{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreieckonh2acbwydy3fodsiyjkvgtkr3k5v6dyn5oddozm4sbxaigoi",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mj3cqgitjtu2"
  },
  "path": "/t/me-and-copilot-are-done/175125#post_1",
  "publishedAt": "2026-04-09T14:59:52.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "Every time I use copilot for anything intelligent it collapses in a heap… I get it, it’s an LLM designed to mirror, infer and basically make stuff up so you get drawn in. Good for writing poems and rewording emails but beyond that, I can’t find a use for it (not that I write poems btw)\n\nSo, I gave it 1 more shot.. “provide the best GPU solution for AI video generation and creation on my PC”\n\nCopilot (local) came up with “RTX 4090 (24GB) is the correct choice.” but they are now no longer supported! It also went on to claim “The 5070/5080 are “inferior” for AI because NVIDIA wants it that way\n\nSo I did my own online search and the opening page statement on the search page was, by Copilot “\n\n**The NVIDIA GeForce RTX 5090 is currently the top GPU for AI video generation, offering 32GB VRAM, 10,240 CUDA cores, and advanced multi-frame generation for high-resolution workflows.**\n\nI’m now confused and I think me and copilot are done… fairly sure there are smarter folks on here…",
  "title": "Me and Copilot are done"
}