🔥 The ushering of agentic web
Complex tasks yet to perform
🐾 IN TODAY'S WILD
Anthropic has a new method to stop AI from turning harmful, tackling a core safety concern.
This era is also one of increased collaboration. alphaXiv is using AI to create interactive chat and community spaces for millions of researchers, transforming how academic work is shared. This move toward more autonomous systems is ushering in the "Agentic Web," where AI agents will perform complex tasks for users.
As AI becomes more integrated, research on "noosemia" is exploring why we attribute intentions to these systems. Google's Gemini now offers "Guided Learning" for deeper understanding, while OpenAI is studying the worst-case risks of open-weight LLMs. The financial world is also shifting, with AI funding surging while fintech investment slows, making AI adoption a matter of survival for many companies.
🦾 AI daily pulse
Anthropic says they've found a new way to stop AI from turning evil
AI is a relatively new tool, and despite its rapid deployment in nearly every aspect of our lives, researchers are still trying to figure out how its "personality traits" arise and how to control them. [LINK]A Comprehensive Behavioral Dataset for the Abstraction and Reasoning Corpus
The Abstraction and Reasoning Corpus (ARC) is a visual program synthesis benchmark designed to test out-of-distribution generalization in machines. Comparing AI algorithms to human performance is essential to measure progress on these problems. [LINK]alphaXiv AI tools transform research collaboration: paper blogs, chat, and community spaces for millions
According to Stanford AI Lab (@StanfordAILab), alphaXiv has rapidly expanded its platform over the past year by integrating advanced AI tools, including paper blogs and interactive chat features for arXiv papers. These tools are now used by millions of researchers globally, enabling more dynamic discussions and collaborative analysis of academic content. [LINK]
⚡️ Top trends
[Research] Agentic web: weaving the next web with AI agents
The emergence of AI agents powered by large language models (LLMs) marks a pivotal shift toward the Agentic Web, a new phase of the internet defined by autonomous, goal-driven interactions. In this paradigm, agents interact directly with one another to plan, coordinate, and execute complex tasks on behalf of users. [LINK]
Noosemia explains human tendency to attribute intentionality to generative AI systems
The increasing sophistication of generative artificial intelligence raises fundamental questions about how humans perceive and interact with these systems, and new research explores a fascinating phenomenon called ‘noosemia’, the tendency to attribute intentionality and even inner life to AI. [LINK]
💻 Top techies
[Google AI] Guided Learning in Gemini: From answers to understanding [LINK]
[Open AI] Estimating worst case frontier risks of open weight LLMs [LINK]
🔮 What else
Why open source matters in fintech AI In the first half of 2025, AI accounted for over half of global VC deals and nearly two-thirds in the U.S. But while funding for AI accelerates, investment in fintech has seen a marked slowdown. For fintech companies, navigating this moment isn't just a strategic priority, it’s a matter of relevance and survival. [LINK]
What can GPT-5 do?