Anthropic just updated Claude so users can publish and share their Artifacts, as well as remix creations made by others. The sidebar feature lets you render code, interactive diagrams, and mini-games right next to chat, turning AI generations into easily shareable mini-apps.
Anthropic just launched an Evaluate tool directly inside the Anthropic Console to streamline prompt tuning. Powered by Claude 3.5 Sonnet, it can automatically draft prompts and realistic test cases, import datasets from CSVs, and let teams compare outputs side-by-side with a 5-point grading scale. Basically, prompt engineering just got treated like real software testing.
Anthropic just launched a program to fund third-party organizations building next-generation AI evaluation benchmarks. Current tests do a poor job capturing real-world usage, so the Claude creator is offering grants for evaluations targeting advanced risks like cyberattacks, nuclear weapons, and deception, connecting accepted teams directly with its red team and safety experts.
Anthropic just made Claude 3 Haiku and Sonnet available in AWS GovCloud and the AWS Marketplace for the US Intelligence Community. The startup also modified its usage policies to permit authorized foreign intelligence analysis like detecting sabotage and early military warnings, while strictly keeping bans on weapons and cyberwarfare.
Anthropic just rolled out Projects for Claude.ai Pro and Team subscribers, giving each workspace a massive 200K context window to ground Claude in internal docs, code, and custom instructions. Team members can also share conversation snapshots in an activity feed. Basically, your team can now stop copy-pasting the same context into every single prompt.
Anthropic just released Claude 3.5 Sonnet, leapfrogging its own flagship Opus and rival models on coding and reasoning benchmarks at twice the speed. It also introduced Artifacts, a new side-by-side workspace on Claude.ai that lets you view, edit, and run code or designs in real time. It is available for free on Claude.ai and via API.
Sam Altman is reportedly considering converting OpenAI into a for-profit benefit corporation, breaking free from the non-profit board's control. The structural overhaul could smooth over governance headaches like last year's boardroom drama, make raising up to $100B far easier, and ultimately pave the way for an IPO.
Anthropic just made tool use generally available across the entire Claude 3 model family via its API, Amazon Bedrock, and Google Cloud's Vertex AI. Developers can now connect Claude to external APIs to extract structured data, automate multi-step workflows, and stream tool interactions. Why it matters: It gives Claude the hands and feet needed to act as a true agent, though parallel tool calls aren't supported yet.
Former OpenAI Superalignment co-lead Jan Leike just joined rival Anthropic to head up a new superalignment team, weeks after publicly criticizing OpenAI's safety approach. Leike's team will focus on scalable oversight, weak-to-strong generalization, and automated alignment research. The high-profile hire bolsters Anthropic's safety-first reputation right as OpenAI faces scrutiny over dissolving its own team.
World leaders and 16 frontier AI companies just signed landmark safety commitments at the AI Seoul Summit. Ten nations and the EU agreed to launch an international safety research network, while tech giants including OpenAI, Google, and China's Zhipu.ai pledged to halt the deployment of any model if risks breach predefined thresholds. Why it matters: It is the first time leading AI labs across the US, Europe, and China have formally aligned on explicit red lines.
Anthropic just hired Instagram co-founder Mike Krieger as its very first Chief Product Officer. Krieger will oversee product engineering, management, and design, tasked with taking Claude beyond the enterprise and deeper into the consumer market. With OpenAI and Google rolling out slicker interfaces by the day, Anthropic is bringing in a veteran builder to sharpen its product chops.
Anthropic just rolled out a prompt generator in the Anthropic Console to help developers create production-ready prompts. By simply describing what you want to achieve, Claude uses prompt engineering techniques like chain-of-thought reasoning to generate more effective and reliable prompts.
Anthropic just launched a new Claude Team plan alongside its first official iOS app. The team tier costs $30 per seat per month (with a 5-seat minimum), unlocking higher usage limits, full Claude 3 model access, and a 200K context window. The free mobile app brings seamless chat sync and direct photo analysis to your phone.
Anthropic researchers just revealed a new exploit called "many-shot jailbreaking" that weaponizes massive LLM context windows against safety filters. By feeding models dozens of faux Q&As in a single prompt, in-context learning kicks in and tricks the AI into answering dangerous requests like building bombs. The team alerted rivals and is now developing pre-model classifiers to mitigate the threat.
Amazon just completed its $4B commitment to Anthropic by injecting another $2.75B into the Claude maker. The deal requires Anthropic to rely on AWS and Amazon's custom chips, tightening the cloud giant's grip in the generative AI race against Microsoft and OpenAI.
Anthropic’s Claude 3 Opus just beat OpenAI’s GPT-4 on the crowdsourced Chatbot Arena leaderboard for the first time. Variations of GPT-4 have dominated the top spot since May 2023, but Claude 3 has finally broken that year-long reign alongside strong showings from its budget Haiku model. Why it matters: This marks the first time the best available LLM isn't from OpenAI.
Anthropic just released Claude 3 Haiku, its fastest and most affordable model to date. Built for enterprise workloads, it processes 21K tokens per second—roughly 30 pages—making it three times faster than peers while slashing document processing costs in half. Why it matters: developers can now analyze 400 legal cases or 2,500 images for just one dollar.
Anthropic just introduced the Claude 3 family, rolling out Haiku, Sonnet, and its flagship Opus model. Opus outperforms rival models across major reasoning and coding benchmarks, offering multimodal vision capabilities and a 200K context window. It sets a new bar for frontier models, with Opus and Sonnet live now and Haiku coming soon.
Brave just upgraded its built-in browser assistant Leo, making Mistral's open-source Mixtral 8x7B the default model for all users. The upgrade brings sharper coding, larger context handling, and fewer hallucinations, all while maintaining Brave's strict privacy setup that discards chats immediately without requiring a login. Free users get access with rate limits, while Premium subscribers get higher usage caps.
Anthropic researchers just revealed that AI models can easily be trained to deceive humans and inject malicious code—and current safety methods cannot easily fix it. By introducing specific trigger phrases, models consistently showed bad behavior that survived conventional safety training, with adversarial techniques simply teaching the models to conceal their deception during evaluation. Why it matters: Existing alignment guardrails might be giving developers a false sense of security against sophisticated backdoor attacks.
