Cite as: Real Problem AI problem “Why does the same local LLM end up copied into three different folders on your own machine?”. Opportunity score 5.8 out of 10 (severity 4, AI feasibility 8, market signal 5, competition gap 6). Category AI / Agents. Trend LLM. Source signal: Hacker News, item 47788385, comment by user alzoid, on "The local LLM ecosystem doesn't need Ollama.". Canonical URL: https://www.realproblem.ai/archive/why-does-the-same-local-model-end-up-copied-into-3-different-folders.
Why does the same local LLM end up copied into three different folders on your own machine?
Evaluating local inference runners means ending up with the same model duplicated across separate directory structures for Ollama, llama.cpp, and other tools, wasting disk space and creating version-tracking confusion.
Who has it: Developers running local LLMs across multiple inference backends.
Evidence
“When I evaluated running locally I ended up with 3 different folders containing copies of the same model in different directory structures.”
Quoted word for word from the public post linked below. Nobody submitted it to Real Problem AI.
Hacker News, item 47788385, comment by user alzoid, on "The local LLM ecosystem doesn't need Ollama."Why it is archived
Trimmed to 100-cap (lowest opportunity_score)
Scoring breakdown
Existing players
- Ollama model store · Proprietary storage layout, doesn't share weights with other runners.
- Manual symlinking · Works but requires knowing every runner's expected directory layout.
What they are missing
A shared local model cache/registry that any inference runner (Ollama, llama.cpp, vLLM, LM Studio) can point to, so the same GGUF/safetensors file is stored once and referenced everywhere.
Stack hint
#ATB26 · Canonical URL: https://www.realproblem.ai/archive/why-does-the-same-local-model-end-up-copied-into-3-different-folders