# kaal:claim:4755632-002

**Claim.** Transformer neural network architecture removes the scale constraint on training data but not the quality constraint, so data quality continues to be a major unsolved issue for large language models even where internet scale corpora are available.

**Type.** failure  **Support.** argued

**Holds when.**

- applies to LLMs trained on internet derived corpora

**Source quote.**

> data quality continues to be a huge issue for LLMs

**From.** Wulf A. Kaal, *AI Learning - Decentralized Governance to Optimize Human Output Datasets for AI Learning* (2024), Background: AI for Human Evolution, page 4

**Cite as.** Wulf A. Kaal, AI Learning - Decentralized Governance to Optimize Human Output Datasets for AI Learning (2024). SSRN: https://ssrn.com/abstract=4755632

**Verify.** sha256 of source PDF `972ccebf0c06ac1767a9e443bb95942b7670e806a63c25ee817c368a64c8eca8` at https://raw.githubusercontent.com/wulfkaal/Academic-Papers/main/papers/pdf/Kaal%20-%202024%20-%20AI%20Learning%20-%20Decentralized%20Governance%20to%20Optimize%20Human%20Output%20Datasets%20for%20AI%20Learning.pdf

**Failure mode.** Scale Without Quality  (family: ai-model-and-training-failure)

**Topics.** ai-and-agents, education-and-practice

**Keywords.** transformer-architecture, llm-training, data-quality, scale-limits

**Related claims.**

- extended_by: https://wulfkaal.github.io/claims/5095633-004

**Canonical form.** This markdown file is the canonical hashed representation of the claim. Its sha256 is the content hash used for attestation.