Quote from: davidm on Today at 16:31:23I don't think it's reasonable for "soldered" RAM on a tablet to be the first item on the cons.
Agreed. Someone didn't pay attention.
Since the 32 GB RAM in this tablet are not upgradable: If you care about using this for AI extensively, know that with 32 GB of memory (RAM+VRAM) and current SOTA AI LLM Qwen3.6-35B-A3B model[1], for its class and size (this is like the smallest size where a model become usable at all kind of tasks), and its best bang for the buck, 4-bit quant, you will be restricted to about this amount of context:
Quote from: reddit.com/r/LocalLLaMA/comments/1sq94qx/is_anyone_getting_real_coding_work_done_with.. I've come to the conclusion that (1) 32768 is the biggest context I can get away with in an adequately smart model, and (2) it just ain't enough.
More about running AI locally: notebookchat.com/index.php?topic=315954.0 ("Your own ChatGPT, offline: AI without the cloud on your laptop") and comments.
[1] artificialanalysis.ai/models/open-source?models=qwen3-6-35b-a3b%2Cqwen3-6-27b#artificial-analysis-intelligence-index (as of this writing, there's no other model that outperforms it at same speed and tasks, the dense 27B model is better, but it's much slower) (you can search for other models in the same table)