Inc42

From LLMs To Verticalisation: India’s Sovereign AI Stack Takes Shape

India’s sovereign AI push is taking shape through a layered stack spanning foundation models, public digital infrastructure, and applied AI systems.
InfoWorld

Multi-token prediction technique triples LLM inference speed without auxiliary draft models

With reported 3x speed gains and limited degradation in output quality, the method targets one of the biggest pain points in production AI systems: latency at scale.

Experiments Find LLMs Rely on Training Data & Lose Mid-Document Details

Tests on GPT and Claude found they ignored invented spells Fumbus and Driplo; training data can override new input, trust ...

Researchers baked 3x inference speedups directly into LLM weights — without speculative decoding

Researchers from the University of Maryland, Lawrence Livermore, Columbia and TogetherAI have developed a training technique that triples LLM inference speed without auxiliary models or infrastructure ...

Bringing quantum ideas to the messy world of disordered proteins

Imagine trying to design a key for a lock that is constantly changing its shape. That is the exact challenge we face in ...
Unite.AI

Easy Rewording Breaks AI Safety, Even for Gemini and Claude

AI safety tests found to rely on 'obvious' trigger words; with easy rephrasing, models labeled 'reasonably safe' suddenly fail, with attacks succeeding up to 98% of the time. New corporate research ...