@tryharness

AI privacy has never been more important

We’ve spent a lot of time thinking about how the average person interacts with AI systems. Most people don’t want complex dashboards. They don’t want to manage multi agent workflows. They don’t want to jump between windows and endless tabs to ask questions. The decision to put the agent interface in a PiP window, and layering in screen awareness has been pretty powerful for our non technical users. There’s a cohort of medical writers who use harness for a lot of their work. It spread by word of mouth in their community, and they cited choosing harness over other AI tools because of the intuitive easy-to-use UX, and the platform being private by design. One user in particular commented writing their manuscript took orders of magnitude less time due to the amount of unconnected information to process that couldn’t be fetched programmatically. The frames harness sees, memories it makes and chat history accumulated over time all being stored on their device, along with TEE models being tagged and labeled made harness a clear choice for them due to the sensitive nature of their work. We’ve done not so great of a job highlighting the privacy first design of harness as a platform, but the goal is to meaningfully take a step forward in how building a consumer AI platform is built. From design and UX, to funding, to privacy

We did it! 🎉🎉 https://t.co/96v3RtplKF

$10 of inference available for $4.37 https://t.co/gon8apB1Ig

@bankrbot Can you please setup an automation to buy 0.25 WETH worth of CAPU over the next 6 hours? Split up buys evenly. After each buy order, use the harness-capu skill to stake the CAPU for daily compute. Report to me on telegram with each successful (or failed)