🕶️ TikTok’s $60B drama, OpenAI’s teen trouble

Robinhood has filed with the SEC to launch “Robinhood Ventures Fund I,” a new publicly traded fund designed to give retail investors access to startup shares before they go public. While details like fees, share counts, and specific investments remain unclear, the filing suggests a focus on sectors such as aerospace, AI, fintech, robotics, and enterprise software.

The move is pitched as a way to level the playing field for everyday investors often excluded from the gains of venture-backed startups, though similar funds like ARK’s Venture Fund already exist. Robinhood’s last attempt at offering retail access to private companies — through controversial “tokenized” stocks — drew backlash, making this fund a more traditional, mutual-fund-style approach.

Don't wait.

Sign up for The Hustle to unlock our side hustle database. Unlike generic "start a blog" advice, we've curated 100 actual business ideas with real earning potential, startup costs, and time requirements. Join 1.5M professionals getting smarter about business daily and launch your next money-making venture.

AI agents today remain limited, and researchers increasingly see reinforcement learning (RL) environments—simulated digital workspaces where agents practice multi-step tasks—as the key to advancing them. Unlike static datasets, these environments let agents interact with tools and software, receive rewards for success, and learn from mistakes, though they are far more complex and resource-intensive to build. Major AI labs like OpenAI, Anthropic, and Google are investing heavily, with Anthropic reportedly considering spending over $1 billion, while startups such as Mechanize Work and Prime Intellect, alongside established players like Scale AI, Surge, and Mercor, are racing to supply these environments.

The hope is that RL environments will become as foundational as labeled datasets were in the chatbot era, potentially minting the next “Scale AI for environments”. Yet, challenges remain: reward hacking, computational costs, and skepticism from experts who doubt RL will scale effectively. Some, like Andrej Karpathy, are bullish on environments but cautious about reinforcement learning itself. Ultimately, the field is crowded, competitive, and uncertain—yet many believe RL environments could unlock the next big leap in AI capability.