- The Rundown AI
- Posts
- OpenAI's new agent toolkit
OpenAI's new agent toolkit
PLUS: Manus teaming up with Alibaba's Qwen for China push
Rowan Cheung
March 12, 2025
Good morning, AI enthusiasts. The year of AI agents just got a major push forward — with OpenAI releasing powerful new tools that let businesses build their own autonomous assistants.
With capabilities like web browsing, file management, and computer use in a single API, is this the breakthrough that finally transforms the agent hype into an enterprise-ready reality?
In today’s AI rundown:
OpenAI releases new DIY agent tools
Manus, Qwen team up for China push
Connect AI coding assistants with external tools
Meta testing its own AI training chip
4 new AI tools & 4 job opportunities
LATEST DEVELOPMENTS
OPENAI
Image source: OpenAI
The Rundown: OpenAI just launched new tools that let businesses build their own AI agents—enabling custom bots to handle tasks like web browsing and file management and marking a major push toward bringing autonomous AI assistants into the enterprise.
The details:
The new Responses API combines web search, file scanning, and computer use capabilities, replacing the older Assistants API, which will sunset in 2026.
It allows companies to develop agents using the same tech powering Operator, with built-in tools for searching the web and navigating computer interfaces.
A new open-source Agents SDK will help developers orchestrate single and multi-agent systems while also providing safety guardrails and monitoring tools.
Early adopters include Stripe, which built an agent to handle invoicing, and Box, which created agents to search through enterprise documents.
Why it matters: 2025 has already been declared the year of AI agents, and China’s Manus took the hype to another level in the past week. While most agents have generated more hype than results, OpenAI expanding the ability for users to build and customize agentic tools may help bridge the gap between demos and real-world utility.
TOGETHER WITH TELY AI
The Rundown: With Tely AI, you don’t need an SEO or content team — AI does it all. It researches your industry, competitors, and website to find the best keywords to rank on Google and produces expert-level articles promoting your company & product.
With Tely AI, you can:
Publish up to 60 articles every month
See Google index results in just 2 weeks
Enjoy full automation for writing, SEO, and publishing
Convert traffic into valuable business leads
MANUS & ALIBABA
Image source: Manus
The Rundown: Manus just announced a strategic partnership with Alibaba's Qwen team to develop a Chinese version of its autonomous agent platform, following the company’s viral success over the past week.
The details:
The collaboration will integrate Manus's agent capabilities with Qwen's open-source language models and computing infrastructure.
Manus, which currently runs on both Anthropic’s Claude and Qwen, will adapt its full feature set for Chinese users and domestic platforms.
The partnership follows Manus' invitation-only preview that demonstrated capabilities surpassing OpenAI’s DeepResearch on agentic benchmarks.
Qwen has also had a busy month, launching a new open-source reasoning model (QwQ-32B) and major upgrades to its chat platform.
Why it matters: We’ve seen plenty of viral AI products fade fast, but a partnership with one of China’s top AI labs suggests there’s more here than just hype. While many dismissed Manus for running on Claude, it’s becoming clear that real value lies in packaging top models with the right tools, workflows, and interface.
AI TRAINING
The Rundown: In this tutorial, you will learn how to connect your favorite AI coding assistants, like Cursor or Windsurf, to powerful external tools using MCP (model context protocol) servers to tackle more complex coding tasks.
Step-by-step:
Choose your AI coding assistant (Windsurf or Cursor) and locate the MCP configuration area.
For Windsurf: Click the tools icon, select "Configure," and add the JSON code that connects to your desired service.
For Cursor: Open Settings > Features > MCP Servers and add a new server with the command that includes your API key.
Start using enhanced capabilities by asking your AI assistant to access external tools in your prompts.
Pro tip: Many MCP tool providers offer detailed documentation on GitHub repositories. Check these resources for specific commands and capabilities available for each tool.
PRESENTED BY INNOVATING WITH AI
The Rundown: Innovating with AI’s new program, AI Consultancy Project, trains AI enthusiasts with all the skills and strategies they need to capitalize on the rapidly growing AI consulting market – which is set to 8x by 2032.
The program offers:
The tools and frameworks to find clients and deliver top-notch services
A 6-month plan to build a 6-figure AI consulting business
Students getting their first AI client in as little as 3 days
META
Image source: Ideogram / The Rundown
The Rundown: Meta just began testing its first in-house AI training chip, according to a new report from Reuters — with the company hoping to reduce dependence on Nvidia and control its soaring AI infrastructure costs.
The details:
The chip is being manufactured by TSMC and is part of Meta's MTIA series—aimed specifically at AI training and inference workloads.
Its test follows the company’s successful first "tape-out" — a crucial development milestone that proves a chip design can be manufactured at scale.
Meta already uses in-house chips for Facebook and Instagram’s recommendation systems, with plans to expand to power genAI products.
It plans to start deploying the new training chips at scale by 2026, potentially saving billions on its projected $65B AI infrastructure spend.
Why it matters: With infrastructure costs skyrocketing and AI systems scaling rapidly, tech giants are increasingly investing in in-house silicon to cut costs and gain greater control. Meta is the latest to join this shift, aligning with OpenAI, Amazon, ByteDance, and others in reducing reliance on Nvidia—the chipmaking king.
QUICK HITS
💼 Fyxer AI - An AI Executive Assistant to help you gain an hour daily with organized emails, tone-matched replies, and crisp meeting notes*
🤖 Hunyuan-TurboS - Tencent AI that beats SOTA models on math, reasoning
🔎 Perplexity Windows App - New desktop app for AI-powered web search
🦉OWL - An open-source alternative to Manus AI’s agentic platform
*Sponsored Listing
Anthropic has seen its annualized revenue grow to $1.4B this month, coming on the heels of Claude powering Manus and the recent Claude Code and 3.7 Sonnet release.
Reka open-sourced Flash 3, a 21B parameter reasoning model rivaling OpenAI's o1-mini with a 32k context length and size for on-device deployment.
Anthropic CEO Dario Amodei predicted that AI will be writing 90% of the world’s code in the next 3-6 months and reaching nearly 100% in the next year.
Voice AI platform Cartesia announced $64M in Series A funding and unveiled Sonic 2.0 and Sonic Turbo AI with speed, ultra-realistic speech, and support for 16 languages.
Sam Altman revealed that OpenAI has trained a new model that excels at creative writing, saying it is the “first time I have been really struck by something written by AI”.
Harvey unveiled new AI agents with planning and adaptation capabilities that match or exceed lawyer performance on key legal tasks.
Luma Labs introduced Inductive Moment Matching, a new pre-training technique to generate superior quality images 10x more efficiently than current approaches.
COMMUNITY
Join our next workshop this Friday at 4 PM EST to learn about how to use Manus AI to boost productivity and automate everyday tasks easily with Dr. Alvaro Cintas, The Rundown’s AI professor.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.
That's it for today!Before you go we’d love to know what you thought of today's newsletter to help us improve The Rundown experience for you. |
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team
Reply