← back
What's new from Anthropic and what's next: Alex Albert
Takeaway
Treat 3.5 Sonnet plus Artifacts/Projects as the first product surface designed natively for LLMs, and rebuild experiences around that rather than tacking AI onto existing UIs.
Summary
- AI products today are stuck in a 'magic star icon' phase — bolting LLMs onto existing surfaces instead of redesigning from the ground up, analogous to early electric factories and shrunken mobile websites
- Claude 3.5 Sonnet beats Claude 3 Opus on MMLU, HumanEval, GPQA and tool-use; on internal pull-request evals it scored 64% vs Opus 38%
- 3.5 Sonnet is 5x cheaper than 3 Opus ($3/$15 per M tokens), available on Anthropic API, AWS Bedrock and Vertex AI, with 200K context and near-perfect recall plus SOTA vision
- New Artifacts feature separates generated content from chat dialogue, enabling collaborative editing of essays, SVGs and React apps
- Projects ground Claude in user knowledge (style guides, code, transcripts); roadmap includes Claude 3.5 Haiku and 3.5 Opus later this year plus tool use API improvements and a console evaluate feature
anthropicclaudellm-apps
Original description
Explore Anthropic's latest strides in large language models, emphasizing enhanced reasoning and multimodal capabilities. We'll showcase how these advancements translate into powerful developer tools, APIs, and best practices for building sophisticated, RSP-aligned AI applications. Recorded live in San Francisco at the AI Engineer World's Fair. See the full schedule of talks at https://www.ai.engineer/worldsfair/2024/schedule & join us at the AI Engineer World's Fair in 2025! Get your tickets today at https://ai.engineer/2025 About Alex Albert Alex Albert is the Head of Developer Relations at Anthropic. Prior to his current role, he spent a year as a Prompt Engineer on Anthropic's Product Research team.