@thund3rstruck.bsky.social
Reposted
For Llama 3.1 8B (Q3), the M3 Ultra delivers around 78.61 tokens/second with large context windows (40k tokens)
October 16, 2025 at 5:55 AM
For Llama 3.1 8B (Q3), the M3 Ultra delivers around 78.61 tokens/second with large context windows (40k tokens)
What are the benchmark of tokens per second for the M3 compared to it and model size?
October 15, 2025 at 10:46 PM
What are the benchmark of tokens per second for the M3 compared to it and model size?
McDonald’s is too expensive.
October 15, 2025 at 10:33 PM
McDonald’s is too expensive.
Here is an example of what I said. www.tiktok.com/t/ZTMABu9jp/
TikTok video by @
www.tiktok.com
October 10, 2025 at 3:10 AM
Here is an example of what I said. www.tiktok.com/t/ZTMABu9jp/
They are quiet because they want to maintain the duopoly of power with immoral maniacs, no matter how bad this devolves into. Also they are paid by Israel to be feckless.
October 9, 2025 at 5:41 AM
They are quiet because they want to maintain the duopoly of power with immoral maniacs, no matter how bad this devolves into. Also they are paid by Israel to be feckless.
What we should take away from this, is that Dinesh is into Bestiality. File that away under gross. 🤢
October 9, 2025 at 5:33 AM
What we should take away from this, is that Dinesh is into Bestiality. File that away under gross. 🤢