"For years, Big Tech CEOs have touted visions of AI agents that can autonomously use software applications to complete tasks for people. But take today's consumer AI agents out for a spin, whether it's OpenAI's ChatGPT Agent or Perplexity's Comet, and you'll quickly realize how limited the technology still is. Making AI agents more robust may take a new set of techniques that the industry is still discovering."
"One of those techniques is carefully simulating workspaces where agents can be trained on multi-step tasks - known as reinforcement learning (RL) environments. Similarly to how labeled datasets powered the last wave of AI, RL environments are starting to look like a critical element in the development of agents. AI researchers, founders, and investors tell TechCrunch that leading AI labs are now demanding more RL environments, and there's no shortage of startups hoping to supply them."
Consumer AI agents remain limited in autonomous software use despite high-profile visions. Reinforcement learning (RL) environments simulate workspaces for multi-step task training and are emerging as critical infrastructure for agent development. Leading AI labs build RL environments in-house but also seek third-party vendors because creating such datasets is complex. Startups like Mechanize and Prime Intellect and data-labeling firms such as Mercor and Surge are investing in interactive simulations. Major labs are considering large investments in environments, with some discussing spending over $1 billion. Investors hope a dedicated provider will become the industry-scale leader for RL environments.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]