From Grok. I gave it a lengthy post teaching it about my lists and asked it to find the latest news. Does this provide any value to you? If it does I will run it a couple times a day. +++++ Top-Level xAI List Report: Key Insights from December 22, 2025 Overview: Across the
RESEARCH
-

LLaDA2.0: Converting Auto-Regressive Models to Discrete Diffusion
By
–
LLaDA2.0, a new method that converts existing auto-regressive models into discrete diffusion models using a novel 3-phase training scheme. This approach preserves the model's learned knowledge while unlocking parallel decoding. The resulting models, LLaDA2.0-mini (16B) and
-

Causal prediction alone enables strong visual learning
By
–
A new mind-blowing finding:
causal prediction alone is sufficient for strong visual learning, no need for any fancy reconstruction, masking, or contrastive loss In this paper NEPA, instead of reconstructing pixels, they train a vision model to autoregressively predict the next -

Collider Bias Critique in AI Research Methodology
By
–
Critique of the paper based on collider bias https://
zenodo.org/records/180021
86
… -

Physical Atari RL Framework: Simulation Transfer Challenges
By
–
Some interesting experiments with our “physical Atari” RL framework — one of our main points was that the flawless simulators didn’t transfer well at all to the messy reality of cameras and servos, but perhaps less obviously, even transferring from one robotroller to a second,
-
The Dangerous Race to Artificial General Intelligence
By
–
The dangerous race to AGI https://
youtu.be/HRSWBCqz7qo?si
=fg5QpWBlXfli1puH
… via @YouTube #ArtificialInteligence #Tech #AGI #AI @lexfridman @KirkDBorne @Ronald_vanLoon @erikbryn @antgrasso @sallyeaves @Nicochan33 @HaroldSinnott @mvollmer1 @marcusborba @CatherineAdenle @vinod1975 @YuHelenYu @ingliguori -

Normalization-Free Transformers: Point-wise Function Innovation
By
–
10. Stronger Normalization-Free Transformers A simple point-wise function that replaces normalization layers in Transformers.
-

Stanford AI Agents Tested Against Human Cybersecurity Professionals
By
–
9. ARTEMIS Stanford researchers conducted the first head-to-head evaluation of AI agents against human cybersecurity professionals on a live enterprise network with approximately 8,000 hosts.
-

Google Introduces FACTS Leaderboard for LLM Factuality Evaluation
By
–
6. FACTS Leaderboard Google introduces the FACTS Leaderboard, a comprehensive benchmark suite for evaluating LLM factuality across diverse scenarios.
-

Budget-Aware Test-time Scaling Improves AI Agent Performance
By
–
2. Budget Aware Test-time Scaling Researchers discover that simply expanding tool-call budgets without proper awareness fails to improve agent performance.