# Keller Jordan **Entity class:** Person **Domain:** Artificial intelligence / Language-model training **Maturity:** Developed ## Definition Keller Jordan is an artificial-intelligence researcher and software author associated with the NanoGPT Speedrun and its open experimentation on faster, more efficient language-model training. ## Mechanism and significance The speedrun turns a compact training system into a measurable experimental organism for comparing optimization, architecture, data use, and training procedure. This makes algorithmic progress visible independently of simply adding more hardware or data. ## Relationships - **Research dossier:** [[research/ASI and RSI Timeline Research Moonshots|ASI and RSI Timeline Research Moonshots]] - **Ontology route:** [[ASI and RSI Timeline Ontology#Model Architecture and Compute Economics|Model Architecture and Compute Economics]] - **Software and benchmarks:** [[NanoGPT Speedrun]] · [[NanoGPT]] - **Research context:** [[Algorithmic Improvement Without Data Scaling]] · [[Cost-Performance Frontier]] · [[Intelligence per Dollar]] ## Sources and provenance - [[research/ASI and RSI Timeline Research Moonshots|ASI and RSI Timeline Research Moonshots]] — immediate source for this person's role in the broadcast research map. - [modded-nanogpt repository](https://github.com/KellerJordan/modded-nanogpt) — primary software record. ## Evidence boundary The research dossier establishes why Jordan belongs in the Moonshots ontology. Benchmark comparisons should preserve the task, hardware, target metric, code revision, and experimental conditions.