Understand this story
David Hall, author of Open Athena, argues that there is no public pool for post-training data comparable to pre-training. He points out that pre-training has ready-made public corpora such as Common Crawl, whereas post-training consumes tasks rather than documents, so there is no corresponding data commons. He uses SWE-bench as an example to illustrate the cost of manufacturing tasks: of more than 90,000 PRs from 12 Python repositories, only 2,294 became original tasks, and after OpenAI manually reviewed 1,699 of them, only 500 Verified tasks remained. David Hall believes that the open-source community needs to systematically transcribe human activity into RL tasks; otherwise, the ability to train frontier models will be concentrated in the hands of a few companies.
AI-generated from public reports
Latest development · Report
Open Athena author David Hall analyzes why there is no public pool for post-training dataDevelopments
1 confirmed developments
David Hall proposes that there is no post-training data commons
Source:English text is being prepared.
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 15
Hacker News Popular
Open Athena author David Hall analyzes why there is no public pool for post-training data
Open Athena author David Hall points out that pretraining has ready-made public corpora such as Common Crawl, whereas post-training consumes tasks rather than documents, so there is no corresponding public data pool. He uses SWE-bench as an example to illustrate the cost of task creation: among more than 90,000 PRs from 12 Python repositories, only 2,294 became original tasks, and after OpenAI manually reviewed 1,699 of them, only 500 Verified tasks remained. He believes the open source community needs to systematically transcribe human activity into RL tasks; otherwise, the ability to train frontier models will be concentrated in the hands of a few companies.