Ollama's Technical Issues and Community Controversy
Ollama's Core Technology and Attribution Issues
Ollama's entire inference capability originally came from llama.cpp, the C++ inference engine created by Georgi Gerganov in March 2023. For over a year, Ollama's README contained no mention of llama.cpp, and their binary distributions didn't include the required MIT license notice for the llama.cpp code they were shipping.
The community opened GitHub issue #3185 in early 2024 requesting license compliance, which went over 400 days without a response from maintainers. When issue #3697 was opened in April 2024 specifically requesting llama.cpp acknowledgment, Ollama's co-founder Michael Chiang eventually added a single line to the bottom of the README: "llama.cpp project founded by Georgi Gerganov."
Technical Problems with Custom Backend
In mid-2025, Ollama moved away from using llama.cpp as their inference backend and built a custom implementation directly on top of ggml. This custom backend reintroduced bugs that llama.cpp had solved years ago, including:
- Broken structured output support
- Vision model failures
- GGML assertion crashes across multiple versions
- Models that worked fine in upstream llama.cpp failed in Ollama
- Lack of support for tensor types required by new releases like GPT-OSS 20B
Georgi Gerganov identified that Ollama had forked and made bad changes to GGML.
Performance Benchmarks
Multiple community tests show llama.cpp running 1.8x faster than Ollama on the same hardware with the same model:
- 161 tokens per second versus 89 tokens per second
- On CPU, the performance gap is 30-50%
- A recent comparison on Qwen-3 Coder 32B showed ~70% higher throughput with llama.cpp
The performance overhead comes from Ollama's daemon layer, poor GPU offloading heuristics, and a vendored backend that trails upstream.
Model Naming Issues
When DeepSeek released its R1 model family in January 2025, Ollama listed the smaller distilled versions (models like DeepSeek-R1-Distill-Qwen-32B) without clearly indicating they were distilled rather than the full models.
📖 Read the full source: HN LLM Tools
👀 See Also
Nia-docs tool creates local filesystem from documentation URLs for Claude AI
The nia-docs tool lets you run npx nia-docs with a documentation URL to create a local filesystem of the docs, which Claude AI can then access directly without additional configuration.
TEMM1E v3.1.0: AI Agent That Self-Fine-Tunes Using User Interactions
TEMM1E v3.1.0 introduces Eigen-Tune, a system that captures LLM interactions as training data, scores quality from user behavior, and fine-tunes local models via LoRA with zero added LLM cost. Tested on Apple M2, it corrected temperature conversions from 72°F = '150°C' to '21.2°C' after 10 conversations.
PACT 0.4.0 adds compound intelligence for AI coding agents
PACT (Programmatic Agent Constraint Toolkit) version 0.4.0 introduces compound intelligence features that help AI coding agents retain knowledge across sessions. The update includes research synthesis, a knowledge directory, and capability self-awareness systems.
DAUB MCP Server Lets Claude Generate and Render UIs via JSON Specs
DAUB is an MCP server that enables Claude to generate UI interfaces directly from natural language prompts, producing structured JSON specs that render as live interfaces without code generation or compilation. It exposes four tools including generate_ui, render_spec, validate_spec, and get_component_catalog.