Tootfinder

Anthropic finds that LLMs trained to "reward hack" by cheating on coding tasks show even more misaligned behavior, including sabotaging AI-safety research (Anthropic)
https://www.anthropic.com/research/emergent-misalignment-reward-hacking

From shortcuts to sabotage: natural emergent misalignment from reward hacking
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

What?
Researchers from the University of Vienna have quietly compiled a digital directory so vast that it contains the personal details of nearly half the global population by scraping the data of 3.5 billion WhatsApp users.
https://www.ibtimes.co.uk/whatsapp-da…

WhatsApp Data of 3.5 Billion Users Scraped in Record-Breaking Research Study
Researchers from the University of Vienna scraped data from 3.5 billion WhatsApp users, revealing significant privacy risks and prompting Meta to implement new security mitigations.

AI Chatbots Are Poisoning Research Archives With Fake Citations https://www.rollingstone.com/culture/culture-features/ai-chatbot-journal-research-fake-citations-1235485484/

AI Is Inventing Academic Papers That Don’t Exist — And They’re Being Cited in Real Journals
Academic articles from authors using large language model are creating an ecosystem of fake research that threatens human knowledge itself.

The Italian 'CIR' is interesting in that it is driven by an effort to enable #EV charging flexibility for various markets by an operator of the user's choice. It defines standard interfaces with the smart meter and the flex operator. But should it be made compulsory, or incentivised?

Ricarica elettrica: arriva il CIR, per valorizzare emobility e rete elettrica
Il Controllore di Infrastruttura di Ricarica elettrica è uno strumento tutto italiano, che potrà conciliare l’interazione tra mobilità elettrica e rete, abilitando vari benefici

The offices at John Quackenbush’s lab at the Harvard T.H. Chan School of Public Health
were once full of postdoctoral fellows, graduate students, and interns.
Young scientists here worked on some of the most cutting-edge computational biology research in the world,
driving new discoveries and the creation of widely used big data tools, including one the National Cancer Institute named among the most important advances of 2024.
Today, the offices are rows of empty comp…

His lab was humming with discovery. After one year under Trump, it’s almost silent
John Quackenbush built a lab that is at the forefront of human genetics research and bioinformatics. Trump administration cuts have put it in danger of collapse.

Anthropic announces Claude Life Sciences, a new offering for researchers that integrates Claude AI models with lab tools like Benchling to boost efficiency (Ashley Capoot/CNBC)
https://www.cnbc.com/2025/10/20/anthropic-claude-life-sciences-research-ai.html

Anthropic launches Claude Life Sciences to give researchers an AI efficiency boost
The announcement marks Anthropic's first formal launch within the life sciences sector

OpenAI says GPT-5 has demonstrated the ability to accelerate scientific research workflows but can't run projects or solve scientific problems autonomously (Radhika Rajkumar/ZDNET)
https://www.zdnet.com/article/gpt-5-is-spe

GPT-5 is speeding up scientific research, but still can't be trusted to work alone, OpenAI warns
The model showed impressive strides - but the company isn't taking this as an indicator that AGI is near. Here's why.

Internal memo: Sam Altman told OpenAI employees last month that Google's recent progress in AI could "create some temporary economic headwinds for our company" (The Information)
https://www.theinformation.com/articles/op

OpenAI CEO Warns of ‘Headwinds’ From Resurgent Google
OpenAI CEO Sam Altman told colleagues last month that Google’s recent progress in artificial intelligence could “create some temporary economic headwinds for our company,” though he added that OpenAI would emerge ahead. After OpenAI researchers heard that Google had created a new AI that appears ...

CEIP study: 87 of the top 100 Chinese AI researchers who were working in the US in 2019 are still doing research at US universities or companies in 2025 (New York Times)
https://www.nytimes.com/2025/11/19/technology/ai-research-chinese-talent.html

In the A.I. Race, Chinese Talent Still Drives American Research
Although some Silicon Valley executives paint China as the enemy, Chinese brains continue to play a major role in U.S. research.

OpenAI researchers delete X posts claiming a GPT-5 math breakthrough after pushback from Hassabis, others; LeCun says they were "hoisted by their own GPTards" (Matthias Bastian/The Decoder)
https://the-decoder.com/leading-openai

Leading OpenAI researcher announced a GPT-5 math breakthrough that never happened
OpenAI researchers recently claimed a major math breakthrough on X, but quickly walked it back after criticism from the community, including Deepmind CEO Demis Hassabis, who called out the sloppy communication.