bdvd 🇦🇷 (@bdvd_25) header
bdvd 🇦🇷 (@bdvd_25) profile picture

Diseño + 3D + agentes. Muestro cómo laburo con Grok Bot y Google Flow en el día a día.

Argentina
Joined July 2021
Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)Photo by bdvd 🇦🇷 (@bdvd_25)
Using Antigravity with Computer Use, custom CLIs, and local MCP servers, I turned my setup into my sidekick: • Blender & Concat via MCP • Headless ComfyUI & DaVinci automation • Native desktop navigation for non-API tools • Flow & Stitch via terminal
4
4
424
bdvd 🇦🇷 retweeted
this might be the best thing i've built with opus 5.5 every frame is python, no 3D assets were used
39
13
7
502
17,384
A tener en cuenta los que preparan sus post en un editor de texto
Replying to @bdvd_25 @b1rkh0ff
Yes. The open-source xai-org/x-algorithm code includes an is_pasted client flag that gets passed to the Grok reply-quality scorer for replies under accounts with 250k+ followers. Scores run 0-3; a 0 can trigger the RiskyHighVizReply label and bury the reply from non-followers. It has no effect on original posts or the main For You ranking.
2
I was tired of searching for the browser each time I had to ask something to the AI chat, so I built a mini app that calls my provider with a global hotkey, like yakuake. It opens on the monitor with the cursor, ready to chat right away.
2
1
65
Is for personal use, but i share it there, full vibecode in 30 minutes, so, many bugs to be expected! XD
6
Lindo laburo armaron en ComfyUI para MiniMax H3: resuelve la fórmula mañosa de fotogramas (17n+5), evita que se pierda el personaje en tomas largas y mete agent skills para sacar clips en batch a 1080p sin volverte loco. Directo a probarlo. #ComfyUI
1
1
56
MiniMax H3 con edición de un solo frame mantiene la identidad del personaje directo con imágenes de referencia. Si el prompt following se la banca así de bien, te ahorrás horas entrenando LoRAs dedicados para cada setup. Abre lindo debate. #StableDiffusion
1
48
Desmitificando un poco el tema de Moss: hicieron pruebas y con INT8-convrot anda de diez y respeta el prompt igual o mejor que en BF16, gastando mucha menos VRAM. Golazo para GPUs de consumo medio.
1
8
Salió un nodo nuevo para ComfyUI (TabCapture) para meter cualquier pestaña del navegador directo al canvas y capturar frames o video con audio de una. Súper cómodo para mandar referencias sin andar grabando pantalla con apps externas.
2
1
1
21
Para los que vienen renegando con la consistencia de personajes en video largo: armaron un workflow en ComfyUI con MiniMax H3 y LoRA turbo encadenando 42 escenas usando la guía de audio original. Queda súper sólido y sin saltos raros entre tomas.
2
1
1
48
It's Free!
Resetting everyone’s Gemini and Argon quota on Antigravity. Lots of updates this week so want to make sure everyone can build for the weekend. Lots more to come next week!
9
Now we wait for @FlowbyGoogle
introducing gemini omni 1.2 flash, now available in ai studio and the gemini api what's new: - fast & 30% lower cost - longer clips - edit in conversation - more natural speech - more controllable shots
5
it's free!
The full Quixel Megascans library is going free. Fab listings will update between now & Nov 12, so check back if an asset you want isn’t free yet. We’ve updated terms for all Quixel assets so you can train AI models on Megascans & Megaplants. Learn more: epic.gm/quixel-megascans-goi…
14