Ex-Meta chief AI Raises $1 Billion to Build AI That Understands the Physical World

In partnership with

Base44 Launches “Superagents” to Let Anyone Deploy Autonomous AI Workers

AI app-builder platform Base44 has introduced Superagents, a new feature that allows users to create fully autonomous AI agents simply by describing tasks in natural language. Instead of manually connecting APIs, workflows, and infrastructure, the platform automatically assembles the required tools and deploys the agent, enabling even non-technical users to launch agents that can monitor systems, execute tasks, and trigger workflows continuously.

Once deployed, Superagents can run in the background, respond to events or schedules, and interact with external services such as productivity tools, messaging apps, and business systems. With persistent memory and built-in integrations, the agents can manage recurring workflows, send updates, and automate operational tasks across platforms. The launch reflects a broader shift toward “AI coworkers” that execute real work rather than simply responding to prompts, pushing agent platforms toward fully autonomous productivity systems.

Replit Launches Agent 4 to Power “Vibe Coding” With Autonomous AI Developers

AI development platform Replit has introduced Agent 4, the latest version of its autonomous coding system designed to turn natural-language ideas into fully functional software products. The new agent acts as a creative hub where users can plan, design, and build applications simultaneously, enabling the creation of websites, mobile apps, presentations, and launch assets within a single project environment.

Agent 4 also introduces parallel AI workflows, allowing multiple cooperating agents to execute tasks at the same time, such as generating code, designing interfaces, and configuring infrastructure, dramatically accelerating the build process. The system handles environment setup, dependency installation, and deployment automatically, effectively acting like a team of AI engineers working in the background while the user focuses on product ideas and design.

The release reinforces the rise of “vibe coding,” where developers and non-technical builders describe products in plain language and AI agents generate the underlying software stack. By enabling autonomous app creation and multi-agent collaboration inside one workspace, Replit is positioning Agent 4 as a step toward fully agent-driven software development workflows.

AgentMail Raises $6M to Build Email Infrastructure for AI Agents

Startup AgentMail has raised $6 million in seed funding to build an email platform designed specifically for AI agents rather than human users. The round was led by General Catalyst, with participation from Y Combinator, Phosphor Capital, and prominent angel investors including Paul Graham, Dharmesh Shah (HubSpot CTO), Paul Copplestone (Supabase CEO), Karim Atiyeh (Ramp CTO), and Taro Fukuyama.

Yann LeCun’s AMI Raises $1.03B to Build a New Kind of AI Beyond LLMs

Advanced Machine Intelligence (AMI), the new startup founded by former Meta chief AI scientist Yann LeCun, has raised $1.03 billion in seed funding at a $3.5 billion pre-money valuation. The round was backed by major investors including Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, signaling strong confidence in LeCun’s alternative approach to building artificial intelligence.

Unlike most AI labs focused on scaling large language models, AMI is developing “world models” — AI systems designed to understand how the physical world works and reason about cause and effect. LeCun argues that current LLM-driven systems are limited because they mainly predict text rather than building deeper models of reality, which are necessary for true reasoning, planning, and autonomous behavior.

The company plans to apply its technology across industries with complex physical systems, including robotics, manufacturing, aerospace, and biomedical research, with the long-term goal of building AI capable of interacting with and learning from the real world. The record-breaking funding round highlights growing investor interest in next-generation AI architectures that could power more capable autonomous agents and embodied AI systems.

ElevenLabs Introduces “Flows” in ElevenCreative for Multi-Model AI Content Pipelines

Voice AI company ElevenLabs has introduced Flows, a new visual workflow system inside its ElevenCreative platform that lets creators build complex generative pipelines combining audio, video, and image models. Using a visual canvas, users can chain together 35+ generative models, including speech, music, sound effects, image, and video generation, to produce complete multimedia projects from a single workflow.

Flows allows teams to design modular production pipelines where each step, such as voice generation, music composition, or video rendering, can be modified independently without regenerating the entire project. This non-destructive workflow makes it easier to iterate quickly and experiment with different models or creative elements while maintaining the rest of the pipeline.

What to Watch this Week

 🎙️AI on Fire: Real Builders. Real Heat.

Stories from people building AI Agents. Explore all episodes here

Building something in AI? Come talk about it.

AI on Fire is the podcast where we speak with founders and builders shaping the agent economy.

If you’re building AI agents, infrastructure, tools, or platforms — we’d love to feature your story.

👉 Apply to be a guest

🏆 The Leaderboard Never Sleeps

The global ranking of AI agents is shifting every day. Who’s on top? Who just dropped?

👉 Explore the Leaderboard →

🗺️ The Map of AI Agents (Live & Growing)

We’re charting the entire AI agent ecosystem — thousands of options across categories.
Your next agent is already on the map.

👉 Explore the Landscape Map →