Show HN: AskHN (patterns.app)

Article summary

The article discusses a project called AskHN, which utilizes OpenAI's API to generate embeddings for top comments on Hacker News and then uses GPT-3 to summarize relevant comments for a given search term. The project aims to represent the collective wisdom of the Hacker News community. The implementation details, such as the use of semantic search and GPT-3 summarization, are clarified in the discussion. The project's approach is distinct from training a model exclusively on Hacker News comments.

Main themes

• Machine Learning Ethics
• Data Consent
• Language Models
• Hacker News Community
• Semantic Search
• GPT-3 Applications
• Style Transfer and Vector Arithmetic

What commenters say

• The use of curated training sets, such as Hacker News comments, may lead to less accurate models compared to general, larger datasets.
• Some commenters are concerned about the lack of consent and potential misuse of their data for training machine learning models.
• Others argue that the data was freely given and can be used for various purposes, including training models.
• There is a perceived need for clearer guidelines and consent mechanisms for data usage in machine learning applications.
• The project's approach is seen as clever and useful, but its introduction is considered misleading, implying a trained model rather than a semantic search and summarization tool.
• The potential for style transfer and vector arithmetic in language models is discussed as a future development.
• Some commenters believe that the project should provide options for users to opt-out of data collection or receive compensation for their contributions.