MACHINE LEARNING

–

29 January 2026 1h26

Youtu-VL Unleashing Visual Potential via Unified Vision-Language Supervision

→ View original post on X — @latentspacepod

29 January 2026

Andrew White: ChemCrow, ether0, PaperQA AI Breakthroughs

By

@latentspacepod

–

28 January 2026 22h52

@andrewwhite01 of @EdisonSci is cofounder and head of science at @FutureHouseSF
, known for many things incl ChemCrow (first LLM chemistry agent), ether0 (first reasoning model in a scientific domain), and paperqa (first superhuman literature agent).

→ View original post on X — @dfintelligence

Open-source Kimi 2.5 on own infrastructure with no limits

By

@dfintelligence

–

28 January 2026 19h32

Damn, but open-source models like Kimi 2.5 on your own infrastructure WITH NO LIMITS AT ALL, but WHAT A BLAST !!!!! That's some crazy shit !!!!!

Visual Generation Enables Human-Like Reasoning via Multimodal Models

By

–

28 January 2026 18h05

Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models

Pragmatic VLA Foundation Model

By

–

28 January 2026 18h05

A Pragmatic VLA Foundation Model

→ View original post on X — @jiqizhixin

FutureOmni: Multimodal AI Predicts Future From Video Audio

By

@jiqizhixin

–

28 January 2026 17h13

What if AI could predict the future from sights and sounds? Researchers from Fudan University & NUS introduce FutureOmni, the first benchmark to test this. It pushes multimodal models to reason across audio & video to forecast what happens next, not just describe the past.

AdaReasoner: Dynamic Tool Orchestration for Visual Reasoning

By

–

28 January 2026 15h27

AdaReasoner Dynamic Tool Orchestration for Iterative Visual Reasoning

→ View original post on X — @askalphaxiv

Patterning: New AI Interpretability Approach for Circuit Analysis

By

@askalphaxiv

–

28 January 2026 15h13

BIG new idea in interpretability called Patterning The basic idea: given a desired generalization/structure, determine what training data produces it So they treat what circuits/algorithms the model learns as something you can solve for by measuring how sensitive those internal

→ View original post on X — @ingliguori

AI Stack Layers: From Rules to Agentic Intelligence

By

@ingliguori

–

28 January 2026 14h17

AI isn’t one thing — it’s a stack. Rules → ML → Neural Nets → Deep Learning → GenAI → Agentic AI Agents don’t replace the layers below.
They orchestrate them. If GenAI answers,
Agentic AI executes.

→ View original post on X — @godofprompt

Specific personas boost LLM output quality to 94%

By

@godofprompt

–

28 January 2026 12h54

Telling an LLM to "act as an expert" is lazy and doesn't work. I tested 47 persona configurations across Claude, GPT-4, and Gemini. Generic personas = 60% quality
Specific personas = 94% quality Here's how to actually get expert-level outputs: