TechCrunch Publishes Glossary of Common AI Terms
TechCrunch has released a guide defining key AI terminology, from LLMs to hallucinations, to help readers navigate the expanding AI vocabulary.
Sources · TechCrunch AI
The news desk
Follow what’s changing.
Recent developments in artificial intelligence, with the original sources one click away.
All news
961–980 of 1,902Sam Altman Responds to New Yorker Profile and Reported Attack on His Home
OpenAI CEO Sam Altman addressed a New Yorker article questioning his trustworthiness and an apparent attack on his residence in a new blog post.
Sources · TechCrunch AI
Anthropic Temporarily Bans OpenClaw Creator Following Pricing Changes
Anthropic temporarily restricted OpenClaw's creator from Claude access after pricing adjustments affected the third-party tool's users.
Sources · TechCrunch AI
Anthropic's Mythos AI Model Raises Cybersecurity Concerns for Developers
Anthropic's new Mythos AI model is prompting calls for better security practices among software developers, according to experts.
Sources · WIRED AI
Molotov Cocktail Thrown at OpenAI CEO Sam Altman's San Francisco Home
A Molotov cocktail damaged the exterior gate of Sam Altman's San Francisco residence, with a suspect arrested by police.
Sources · NYT Technology
New York Times Seeks Reader Input on AI Chatbot Use for Nutrition Advice
The New York Times is soliciting stories from readers who have used AI chatbots for nutrition guidance, including diet management and health conditions.
Sources · NYT Technology
Suspect Arrested for Alleged Molotov Cocktail Attack on Sam Altman's Residence
A suspect was arrested for allegedly throwing a molotov cocktail at OpenAI CEO Sam Altman's home and making threats at the company's headquarters.
Sources · WIRED AI
Amazon Bedrock AgentCore Runtime Adds Stateful MCP Client Capabilities
AWS introduces stateful MCP server capabilities for Amazon Bedrock, enabling user input requests, LLM sampling, and progress streaming.
Sources · Amazon AWS AI
OpenAI Introduces $100/Month ChatGPT Pro Subscription Tier
OpenAI launches new $100/month ChatGPT Pro plan, filling gap between $20 and $200 subscription tiers in response to user demand.
Sources · TechCrunch AI
OpenAI Supports Illinois Bill to Limit AI Company Liability in Cases of Critical Harm
OpenAI testified in favor of Illinois legislation that would restrict when AI companies can be held liable, even in cases involving mass deaths or disasters.
Sources · WIRED AI
Researchers Propose New Methods for Privacy-Preserving Large Language Model Inference
New research introduces techniques to protect user privacy when using cloud-based LLMs without transmitting raw text prompts.
Sources · arxiv.org
Tool-MCoT Uses Small Language Models for Efficient Content Moderation
Researchers introduce Tool-MCoT, a fine-tuned small language model that uses external tools for content safety moderation with improved efficiency.
Sources · arxiv.org
AWS Launches Reinforcement Fine-Tuning on Amazon Bedrock; Researchers Release Pramana Reasoning Framework
Amazon Web Services introduces reinforcement fine-tuning for models on Bedrock, while researchers develop Pramana using ancient logic systems.
Sources · aws.amazon.com / arxiv.org
Conflicting Court Rulings Create Uncertainty Over US Military's Use of Anthropic's Claude
Contradictory decisions from a US appeals court and lower court have left ambiguous whether the US military can deploy Anthropic's AI model.
Sources · WIRED AI
New Research Explores LLM Capabilities in Knowledge Reasoning, Algorithm Innovation, and Network Security
Four new papers examine how large language models handle knowledge graphs, reinvent algorithms, impact planetary resources, and detect network intrusions.
Sources · arxiv.org
New Research Tackles Knowledge Gaps and Alignment Issues in Large Language Models
Four recent papers address LLM hallucinations, sycophancy, capability-targeted training, and attention architecture optimization.
Sources · arxiv.org
Tubi Launches First Native Streaming App Integration in ChatGPT
Tubi becomes the first streaming service to integrate a native app within ChatGPT, expanding the AI chatbot's functionality.
Sources · TechCrunch AI
AI Agents Require Process Redesign to Unlock Full Potential, Says MIT Technology Review
Unlike static systems, AI agents can learn and adapt dynamically, but require fundamental process redesign to realize their autonomous capabilities.
Sources · MIT Technology Review
Amazon Bedrock Projects Introduces Cost Management Tools for AI Workloads
AWS launches Bedrock Projects to help organizations track and analyze AI inference costs across specific workloads using Cost Explorer and Data Exports.
Sources · Amazon AWS AI
Anthropic Announces New AI Model Mythos for Cybersecurity Applications
Anthropic unveils Mythos, a new AI model for defensive cybersecurity, but holds back public release while testing with select companies.
Sources · NYT Technology / TechCrunch AI