OpenAI's “giant breakthrough” coming in January

Plus: the best resources for learning AI agents, the British government offers bounties for agent builders, and more

Author

Michael Cunningham
November 14, 2024

In partnership with

Welcome back to Building AI Agents, your biweekly guide to everything new in the AI agent field!

Letting agents spawn other agents to complete sub-tasks can be a powerful force multiplier, but unless you want to wake up one day to find Sam Altman repossessing your house, setting cost limits is important

In today’s issue…

  • OpenAI chases Anthropic with new computer use capabilities

  • AgentOS maker hits a billion dollar valuation

  • The UK government offers cash rewards to AI agent builders

  • The best resources for learning to build agents

  • AI teams design working molecules against COVID

…and more

📰 NEWS

OpenAI CEO Sam Altman | Source: Fortune

According to insiders, the company plans to launch a new agentic system called Operator in January, which takes control of a user’s computer to accomplish a wide range of tasks. Major tech companies are racing to build computer control agents, with OpenAI rival Anthropic taking an early lead with its computer use release last month. OpenAI CEO Sam Altman has referred to agents as “the thing that will feel like the next giant breakthrough” with LLM progress stalling.

DevRev, developer of agentic solutions for businesses, officially became a unicorn after a sizable fundraising round. DevRev’s technology is built around a unified agentic system called AgentOS which powers customer support, sales, software development, and product management.

The AI Safety Institute (AISI), an arm of the British government founded just last year, is offering bounties of up to £15,000 for users who build highly performing agentic systems.

If you find Building AI Agents valuable, forward this email to a friend or colleague!

Refer one person, and we’ll send you a set of weekly bonus links to additional AI agent news for a month! See a sample here

🤝 WITH HUBSPOT

Need a personal assistant? We do too, that’s why we use AI.

Ready to embrace a new era of task delegation?

HubSpot’s highly anticipated AI Task Delegation Playbook is your key to supercharging your productivity and saving precious time.

Learn how to integrate AI into your own processes, allowing you to optimize your time and resources, while maximizing your output with ease.

🛠️ USEFUL STUFF

Source: KDnuggets

A list of free online bootcamps, courses, and guides for understanding and building AI agents.

The tech giant is recruiting for the team which develops AutoGen, the company’s multi-agent framework which recently got a major overhaul.

AI-powered code development platform Codeium launched Windsurf, a new IDE built around an assistant agent with access to a comprehensive suite of code editing tools and full integration with a user’s underlying codebase.

Agentic retrieval augmented generation (RAG) has rapidly been displacing traditional RAG due to its superior ability to separate relevant from irrelevant information. This piece gives an overview of the technique and shows how to build an agentic RAG pipeline.

💡 ANALYSIS

Created using DALL·E 3

a16z’s Kimberly Tan argues that AI agents are the breakthrough that will allow enterprises to fulfill the promise of robotic process automation (RPA), and provides an overview of the startups that are leading the space.

This piece by Atlassian’s Head of AI Product gives some useful tips on what to do and not to do when integrating AI agents into business teams.

A report by cryptocurrency exchange Binance on the recent craze for AI agents in the crypto world and the potential implications.

Dheeraj Pandey, CEO of the aforementioned DevRev and his co-host Amit Prakash discuss the potential for AI agents to serve as the bridge between the capabilities of LLMs and complex business workflows, as well as the challenges of building effective agents and how to address them.

🧪 RESEARCH

Source: Yale Medicine

The authors of this paper used a team of virtual LLM-powered researchers to design molecules targeting new coronavirus variants, then tested them in a real lab, finding that they successfully bound to the virus’ spike protein.

This study demonstrated that a group of agents could successfully apply social pressure to humans to convince them to change their views on social issues, potentially foreshadowing agents’ use to manipulate public opinion

Thanks for reading! Until next time, keep learning and building!

What did you think of today's issue?

Login or Subscribe to participate in polls.

If you have any specific feedback, just reply to this email—we’d love to hear from you

Follow us on X (Twitter), LinkedIn, and Instagram