Forcing LLMs to be evil during training can make them nicer in the long run
A new study from Anthropic suggests that traits such as sycophancy or evilness are associated with specific patterns of activity in large language models—and turning on those patterns during training
LangChain’s Align Evals closes the evaluator trust gap with prompt-level calibration
LangChain allows enterprises to make and calibrate a model to evaluate applications and get it close to human preferences.Read More
You’ve heard of AI ‘Deep Research’ tools…now Manus is launching ‘Wide Research’ that spins up 100+ agents to scour the web for you
The implication seems to be that running all these agents in parallel is faster and will result in a better and more varied set of products.Read More
Informatica advances its AI to transform 7-day enterprise data mapping nightmares into 5-minute coffee breaks
Informatica’s data platform evolution shows how it uses AI to actually serve enterprise needs.Read More
The Download: OpenAI’s future research, and US climate regulation is under threat
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. The two people shaping the future of OpenAI’s
The two people shaping the future of OpenAI’s research
For the past couple of years, OpenAI has felt like a one-man brand. With his showbiz style and fundraising glitz, CEO Sam Altman overshadows all other big names on the
Google DeepMind says its new AI can map the entire planet with unprecedented accuracy
Google DeepMind unveils AlphaEarth Foundations, an AI system that processes satellite data 16x more efficiently to create detailed Earth maps for tracking deforestation, climate change, and environmental shifts.Read More
C8 Health started with an AI that gives anesthesiologists guidance on demand — now it’s targeting whole hospitals
A friendly red panda avatar serves up the knowledge from the organization’s own siloed databases, complete with citationsRead More
Mark Zuckerberg says ‘developing superintelligence is now in sight,’ shades OpenAI and other firms focused on automating work
So perhaps, these competing visions of superintelligence are actually far more similar than they are opposed.Read More
An EPA rule change threatens to gut US climate regulations
This story is part of MIT Technology Review’s “America Undone” series, examining how the foundations of US success in science and innovation are currently under threat. You can read the rest