Anthropic News | Sep 06, 2024

Brought to you by RivalSense - an AI tool for monitoring any company.

RivalSense tracks the most important product launches, fundraising news, partnerships, hiring activities, pricing changes, tech news, vendors, corporate filings, media mentions, and other developments of companies you're following 💡

---

Company Overview

Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems. Their first product is Claude, an AI assistant for tasks at any scale. Their research interests include natural language, human feedback, scaling laws, reinforcement learning, code generation, and interpretability.

Latest Updates

• Anthropic has introduced a new 'Enterprise' pricing tier for businesses operating at scale, featuring expanded context window, single sign-on, and data source integrations.
• Salesforce teams up with Anthropic to enhance Einstein capabilities with Claude.
• Anthropic's integration enhances specialized work environments and speeds up coding without needing extra installations.
• Anthropic's Claude models are now integrated with BigQuery, enabling advanced AI capabilities directly on BigQuery data.
• Anthropic is introducing a 500k context feature for Claude.
• Anthropic released native GitHub support for its chatbot Claude, available only in the Enterprise version launched today.
• Anthropic is adding function calling to their AI provider.
• Anthropic recently launched the Claude Quick Start Project Library, featuring ready-made projects like the Claude AI Customer Service developed with Pietro Schirano to help developers quickly build AI applications using the Anthropic API.
• On September 5, Anthropic launched an enterprise solution for secure team collaboration with SSO and IP management.
• Anthropic is among the California-based companies whose high-performing generative AI systems are subject to new U.S. government reporting and potential tougher regulations.
• Jan Lieke, former co-leader of OpenAI's Superalignment team, has joined Anthropic.
• Canadian artist Tim Boucher sent a letter to a San Francisco judge to object to the mischaracterization of his work in a lawsuit against Anthropic.
• Anthropic's new mobile app version of Claude allows users to build a working bill splitting app directly from their phone in less than 60 seconds.
• Amazon is partnering with Anthropic to upgrade Alexa using Anthropic's advanced Claude AI models, aiming to enhance Alexa's conversational abilities, personalized shopping suggestions, and smart home controls.
• Anthropic announced that artifacts are now generally available as of August 27, 2024.
• The US AI Safety Institute partnered with Anthropic and OpenAI for pre-release access to new AI models to focus on safety research and risk reduction.
• Anthropic and OpenAI agreed to share models with a U.S. government safety institute.
• Amazon will release a new AI-powered Alexa using Anthropic's Claude models this October.
• Anthropic's chatbot Claude is now leading the AI assistant market with its new 'Artifacts' feature, which allows users to create visual projects from AI outputs.
• Anthropic's Artifacts feature, which allows users to visualize and collaborate on AI-generated content, is now generally available.
• Anthropic has published the 'system prompts' that drive Claude and released a feature allowing end-users to reduce in-cache token costs by 90%.
• Anthropic released Claude 3.5 Sonnet on 20 June 2024, which outperforms other LLMs for coding tasks, and introduced Artifacts, a feature enabling the creation and sharing of various digital assets.
• As of 26 August 2024, the Elasticsearch Open Inference API now supports Anthropic’s Claude models, including Claude 3.5 Sonnet, Claude 3 Haiku, and Claude 3 Opus.
• In May, Chu Ruisong, President of AWS Greater China, announced that their cloud platform Bedrock offers access to leading LLMs, including Anthropic's closed-source model Claude 3.
• On 25 August 2024, Anthropic released Claude 3.5, which can generate calendar entries from an image.
• Anthropic enabled CORS support for their JSON APIs, allowing direct calls to Claude LLMs from a user's browser.
• Anthropic started supporting LaTeX rendering in Claude.
• Zed announced new AI features powered by Anthropic's Claude and is collaborating with Anthropic's Rust engineers to further develop its open-source codebase.
• Anthropic's Claude model is supported by Centric's custom agentic framework, which can be deployed on AWS, Azure, or GCP.
• Anthropic, the Amazon-backed AI startup, was hit with a class-action lawsuit in California federal court over alleged copyright infringement involving the use of pirated books in its AI models.
• Three writers filed a class-action lawsuit against Anthropic, claiming the company misused their works to train its AI chatbot, Claude.
• A group of authors sued Anthropic for copyright infringement, alleging the company used pirated books to train its Claude AI chatbot.
• Anthropic asked the court to dismiss all but the core direct copyright infringement claim in the lawsuit filed by music publishers, arguing that other claims are implausible.
• Intuit partnered with Anthropic to use Claude AI for generating clear tax explanations, improving helpfulness ratings and conversion rates during peak tax filing season.
• Anthropic's new prompt caching with Claude can reduce costs by 90% and latency by 85%.
• Pulpit AI is partnering with Anthropic to develop AI models for sermon analysis and personalized content creation.
• Anthropic partnered with Caylent to accelerate GenAI production timelines and enhance AI solutions for businesses.
• Anthropic has expanded its model safety bug bounty program and made Claude available in Brazil as of August 2024.
• Anthropic introduced artifacts and projects for pro users on Claude.ai.
• Anthropic launched Claude finetuning, a new 'Evaluate' tab in the Console, shareable and remixable artifacts, shareable Projects, a developer course on API fundamentals, and a model safety bug bounty with rewards up to $15,000.
• Datadog introduced LLM Observability to monitor the performance and safety of Anthropic-powered applications.
• OpenAI co-founder Chris Olah announced he will join Anthropic.
• OpenAI co-founder Dario Amodei joined Anthropic.
• OpenAI co-founder Greg Brockman is on leave through the end of the year, and another co-founder, John Schulman, has joined Anthropic.
• OpenAI is experiencing significant leadership shifts that could reshape its future direction.
• Anthropic has added an 'Enterprise' section to their website, indicating a focus on larger business customers.
• A new LLC's MVP is 80% powered by Anthropic's LLM layer.
• Amanda Askell, who leads the team at Anthropic responsible for Claude's personality, has been named one of TIME's 100 Most Influential People in AI for 2024.
• Anthropic CEO Dario Amodei and president & co-founder Daniela Amodei are featured on the 2024 TIME100 AI list.
• Anthropic has launched a Quickstarts repository to help developers rapidly build applications using their API.
• FeatherSnap uses Anthropic's Claude in Amazon Bedrock to offer real-time bird identification for its smart bird feeder and app users.
• Anthropic CEO Dario Amodei described the future of AI as a hive-mind with a corporate structure.
• Anthropic's models, including model 4, 4o, and mini4o, are being tested for accuracy in citing case law.
• Mike Sabes has started as Head of Corporate Finance & Strategy at Anthropic.
• Clay integrated Anthropic's Claude models into their platform, significantly boosting user demand and internal efficiency.
• Danny Bauer used Anthropic's Claude to generate 250-350 word preview posts from transcribed course content.
• Anthropic CEO predicts AI could increase the rate of biological discoveries by 100x.
• Anthropic has introduced a new 'Capture Screenshot' feature in Claude, with a record button toggle also observed.
• On 2024-08-29, a time-lapse video generator web app was created using Claude, an AI assistant developed by Anthropic, in collaboration with developer Netsi1964.
• Anthropic's Claude helps Wedia Group process up to 30,000 new images monthly across 20+ languages, significantly reducing tagging time for their clients.
• Michael Gerstenhaber, VP of Product at Anthropic, participated in a panel on generative AI at The Information’s San Francisco AI Summit.
• Cesar C. from Anthropic recommended using Risotto (YC W24) for IT ticketing integrated with Claude/Anthropic.
• Anthropic's AI model Claude reached $1 million in revenue in 16 weeks.
• Anthropic's recent study found that backdoors can persist despite safety alignment techniques like supervised fine-tuning and reinforcement learning from human feedback.
• The University of Florida's NaviGator AI platform now includes Anthropic’s Claude among its hosted large language models.
• Anthropic sent a public letter requesting various amendments, some of which were incorporated.
• Anthropic's hh-rlhf preference dataset showed a 26% incidence of alignment resistance where language model labelers disagreed.
• Researchers at Anthropic mapped portions of the neural network inside their Claude Sonnet model, identifying features related to people and concepts, and developed new metrics for evaluating these features.
• Zed Code Editor started adding AI features powered by Anthropic's Claude.
• Three authors filed a class-action lawsuit against Anthropic for allegedly using pirated works to train its Claude language models and are seeking damages and a permanent injunction.
• LooPIN announced a partnership with Anthropic to allow LooPIN credits to be used for Anthropic's API.
• Anthropic embraced prompt caching functionality.
• Anthropic's new Prompt Caching optimization offers significant practical benefits for Gen AI solutions beyond cost savings.
• Dario Amodei, CEO of Anthropic, testified in front of the US Senate about increasing biological risk from AI models.
• Anthropic introduced a Beta feature for cached prompts to save costs and reduce carbon footprint.
• Anthropic hosted a Founder Day with Factorial Funds and New Enterprise Associates (NEA), featuring insights from co-founder Daniela Amodei and CPO Mike Krieger on their product roadmap and API tools.
• Anthropic released updates on their interpretability research in July 2024.
• Anthropic's API had serious issues causing requests to stop midway, increasing token costs for users.
• Anthropic restricted access to Claude 3.5 Sonnet for free accounts.
• Daniela Amodei from Anthropic will be speaking at Dreamforce 2024 in San Francisco.
• Anthropic expanded its bug bounty program to find universal jailbreaks in its next-generation safety system.
• Anthropic's Claude 3.5 Sonnet and Claude 3 Haiku are now available in more AWS regions, including US West (Oregon), Europe (Frankfurt), Asia Pacific (Tokyo), and Asia Pacific (Singapore).

---

Did you find it useful?

If you liked this report, consider following your own companies of interest. Receive weekly insights directly to your inbox using RivalSense AI.