this way, and it takes a lot of time and compute.

In this work, they skipped training completely. They asked Astra to make Qwen3.5-0.8B better at math, with no gradients and no training. Astra looked at the weights, found an issue linked to the model getting stuck in reasoning loops, and changed only 0.016% of them. Its MATH-500 score went from 7.9% to 48.1%, about 6x higher, and its instruction following score tripled as well.

The researchers also gave Astra the raw weights of four game-playing networks with identical structures. With no tools and no list of options, it correctly named all four games: Breakout, Pong, Pac-Man and Freeway. Then, given the rules and code for a small version of Breakout, it wrote a network's weights from scratch in a single reply. That network scored 517.4, while one trained with PPO for 500 million steps scored 25.3.

With normal software, you can read the code, find the bug and fix that one thing. With neural networks, you have mostly had to retrain them and hope the problem goes away. This work suggests weights might start to behave more like code. If that holds up, you could fix some model problems in one edit instead of a long training run, inspect what a model does by reading it directly, and let AI improve other AI in a much more direct way.

It is still early. The authors themselves call these early signs, the game networks are tiny, and the Qwen fix targeted a flaw its model card already warned about, so this does not mean AI can rewrite any model at will yet. It needs to be replicated and tested on bigger models. But AI reading another network's weights and improving it this much with no training is a big deal, and I think a lot more people should be paying attention.

[jonchi @jonching_]

@hckmstrrahul Would love to try the tool

[jonchi @jonching_]

@ShutterCG Does it work for a cluster of buildings ? Like Hoi An in Da Nang ?

[Erick @ErickSky]

🚨 ¿Te enteraste?

DimensionalOS (dimOS) explotó y ya está en el #3 trending de GitHub.

Miles de devs están sacando sus agentes IA (OpenClaw y otros) del mundo digital y poniéndolos a controlar robots reales: cuadrúpedos, humanoides y drones.

Lo más brutal:
• Navegación autónoma real
• Memoria espacial + temporal
• Percepción y control directo de hardware físico

Esto ya no es simulación.
Esto es agentes viviendo en el mundo real.

REPOOOO👇

[jonchi @jonching_]

renaissance is coming

[Soleio @soleio]

This is the most visceral way to show the era we are now in.

11/10, no notes.

[Aurelien @Aurelien_Gz]

this is how you create a premium dark website..
​
​pure design mastery from google

[jonchi @jonching_]

@vamsibatchuk map UI has so much potential

[kepano @kepano]

currently on a mission to make a piece of furniture for every room in the house that reuses the same pattern of joinery, door handles, material proportions

so far the desk, the credenza, the wardrobe, the library... now working on the bed frame
https://t.co/UkkOFK83Wk

[jonchi @jonching_]

@Xaraphim Do you speak the language ? Happy to help

[Emmett Shear @eshear]

@jxmnop They all have the same core capability (a model shaped like all human cultural knowledge trained to act as an assistant). There is a large unknown about what this powerful thing is good for. But when someone invents a new thing, it’s easy to copy.

[prayingforexits 🏴‍☠️ @mrexits]

Steve Jobs on computer design

[Jason van Luipen @B_H_Creations]

@androo829 @ns123abc

[jonchi @jonching_]

@yamazombie1 Love the work!

[Tim Urban @waitbutwhy]

Whyyy did no math teacher of mine ever show me a visual like this

[Ren Yuan @_renyuan]

Visualization and sonification of strange attractors.

[Taelin @VictorTaelin]

After further experimentation, I say Sonnet is smarter in general, but not by far, and DeepSeek is even better in some aspects, like formatting. It is also faster, and 10x cheaper. This model is absolutely legit, and superior to GPT-4o and Gemini-2.

[jonchi @jonching_]

@HackerHouseTW thank you for the awesome week ! @thorwebdev

[Silicon Jungle @JungleSilicon]

an embedding space for Midjourney srefs.

search for styles using natural language
scrub through aesthetic properties like warm, cool, soft, sharp, muted, vibrant, emotions.

then - copy the sref and use it in a prompt.

[jonchi @jonching_]

@nic_carter goals