The FPS Review Chats with NVIDIA’s Chat with RTX About Its Own Content

[David_Schroth]

Introduction NVIDIA has released its Chat with RTX tech demo that we first got to see running last month at CES. NVIDIA built the app based on its TensorRT-LLM RAG developer reference project that it hosts on GitHub. It requires a GeForce RTX 30 or 40 series GPU that has at least 8GB of video […]

See full article...

 

[Grimlakin]

In tinkering with fedora you can run large libraries but you'll need a ton of memory. If your Library was over 40 gig in size you're pushing it to get that to run well on a 64 gig machine.

I need to figure out how to have Fedora digest external information though. THAT sounds really cool. Too bad it's (Nvidia Chat with RTX) only for RTX cards right now.

Fedora just added Vulcan acceleration to the AMD cards.

It's also the reason I increased my ram to 96 gig.

 

[Brian_B]

I don't feel like my job is in danger just yet.

 

[Grimlakin]

I don't feel like my job is in danger just yet.

Oh no far from it. AI Hallucination is still a HUGE issue.

I question the.. "AI SOLVED THIS IMPOSSIBLE THING" AI today will claim to do ANYTHING you ask of it. It's dumb in that.

 

[Peter_Brosdahl]

Definitely interesting to say the least.