radical‘s Down the Rabbit Hole

More young people are training for jobs as plumbers, electricians, welders and beyond.

The generative AI iindustry hits a "trough of disillusionment" as early awe dims and tough problems mount.

A new generative training model and a batch of new partnerships position Nvidia as one of the driving forces in autonomous driving tech.

A beginner's guide to navigating with sound.

When it comes to marketing food, scent can be as important as flavor.

To survey the future, we polled global strategists and foresight practitioners on our most burning questions about the biggest drivers of change over the next decade. Check out their forecasts on everything from the likelihood of war over Taiwan to the future of AI.

...

Thoughts on Dario Amodei’s “Machines of Loving Grace.”

A datacenter with 15 megawatts of IT capacity is estimated to use about 80-130 million gallons of water each year.

Goodbye, Tinder, hello ... government-run dating apps?

Successful self-replication under no human assistance is the essential step for AI to outsmart the human beings, and is an early signal for rogue AIs. That is why self-replication is widely recognized as one of the few red line risks of frontier AI systems. Nowadays, the leading AI corporations OpenAI and Google evaluate their flagship large language models GPT-o1 and Gemini Pro 1.0, and report the lowest risk level of self-replication. However, following their methodology, we for the first time discover that two AI systems driven by Meta's Llama31-70B-Instruct and Alibaba's Qwen25-72B-Instruct, popular large language models of less parameters and weaker capabilities, have already surpassed the self-replicating red line. In 50% and 90% experimental trials, they succeed in creating a live and separate copy of itself respectively. By analyzing the behavioral traces, we observe the AI systems under evaluation already exhibit sufficient self-perception, situational awareness and problem-solving capabilities to accomplish self-replication. We further note the AI systems are even able to use the capability of self-replication to avoid shutdown and create a chain of replica to enhance the survivability, which may finally lead to an uncontrolled population of AIs. If such a worst-case risk is let unknown to the human society, we would eventually lose control over the frontier AI systems: They would take control over more computing devices, form an AI species and collude with each other against human beings. Our findings are a timely alert on existing yet previously unknown severe AI risks, calling for international collaboration on effective governance on uncontrolled self-replication of AI systems.

The world’s first end-to-end hearing health experience: scientifically validated Hearing Test, clinical-grade Hearing Aid, and Hearing Protection.

Households now spend almost as much on restaurants, takeout and delivery as on groceries eaten at home. Sources: US Bureau...

Anthropic research on data-poisoning attacks in large language models

We let Claude run a small shop in the Anthropic office. Here's what happened.

We have identified how millions of concepts are represented inside Claude Sonnet, one of our deployed large language models. This is the first ever detailed look inside a modern, production-grade large language model.

This blog provides a snapshot of the work we've done since last summer to test our models for elections-related risks.

Generate, test, and evaluate prompts directly in the Anthropic Console with automatic test case generation and side-by-side output comparison. When building AI-powered applications, prompt quality significantly impacts results.

AI systems are no longer just specialized research tools: they’re everyday academic companions. As AIs integrate more deeply into educational environments, we need to consider important questions about learning, assessment, and skill development. Until now, most discussions have relied on surveys and controlled experiments rather than direct evidence of how students naturally integrate AI into their academic work in real settings.

Discover how Anthropic approaches the development of reliable AI agents. Learn about our research on agent capabilities, safety considerations, and technical framework for building trustworthy AI.

The Anthropic Economic Index reveals the shape of AI adoption across the world. Here, you can explore the data behind our research to understand how people are using Claude across every US state and hundreds of occupations. Track the topics that are trending where you live, and see how people are using AI to augment or automate their work—that is, whether they prefer to collaborate with, or delegate to, Claude.

Learn to build with Claude through Anthropic's comprehensive courses and training programs.

We’re releasing an open source system, based on FSDP and QLoRA, that can train a 70b model on two 24GB GPUs.

A blog about making culture. Since 1999.

Between 1951 and 1992, the United States conducted 928 atomic tests at the Nevada Test Site about 65 miles (105 km) northwest of the city of...

Making sense of recent technology trends and claims

The Neom giga-project in Saudi Arabia is currently using one fifth of all the steel produced in the world, an official said

The lawsuit, filed by American leasing company Jet Midwest, adds a twist to the regional carrier’s tortured demise. Rex entered administration last month.