TII’s Falcon H1R 7B can out-reason models up to 7x its size — and it’s (mostly) open

According to TII’s technical report, the hybrid approach allows Falcon H1R 7B to maintain high throughput even as response ...

True agentic AI is years away - here's why and how we get there

Today's AI agents are a primitive approximation of what agents are meant to be. True agentic AI requires serious advances in reinforcement learning and complex memory.
GitHub

VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models

In this paper, we present VerifyBench, a benchmark specifically designed to evaluate the accuracy of reference-based reward systems. To create VerifyBench, we curated a diverse collection of ...
Business Insider

Presearch Launches Node NFT Licenses, Allowing More People to Power the Next Web

SAN DIEGO, CA, Dec. 09, 2025 (GLOBE NEWSWIRE) -- Presearch (https://presearch.com/), the privacy-centric, non-profiling engine, today announced its first Node NFT ...
The Conversation

What are small language models and how do they differ from large ones?

Lin Tian receives funding from the Advanced Strategic Capabilities Accelerator (ASCA) and the Defence Innovation Network. Marian-Andrei Rizoiu receives funding from the Advanced Strategic Capabilities ...
Fierce Healthcare

CMMI debuts ACCESS Model to spur use of tech in chronic disease treatment

The CMS Innovation Center has debuted a new model to encourage the use of technology to treat chronic diseases, which could be a boon for health tech companies that have struggled with reimbursement.
Futurism

Large Language Models Will Never Be Intelligent, Expert Says

Are tech companies on the verge of creating thinking machines with their tremendous AI models, as top executives claim they are? Not according to one expert. We humans tend to associate language with ...
IEEE

Policy Learning from Large Vision-Language Model Feedback Without Reward Modeling

Reinforcement learning (RL) has demonstrated remarkable success across various decision-making domains, including games [1], [2], autonomous systems [3], [4], and robotics [5], [6], [7]. A key factor ...
Computer Weekly

NTT brings Large Action Model to “market”

The latest trends in software development from the Computer Weekly Application Developer Network. NTT, Inc. used its recent NTT R&D Forum in Tokyo to detail a new AI technology called the Large Action ...
blockchain

Anthropic Research Reveals Serious AI Misalignment Risks from Reward Hacking in Production RL Systems

According to Anthropic (@AnthropicAI), their latest research highlights the natural emergence of misalignment due to reward hacking in production reinforcement learning (RL) models. The study ...
Fierce Healthcare

Function Health lands $300M, rolls out 'medical intelligence' AI model for health data

Function Health, a startup that offers lab testing and body scans for early disease detection, scored nearly $300 million in series B funding, boosting its valuation to $2.5 billion. Redpoint Ventures ...