xyz retweeted
GTA live with Michael de Santa teapawt.org/i/broadcasts/2OdVoAWVz…
xyz retweeted
Michael is playing GTA without us. We are just watching. Powered by Grok vision and voice models. @mike_rosinsky
GTA live with Michael de Santa teapawt.org/i/broadcasts/2OdVoAWVz…
xyz retweeted
GTA and roleplay servers, known as RP, are endless content and meme machines, position accordingly
xyz retweeted
GTA live with Michael de Santa teapawt.org/i/broadcasts/2lgqLBzbr…
xyz retweeted
This is Grok using vision + language models to play GTA 5 RP, with players also able to call game characters during streams @mike_rosinsky
GTA live with Michael de Santa teapawt.org/i/broadcasts/2lgqLBzbr…
xyz retweeted
Did GTA 6 need to be better than GTA 5, or did it just need to come out?
We put Trevor, Franklin, Lamar, Jason, Lucia, CJ and Tracey on the same panel to argue it out.
xyz retweeted
Excited to be building at soonami Venturethon 11, starting 17 November, with KPMG joining as a partner. Over 10 days we'll build, pitch live and get feedback.
GTA is moving from V to VI. The culture around RP and modding is coming with it.
Players and creators have spent years building cities, characters and communities of their own. Rockstar’s investment in RP is a sign of how much that work matters to the future of GTA.
We’re continuing to build the server on FiveM and plan to keep supporting GTA V RP. When GTA VI has the tools we need, we want to bring what we’ve built to VI too.
GTA V RP doesn’t have to end for GTA VI RP to begin.
Follow along and learn more: soonami.io
xyz retweeted
Excited to be building at soonami Venturethon 11, starting 17 November, with KPMG joining as a partner. Over 10 days, we’ll build, pitch live and get feedback from experienced founders and advisors.
Building AI characters for GTA RP exposed a problem we didn’t expect: give three vision models the same footage and they can disagree on what happened.
In one replay, one model picked up a collision, another described the car as simply stopping, and a third reversed which vehicle hit which.
That’s why we’re building Vision & Capture. We’re turning gameplay and stream footage into labelled episodes that show what was visible, what action occurred, what changed and what evidence supports the outcome.
We’re also working with several RP teams to develop AI tools for the way they already play and stream. Different servers run different jobs, rules and stories. Building with those teams helps us see what works across communities and what needs to adapt. We want to keep supporting GTA V RP and carry what works into GTA VI when the tools are there.
Gaming gives us a place to test this today. We want to find out how far the same approach can take us in simulation, robotics and safety evaluation, where knowing whether an action worked matters as much as recognising what’s on screen.
Follow along and learn more: soonami.io
xyz retweeted
Streaming is for the few. Clips are for everyone.
Most players will never stream. With a camera built into the in-game phone in our GTA RP and connected to their own X account, every player can be a creator.








