Yared's log ๐Ÿซ†
98 subscribers
593 photos
105 videos
15 files
297 links
10 years of experience

๐Ÿ‡ช๐Ÿ‡น

i like thinking long term

Interested in ML, ASML, Robotics, photonic computing, faith and Humanity.

worked in: ML, Robotics, Full stack Dev..

Here I will log what's in my mind.
Download Telegram
Forwarded from Biniyam
For years, weโ€™ve been trying to understand why we dream.

And one answer might have come from machine learning.

When you train a model on the same data over and over, it starts to overfit. It stops learning the pattern and starts memorizing the examples.

Maybe the brain has the same problem. Every day, we live the same narrow dataset: same routines, same faces, same patterns.

So at night, the brain does something strange. It creates experiences that never happened.

Distorted memories. Impossible places. Random people.

Maybe dreams are not random. Maybe they are the brainโ€™s data augmentation.

Maybe dreams are the brainโ€™s best effort to not overfit

Source: Erik Hoel, โ€œThe overfitted brain: Dreams evolved to assist generalizationโ€, Patterns, 2021.

@b1n1yamBuilds
Forwarded from AI Post โ€” Artificial Intelligence
๐Ÿš€ MiniMax just unveiled M3, an open-weights AI model aiming to compete with the frontier labs on three of the hardest problems in AI at once.

Most models specialize in one area. M3 is trying to combine all three:

โœ… Frontier-level coding and agent capabilities
โœ… Massive 1 million token context window
โœ… Native multimodal understanding from day one

Its benchmark results are impressive:

โ€ข 59.0% on SWE-Bench Pro
โ€ข 66.0% on Terminal Bench 2.1
โ€ข 34.8% on KernelBench Hard
โ€ข 74.2% on MCP Atlas
โ€ข 34.8% SWE-fficiency

The model is powered by MiniMaxโ€™s new Sparse Attention architecture, which allows it to process up to 1 million tokens of context without the usual scaling bottlenecks.

@aipost ๐Ÿด
Please open Telegram to view this post
VIEW IN TELEGRAM
I ๐Ÿ’– Google
Please open Telegram to view this post
VIEW IN TELEGRAM
Google just dropped Gemma 12B; it only needs 16GB VRAM and outperforms its 27B predecessor models. ๐Ÿ’–๐Ÿ’–๐Ÿ’–

Thank you Google for giving us Local AI Supporting Dev community members a golden model

link
Please open Telegram to view this post
VIEW IN TELEGRAM
โค2
แˆ˜แˆแŠซแˆ แ‰€แŠ•
โค3
Holo3.1 is the latest family of Vision-Language Models (VLMs) for computer use agents. Building on Holo3, it expands support beyond browser and desktop automation to mobile environments, introduces native function-calling support for seamless integration with agent frameworks, and enables local deployment through optimized quantized checkpoints.

by China MiMo
https://huggingface.co/Hcompany/Holo-3.1-35B-A3B-GGUF
Forwarded from shego.codes
๐Ÿ’Œ
"What's meant for you doesn't pass you. It circles you again and again until you are finally ready."

#เฑจเงŽโ €
โ €โ € โ €โ € โ €โ € โ €โ € โ €โ € โ €โ € โ €โ €:ยจ ยท.ยท ยจ:
โ €โ € โ €โ € โ €โ € โ €โ € โ €โ € โ €โ € โ €โ € `ยท . ๐™š
โค2
The CBE app is working now and they have made the UI amd overal UX bater
Yared's log ๐Ÿซ†
https://youtu.be/Y9Wz2PV404E?si=PquRzo216IhJsqsa
Claude mythos is now available for public use with security measures and brand new name its now called Fable
แˆฐแˆ‹แˆ
A practical tip for anyone using Codex or AI coding agents on a large codebase:

You must design the codebase in a way that guides the agent.

In CRUder, I made Codex much more effective by setting hard architecture rules and restructuring the project for AI-assisted development:

โœฆ Split huge files into smaller focused files
โ€ข Kept files around 450 to 500 lines max
โœฆ Added folder-level AGENT.md guide files
โ€ข Defined what each folder does
โ€ข Explained what each file cover
- This Reduced unnecessary context loading
โœฆ Built the system in a modular, function-based structure
โ€ข Made the codebase easier to inspect
โ€ข Made edits safer and more predictable
โ€ข Reduced regressions during feature implementation

After this: Codex became more reliable, used context more efficiently.

AI coding agents are powerful, but they need proper engineering guidance.


Good architecture is not only for human developers. It is also how we enable AI collaborators to work at production scale.
โค2
Yared's log ๐Ÿซ†
A practical tip for anyone using Codex or AI coding agents on a large codebase: You must design the codebase in a way that guides the agent. In CRUder, I made Codex much more effective by setting hard architecture rules and restructuring the project forโ€ฆ
This is a significant solution to reduce regressions caused by unrelated fixes.

While building CRUDER initially, the model just kept stacking solutions onto the initial basic files it created, so I lost some progress due to regressions and a disproportionate consumption of context resulting from oversized documents.

So I restructured the project's overall architecture the way I described above, and after that, I have seen no regressions at all, and the token consumption shrank by a good amount.