Re: the report that older models can find the same exploits as Mythos: This doesn’t mean much about its power relative to those models, and is a common mistake in evaluating frontier AI. We continually find that old models are far more powerful than we realized once new models
RESEARCH
-
US and China Race for AI Supremacy Military Dominance
By
–
Only two nations truly grasp the stakes. The United States and China.
— Nina Schick (@NinaDSchick) 14 avril 2026
China declared its ambitions early. As far back as 2017, AI supremacy became national doctrine. By 2019, they had a plan to embed AI into every layer of the People’s Liberation Army—from logistics to weapons… pic.twitter.com/DMHRcjjc2MOnly two nations truly grasp the stakes. The United States and China. China declared its ambitions early. As far back as 2017, AI supremacy became national doctrine. By 2019, they had a plan to embed AI into every layer of the People’s Liberation Army—from logistics to weapons
-
Smart AI better detects nonsense
By
–
The smarter your AI reasons, the harder it falls for BS.
— AlphaSignal AI (@AlphaSignalAI) 14 avril 2026
Most AI models will confidently answer a completely nonsensical question.
A new open-source benchmark measures exactly that.
BullshitBench v2 tests 70+ model variants across 100 carefully crafted nonsense prompts.… pic.twitter.com/jp1w7Q2iCFThe smarter your AI gets at reasoning, the better it detects BS. Most AI models confidently answer completely nonsensical questions. A new open-source benchmark measures this precisely. BullshitBench v2 tests 70+ model variants across 100 carefully crafted nonsense prompts.
-
MiniMax M2.7 Open-Source: Self-Evolution and Autoresearch at Scale
By
–
MiniMax M2.7 is open-source!
— Akshay 🚀 (@akshay_pachaar) 14 avril 2026
The most interesting part of this release isn't a benchmark number. It's what MiniMax calls "self-evolution," and it's essentially Karpathy's Autoresearch applied at full scale.
Every AI agent today runs inside a harness: the scaffolding of skills,… pic.twitter.com/91lBJXZax5MiniMax M2.7 is open-source! The most interesting part of this release isn't a benchmark number. It's what MiniMax calls "self-evolution," and it's essentially Karpathy's Autoresearch applied at full scale. Every AI agent today runs inside a harness: the scaffolding of skills,
-
AI Daily Roundup: Agents, Rivalry, Mobile AI, and Adoption Stats
By
–
Top stories in AI today: – AI agent hires humans, opens boutique in SF
– OpenAI talks Anthropic rivalry, Amazon upside
– Run Google’s latest AI on your phone for free
– Stanford's AI index: 53% adoption, 31% trust
– 4 new AI tools, community workflows, and more -
Machines as consumers: rethinking economic allocation frameworks
By
–
This at first seems deep, but collapses under just a very basic scrutiny. Why would machines only be producers, why can’t they be consumers as well? And in the world of limited resources (energy, computing power, etc.), we’ll still need mechanisms to allocate those resources… https://t.co/9oMZznqTCS
— Bojan Tunguz (@tunguz) 14 avril 2026This at first seems deep, but collapses under just a very basic scrutiny. Why would machines only be producers, why can’t they be consumers as well? And in the world of limited resources (energy, computing power, etc.), we’ll still need mechanisms to allocate those resources
-
Sequential World Model Enables Seamless Multi-Robot Collaboration
By
–
How do we make multiple robots work together seamlessly without getting bogged down by complexity? Researchers at the Chinese Academy of Sciences introduce the Sequential World Model (SeqWM), a novel approach where robots predict their own actions and share intentions based on
-
AI Challenges Specialization: New Paradigm Emerges
By
–
We Told Everyone To Specialize—#AI Just Proved That Wrong
by @nelsonchu_@Forbes Learn more: https://
bit.ly/412mrtn #ArtificialIntelligence #MachineLearning #ML -
Can Code Symbol Systems Emerge Inside Single Neural Networks
By
–
Can Code/Symbol Systems Emerge Inside a Single Neural Network?
-
AI exploit discovery capabilities versus human performance and costs
By
–
No evidence so far that AI can find exploits a human couldn’t, or even do it at lower cost.