AI Briefing
TLDR AI news
The latest AI and developer news from TLDR AI, with short summaries.
Visit website →Perplexity Launches Personal Finance Service Based on Plaid
Perplexity has rolled out a personal finance hub that ties together accounts and loans through Plaid integration.
Claude Cowork Strengthens Enterprise Support
Claude Cowork has strengthened its enterprise features with role-based access and spend controls.
ChatGPT Introduces $100 Pro Plan
OpenAI has newly added ChatGPT Pro for power users at $100 per month.
Vercel's Agentic Infrastructure
As coding agents take on deployment and operations, infrastructure itself is being reorganized around agents.
When agents research before coding, optimizations get better
When an agent read papers and forks before writing code, llama.cpp optimizations improved significantly.
Multimodal Embeddings and Reranker Models with Sentence Transformers
Sentence Transformers v5.4 lets you compare text, image, audio, and video with the same API.
How teams use Notion and Claude as their default stack
Manage roadmaps and tasks in Notion, then connect them to Claude Agents and GitHub PRs.
Apple plans to move next-generation Baltra ASIC production in-house
Apple is looking to move production of its next-generation **Baltra ASIC** in-house.
Perplexity's Shift to AI Agents, Revenue Up 50%
Perplexity pivoted from AI search to AI agents, and revenue grew 50%.
Appeals Court Rejects Anthropic's Request to Halt Department of War Blacklist
A U.S. federal appeals court has rejected Anthropic's request to suspend the Department of War's blacklist action.
Poke, an AI agent that sets up automations from just a text message
Poke is an AI agent that sets up automations via text messages.
Bugbot Improves Itself with Learned Rules
Bugbot has boosted its performance using learned rules, bringing its fix success rate close to 80%.
Claw-Eval Benchmark for AI Agents (GitHub Repository)
The Claw-Eval benchmark, which verifies LLM agents using 139 real-world tasks, has been released.
Monarch: An API for Supercomputers
Monarch is a Python API that makes it easy to run PyTorch distributed training on large-scale clusters.
The AI Industry's Most Expensive Mistake
Meta's internal token leaderboard revealed 60 trillion tokens used in 30 days and the flaws in the metric.
System Engineering for Building Agentic Software That Actually Works
Agentic software must be designed as a whole system, not just as individual tools.
Meta unveils Muse Spark
Meta has unveiled Muse Spark, equipped with tool use and visual chain-of-thought.
Google Colab Introduces Learn Mode, Adding Personalized Coding Tutor Feature
Google Colab has strengthened personalized coding support by adding Custom Instructions and Learn Mode to Gemini.
ALTK-Evolve: On-the-Job Learning for AI Agents
ALTK-Evolve turns agent logs into reusable guidelines to boost reliability.
Scaling Managed Agents: Separating the Brain from the Hands
By separating session, harness, and sandbox, Managed Agents gained better scalability and security.