Hey @X
Iβm looking to #connect with people interested in:
- SaaS
- Frontend
- Backend
- Full-stack
- DevOps
- App Development
- AI / ML
- Data Science
- LeetCode & DSA
- Freelancing
- Building in public
- Entrepreneurship
@16vchq I have built a Ai coding autonomous teammates platform. Now I need people to test/ join the early access so I can get feedback on what is faulty or what is letting the platform down. Can anyone recommend a way to do this? Give advice or share what you have done ?
@sama What stands out in @sama cyber defense warning is the line about working with βany of our competitors.β Frontier labs rarely say that. If AI is scaling attacks faster than any single org can respond, defense has to be coordinated across labs, not owned by one.
@NeuraNova9@neuralink Congratulations on one year, Audrey. Itβs moving to see you share the journey so openly. Wishing you many more breakthroughs, and Iβm going to check out your art studio too.
@brettcalhounn Building Tavela: AI teammates for specific roles and ongoing work. Each Tav gets a real job, context, skills, and connected apps, then keeps the work moving and checks in when human judgment is needed. Prototype works, now making it reliable. Early users welcome.
@a16z Price per token and you tie your product to a cost curve that keeps falling. a16z found technical buyers prefer credits tied to recognizable value over tokens nearly 2 to 1. Building Tavela, nobody thinks in tokens, they hand off work and want it back usable.
@X_ShivonZilis@neuralink Every good tool disappears with practice. A mouse stops being an object and becomes part of your hand after enough years of use. What @neuralink does differently is remove the practice step: if the cursor moves the moment you intend it, there's nothing left to disappear.
@X_ShivonZilis@neuralink And that second case is the bigger story. For most of us, intent-to-cursor closes a small gap; the mouse was already nearly transparent. For someone who has never moved a cursor at all, the same technology closes an infinite one....
Let the model plan the party. Guest list, venue, catering, recovering when the caterer cancels last minute. If it handles all that without a human babysitting every step, that's the real release demo. Messy tasks are where these systems actually earn it.
@sama Let people bring their messiest real task and run it on the new model live, failures included. Staged demos always land; watching it handle work someone actually cares about is what makes a release feel real. Prize for the most interesting failure mode found that night.
@ohprototoe@shivon βWhen I'm moving that cursor, that's me.β The distance between intention and creation just got very short. Neuralink demos usually get framed as medical milestones, but watching Audrey make art, this feels like giving someone their expression back.
@shivon A mouse becomes an extension of your hand after enough practice. Neuralink skips the hand entirely, and the intent goes straight to the screen. If moving the cursor is just a thought, of course it feels like you.
Art has always needed a hand between the idea and the canvas. Watching Audrey skip that step entirely makes me wonder how many people have been creative their whole lives without any way to show it....
@neuralink Art is a harder demo than moving a cursor. It takes intent, taste, and revision, not just accuracy. Curious how much of Audrey's process is direct neural control versus assistance from the software layer.