63 karma · joined March 15, 2024
i'd rather use a nice-looking and well designed (UX) terminal than have to use a clunky ai native one.
These small but mighty GPUs pack a punch, and 10 of these nano GPUs has equivalent VRAM to an H100 GPU, at just a fraction of the cost. (2500 before tax)
I can see a lot of mini clusters or self hosted inference being done with these GPUs, and seeing as they already are being sold out everywhere goes to show that Nvidia isn't going anywhere in this AI trend.
but yeah does end up having a monthly ACU (agent compute unit) that it can use, but we'll see over time if it ends up being a worth it investment for smaller dev teams
But where have you been committing?
Have you really been coding as much as you say you are?
Check it out !!
If you like random projects like this feel free to follow me on twitter @kylejeong21
trying to see if there's demand for this idea. i've seen people build this out internally, but that i guess shows that people are using something like this:
AI agents to automate outbound enterprise sales
lets say exec@company.com signs up for a free trial,
the agents would find info on the person/company, save the data and send personalized outbound
either with human in the loop, or fully automated.
anyone want to use this?
Plus this is sort of a side project for me and users are just an added bonus.