we're launching the alpha version of playback - only 100 people will get in. Here's what to expect: this is the first chance to contribute redacted training data and earn ████ before anyone else. macOS m-series required. looking for heavy AI users: founders,

GM say ██ BACK

What you see is all there is. - ██████

Alpha launching soon™ - ███ Sign up for the waitlist now https://t.co/tKtzbaGfiO We will start sending ██████ and give out alpha security clearance.

We've been heads down building for the last two months - now it's time. the playback alpha is dropping in █ days. we're sending out a questionnaire to classify testers. we're also raffling 2 spots to the nads @monad_xyz & 2 to infinitians @infinitacity stay sharp ██ more

I pitched @TimDraper & @AdamDraper from @BoostVC and @drapervc this is what I learned: ██ imagine prepping for weeks, refining your pitch down to a tight, compelling 3 minutes… and then, 3 days before the pitch, they tell you it’s happening in a steakhouse 🥩 oh, and by

I want to own my fuckin AI

“I want to own my fucking AI”

Jarvis doesn’t run in AWS - always-on computer use assistants that watch your screen have to run locally

Playbacknetwork LLMs prompt text instructions but are unable execute them directly. A LAM(Large action model) can transform these prompts into actual actions(moving mouse,typing keys...) directly in your computer, through @playbacknet 's proprietary software and training data

This @OpenAI dev demo vid is crazzzy. Must watch. I'm most impressed by the latency. I suspect they're moved a lot of the o1 "thinking" to happen in the background while it's the user's turn. plus caching. phone use demo was slick. nice tool ui. https://t.co/nz385vrN9H

Introducing Playback ███████ in @monad_xyz Discord now. Join ███████:

GMONAD ███████

GMONAD ███████ We are █████ excited for the @monad_xyz community call tomorrow ███████

computer use is wild - been working on this area for a while and in principle it's doable but making it work properly is trickkky

Incredible progress but the model has to run locally for speed and privacy reasons. can't be VM because it needs your full context like @RewindAI or ██████ we're shipping something █████ 2024

@TonyGe_dev has to run locally. building this @playbacknet

The cost of running this is massive with Claude API pricing. Model to run locally - for cost and privacy reasons.

decentralized AI █████ █ █████. █████ so you can clone yourself