Hi @X people!
I'm 36 โ ex-founder in transpo / mobility.
Last 3 years I've gone all-in on AI โ agentic workflows, coding agents, open models.
Monday = time to expand my tech circle ๐
Looking to build with:
- Founders / Indie Hackers / Startup Builders
- Vibe Coders / Developers / AI Builders
- Designers / Creators / Growth Operators
- Freelancers / Students
- Anyone shipping on the internet
If you're creating, learning, and actually building โ let's connect.
Drop a Hey + what you're building right now ๐
Tried NVIDIA just dropped Nemotron 3.5 Super VL (FREE): a 120B model that only wakes up 12B parameters so it doesnโt have an existential crisis mid-task.
Tried it for 3 days.
How to use it: download the weights (yes, free), serve it with vLLM, and optionally fine-tune the thing on your own data so it stops being a generalist and becomes dangerously competent at your specific nightmare workflows.
Then sit back and watch it outperform the closed models youโre still renting by the token.
Nemotron 3.5 Super VL early verdict:
Itโs not the smartest open model (the Chinese flash models still win pure coding contests), but itโs the one that actually looks at your PDFs and charts, finishes the task 5ร faster, and lets you post-train it into something dangerously good at your specific job โ all while running on hardware you can put on a desk.
In other words: the model that stopped pretending raw IQ was the only metric that matters.