Next is a web agent, which can collect data online and book services autonomously. Eg., autogenerating expense reports, transferring data between docs, booking travel. High-functioning agents will blast us into a new era of work — and OpenAI is once again leading the charge.
AGENTS
-
OpenAI Building Device Agents for Automated Task Execution
By
–
Per The Information, OpenAI is reportedly building two types of AI agents. The first is a device agent that can take over a user's device to perform tasks requiring text/clicks. I've said it once, and I'll say it again. 2024 is the year of AI Agents.
-
OpenAI Develops AI Agents for Device Automation
By
–
JUST IN: OpenAI is developing AI agents that take over devices and automate tasks. Plus, huge developments in AI from Google, 1X Tech, Midjourney, Brilliant Labs, Perplexity, Stable Audio, and the U.S. AI Safety Institute. Here's EVERYTHING you need to know:
-
Update on OpenAI Personalization, GPT-5, and AI Agents
By
–
Open IA is still cooking Personalisation feature to come in a short term. GPT-5 and agents are rather in the unclear state in regards to the timeline
-
Build AI Agents with Open Source Mixtral Models
By
–
Build Agents with Open Source Models! Explore the latest in context-aware search with our new template by @tjaffri
, powered by @MistralAI 's open-source mixtral model & hosted on @FireworksAI_HQ . It searches academic papers on Arxiv out of the box, and customize it with -
Human-in-the-Loop Agents: Interrupt and Authorize Features
By
–
Human-in-the-loop with OpenGPTs and LangGraph Today we’re launching two “human in the loop” features in OpenGPTs (powered by LangGraph): Interrupt
Authorize What do these do and why is this "human-in-the-loop" important for agents? Interrupt This allows the user to -
GPT-5 and AI Agents Coming Soon: What’s Next
By
–
Yes. GPT-5/Agents just around the corner. Writing more on this in tomorrow’s newsletter.
-
Microsoft pivots Copilot strategy from search to AI assistant
By
–
Love that Microsoft is going all in with Copilot, and positioning it more as an AI assistant rather than a search tool (Bing)
-
Self-Reflective RAG Enhancement Using LangGraph
By
–
Self-Reflective RAG using LangGraph Self-reflection can enhance RAG, enabling correction of poor quality retrieval or generations. Several recent papers focus on this theme, but implementing the ideas can be tricky. Here is a video and two cookbooks that show how to engineer
-
Building LLM Agents with Plugin Access and Personalization
By
–
Connery Plugins + OpenGPTs Blog + video links at the end! Some of the coolest LLM-powered apps involve granting the model agentic access to third-party systems, but there are a few things to think about when shipping: Personalization: How do you manage linked