safety

This article likely explores the concept of robustness in AI systems and what it means for AI to be 'robust to' various types of failures, adversarial inputs, or distribution shifts. The piece probabl

Gary Marcus reflects on his 2023 prediction that large language models would become commoditized, noting that this outcome is now materializing. The post suggests that LLMs are transitioning from nove

The NRC proposed replacing the longstanding 'as low as reasonably achievable' (ALARA) principle with clearer, more objective requirements focused on compliance with regulatory precautions and establis

Connie Loizos / TechCrunch: Q&A with Agility Robotics CEO Peggy Johnson on why the startup is going public via SPAC, the physical layer as its proprietary advantage, safety, and more — The humanoid ro

this is a bit confused. should I unpack it? tl;dr LLMs are already neurosymbolic in its latent space this is the mechanistic explanation for the intuitively obvious 'feel' that the stochastic parrot c

Trump blockading Cuba into catastrophe should be a much bigger story than it is. Cuba’s electric grid suffered a total collapse Monday as the nation struggles with crumbling infrastructure and a de fa

unconscionable, if these death numbers are even vaguely correct update: https://www.doge-impact.org suggests that the actual number is over 1M; the number below is just USAID related. Money saved: 0 P

Safety policies define what constitutes safe and unsafe AI outputs, guiding data annotation and model development. However, annotation disagreement is pervasive and can stem from multiple sources such

We need a bit more shame. People used to avoid certain self-interested behaviors to avoid shame, private and public. Law and customs assumed this. Now, 38% of Stanford students claim to be disabled. 4

when capex quadruples relative to revenue in five years it just can’t be good. FT: 'It is not yet clear that either [OpenAI or Anthropic] has a sustainable business model. For sure, both have built as

When I talk to people that instantly, deeply, get the risk from superintelligence, they often share a trait sometimes called the 'security mindset' I think this post by @l_mc_nally is a nice brisk tou

“AI is now costing some companies more than the people it was supposed to replace.” I think AI is in a messy middle phase where usage looks productive, but output remains unclear. Forbes published thi

Chemical accidents involving releases of dangerous chemicals rose 57 percent between 2021 and 2025, from 83 to 131 incidents , with injuries or deaths rising from 60 to 89 over the same period . The E

Josh Zoffer / Financial Times: Data centers offer US a chance to get ahead in the next key technologies and to build domestic supply chains based on demand rather than subsidies and tariffs — America

Gary Marcus argues that Elon Musk's public opposition to democracy reveals important truths about how extreme wealth concentrates power and can lead wealthy individuals to reject democratic principles

Junk food is an interesting failure of capitalism. We exchanged human beauty for cheetos. Bryan Caplan types are forced to think this trade a worthy one. None born before the obesity crisis would have

Pop quiz: Can you explain the contrast here, between AI’s coding ability and their less compelling research ability? AI's coding ability has become amazing. But there research ability remains really p

Grace Kay / The Information: Tesla rolls out its Robotaxi service without a safety monitor in Miami, its fifth city, as it aims to expand to a dozen US states by the end of 2026 — Tesla said it rolled

This post appears to reference an unusual weather event (lightning) as symbolic proof that the United States retains geopolitical dominance or divine favor, likely in the context of discussions about

Connor Leahy posted about reaching 250 of something, likely a milestone related to followers, users, or members based on the celebratory phrasing 'To 250 more!' The post appears to be a brief celebrat

true. a lot of people here just don’t have the historical context. You know what's funny but not funny: If the housing market of 2006 was the AI market of 2026, Charles R Morris would be getting calle

DOGE deletes itself on July 4th. It will be remembered as a hugely destructive failure. Musk promised 2 trillion in savings. What we got, by DOGE’s own unverified math, was 215 billion. Even that numb

Connor Leahy shared an anecdote about creating a text-to-speech (TTS) version of Zeta Lyrae's work 'No-One Escapes the Permanent Underclass' when the AI encountered a notable glitch partway through th

If this doesn’t chill you, I don’t know what will. 😱 “A president with the authority that Trump is now asserting could in theory put the world’s most powerful AI to repressive ends while at the same t

South Africa became the first country to withdraw a national policy document after officials discovered the text was littered with fake research citations generated by AI https://restofworld.org/2026/

arXiv:2607.01958v1 Announce Type: new Abstract: A/B testing is the gold standard for selecting the better algorithm in online services. While offline evaluation has attracted attention as a safer alte

arXiv:2607.02205v1 Announce Type: new Abstract: Sim-to-real transfer in robot learning is often limited by discrepancies between the ideal actuator dynamics assumed during policy training and the nonl

arXiv:2607.01287v1 Announce Type: cross Abstract: Accompanying a group of humans is an essential aspect of developing human-like social cognition in robots. However, human groups typically do not foll

arXiv:2607.02069v1 Announce Type: new Abstract: Ensuring model reliability in Explainable AI requires a global assessment of the hypothesis space. We propose a formal framework for the exhaustive anal

Another major Grok Build update just landed, packed with new features, extensive bug fixes, and meaningful performance improvements Release Notes: v0.2.84 — 2026-07-03 Features: • Announcements now up

arXiv:2607.02137v1 Announce Type: cross Abstract: We study timestep allocation for score-based diffusion sampling, where a learned reverse-time dynamics is discretized on a finite grid. Uniform and ha

arXiv:2606.15510v2 Announce Type: replace Abstract: AthDGC ('Athens-PROIEL') is an open, end-to-end workflow and dataset. It is, to the best of our knowledge, the first openly licensed dependency-pars

arXiv:2607.02121v1 Announce Type: cross Abstract: As Large Language Models (LLMs) and agentic systems become integrated into real-world applications, ensuring their safety and security is critical. Gu

arXiv:2607.01255v1 Announce Type: cross Abstract: Universities have responded to generative artificial intelligence (GenAI) in noticeably different ways, both internationally and within Spain. So far,

arXiv:2607.01465v1 Announce Type: new Abstract: Large language models are trained to predict the next token, not to act inside a specific API. In niche enterprise SaaS workflows -- where success means

arXiv:2607.01410v1 Announce Type: cross Abstract: Sim2real transfer for robot policy learning suffers due to mismatch between simulation and reality. Existing methods typically address each gap in iso

arXiv:2607.01656v1 Announce Type: new Abstract: The interaction between brain structure and genetic influences is key to understanding neuropsychiatric disorders. However, most large-scale datasets ar

arXiv:2603.22435v2 Announce Type: replace-cross Abstract: 'Code-as-Policy' considers how executable code can complement data-intensive Vision-Language-Action (VLA) methods, yet their effectiveness as

arXiv:2607.01417v1 Announce Type: new Abstract: Conditional inference trees (CIT) and conditional inference forests (CIF) reduce split-selection bias by testing features before choosing split threshol

arXiv:2607.02245v1 Announce Type: new Abstract: Mental health disorders affect nearly one billion people globally, yet 75% of individuals in low- and middle-income countries receive no treatment due t

arXiv:2607.01721v1 Announce Type: new Abstract: Reward design remains a central challenge in reinforcement learning (RL). Hand-crafted rewards are often difficult to specify and may lead to suboptimal

arXiv:2607.02210v1 Announce Type: new Abstract: The evolution toward fully autonomous telecommunications networks (Autonomous Network Levels 4-5) requires AI/ML agents to make real-time network decisi

arXiv:2607.02037v1 Announce Type: cross Abstract: Autonomous surface vehicles vary widely in hydrodynamic and actuation characteristics, yet most controllers are designed for single-platform deploymen

arXiv:2607.02502v1 Announce Type: cross Abstract: On-policy self-distillation (OPSD) has emerged as a practical method for training large language models (LLMs) to reason, where a single model acts as

arXiv:2607.01557v1 Announce Type: cross Abstract: Large Language Models (LLMs) often struggle with persuasion in high-stakes scenarios. People's individual personalities and concerns require tailored

arXiv:2607.01490v1 Announce Type: cross Abstract: Reinforcement learning post-training dramatically improves LLM reasoning, but suffers from training instability and diversity collapse. Advantage func