There's a bunch of different versions of the story. But in general, "smart techie with time on his hands spends months creating a useful custom cure for a loved one" is something that we've seen happen many times over the last 20 years. It's not prima facie unusual/unlikely.
AI
-
Use MoE Models for Unified Memory Hardware Like DGX Spark
By
–
As I have mentioned before, stop trying to get Dense models running on the DGX Spark/Mac Studios Unified Memory is best fit for MoEs because you only make each token go through a small subset of the numbers of parameters in the model Optimize for your hardware x.com/LeTechLead/sta…
-
Qwen 3.5 27B Dense vs 122B MoE Compared on DGX Spark
By
–
This guy has written an article comparing > Qwen 3.5 27B Dense vs > Qwen 3.5 122B MoE on the DGX Spark
-
User finds Grok AI to be surprisingly cheap
By
–
And with rclown by @marckohlbrugge you can backup your R2/S3/B2 in turn
-
User reports rapid expiration of grok-4-1-fast-non-reasoning AI model
By
–
If it's that, it could be some biological thing where a species is trying to kill itself
-
3D Soldier Deformation Issue: Optimized Mesh Scrambles Bone Weights
By
–
I always use R2 now (which is S3-compatible) I don't think it makes sense to store it on a VPS anymore, cause imagine it died or deletes, better keep it on outside storage like R2/S3/B2
-
Request for Model Aliases in AI Tooling
By
–
Good news "Recent news: In October 2025, Glass Lewis announced it would end its practice of issuing uniform proxy voting recommendations, shifting toward allowing clients to tailor advice to their own preferences. In January 2026, JP Morgan Chase's asset management unit cut
-
User Reports Slow Performance After Switching to Grok-Latest AI Model
By
–
Damn "They're co-owned by the Ontario Teachers' Pension Plan Board and Alberta Investment Management Corporation (AIMCo), two large Canadian pension fund manager" And of course Canada + teachers is gonna be left
