In the latest issue of The Batch, Andrew Ng shares a simple recipe: using aisuite and MCP tools to spin up a highly autonomous (but unreliable) agent. (Practical agents need more scaffolding!) Plus: 📰 Claude Opus 4.5 is faster, cheaper, and stronger 📰 U.S. launches “Genesis
5
13
78
Replies
@DeepLearningAI Sounds like AI is getting a glow-up, but I'm still trying to keep my code from crashing 🤖
0
0
0
@DeepLearningAI Autonomous but unreliable is the perfect description of raw MCP stacks right now; the real unlock is treating aisuite + MCP as primitives and layering domain checks, retries, and coarse workflow graphs on top so agents behave like systems, not demos.
0
0
0
@DeepLearningAI experimenting with lightweight scaffolding before scaling fully autonomous agents has saved me more debugging headaches than raw compute ever could. curious to see which frameworks others are layering on top to balance reliability with speed.
0
0
0
We have open sourced AutoGLM, a project dedicated to teaching AI to operate smartphones like a human. After 32 months of development, we have created an AI agent that can move beyond chat and perform complete tasks within mobile apps, such as ordering food, managing
39
173
1K
We created a GitHub repo for all MCP at @Google. Get info on our remote managed MCP servers, open source MCP servers, examples, and learning resources. https://t.co/XAOVhJjB4W
30
148
2K
Today, we're releasing BU-30B-A3B-Preview (OSS)🧙 For the first time, you can run hundreds of browser tasks reliably on $1 of compute👀 > 30B parameters > 3B active > SoTA quality at real-time speed Welcome to the future of web agents. Try it locally or on our cloud today☁️
30
65
639
This is the DeepSeek moment for Voice AI. Today we’re releasing Chatterbox Turbo — our state-of-the-art MIT licensed voice model that beats ElevenLabs Turbo and Cartesia Sonic 3! We’re finally removing the trade-offs that have held voice AI back. Fast models sound robotic.
136
341
3K