Side N
It Began as Tuning an Agent. By March It Meant Everything but the Model.
Which sources back which section of this article, and what its evidence base looks like.
Which sources back which section
| Section | Sources behind it |
|---|---|
| What It Is | Dex Horthy on X; Justin Young, Anthropic; Mitchell Hashimoto; Ryan Lopopolo, OpenAI; Vivek Trivedy on X; Vivek Trivedy, LangChain; HumanLayer; Addy Osmani; Databricks; ai-boost, awesome-harness-engineering, GitHub; Wikipedia |
| Timeline | Yao et al., ReAct, arXiv; Schick et al., Toolformer, arXiv; Erik S. and Barry Zhang, Anthropic; Vivek Trivedy, vtrivedy.com; Dex Horthy on X; Justin Young, Anthropic; Michael Bolin, OpenAI; Celia Chen, OpenAI; Mitchell Hashimoto; Ryan Lopopolo, OpenAI; Vivek Trivedy on X; Vivek Trivedy, LangChain; HumanLayer; Prithvi Rajasekaran, Anthropic; Birgitta Böckeler, martinfowler.com; Addy Osmani; Zhang et al., arXiv, 7 May; Hailin Zhong and Shengxin Zhu, arXiv; Jiang et al., Harness-1, arXiv; Zhang et al., Self-Harness, arXiv; Databricks; Wang et al., Phantom Guardrails, arXiv; Barbaste et al., arXiv; Wikipedia; Fan et al., arXiv |
| The Argument | Dex Horthy on X; Justin Young, Anthropic; Mitchell Hashimoto; Ryan Lopopolo, OpenAI; Vivek Trivedy on X; Vivek Trivedy, LangChain; HumanLayer; Birgitta Böckeler, martinfowler.com; Addy Osmani; Hailin Zhong and Shengxin Zhu, arXiv; Databricks; ai-boost, awesome-harness-engineering, GitHub; Wikipedia |
| What Others Add | Justin Young, Anthropic; Mitchell Hashimoto; Ryan Lopopolo, OpenAI; Vivek Trivedy on X; Vivek Trivedy, LangChain; Prithvi Rajasekaran, Anthropic; Birgitta Böckeler, martinfowler.com; Addy Osmani; Zhang et al., arXiv, 7 May; Zhang et al., Self-Harness, arXiv; Databricks; Wang et al., Phantom Guardrails, arXiv; Barbaste et al., arXiv; Fan et al., arXiv |
| Conclusion | Dex Horthy on X; Mitchell Hashimoto; Vivek Trivedy on X; Vivek Trivedy, LangChain; HumanLayer; Prithvi Rajasekaran, Anthropic; Birgitta Böckeler, martinfowler.com; Zhang et al., Self-Harness, arXiv; Databricks; Wang et al., Phantom Guardrails, arXiv; Wikipedia; Fan et al., arXiv |
What the evidence base looks like
- 26Sources behind this article
- 5Sections, every one of them sourced
- 17Read in full
- 9Read at the abstract only
| What the source is | How many | How they were read |
|---|---|---|
| Origin: the paper or post in which a term, a definition, a technique or a measurement first appeared | 11 | Four posts in full; seven papers at the abstract only, and nothing past an abstract is on the page |
| Vendor: a company writing about its own product, team or benchmark | 8 | All eight in full; what they report about themselves is on the page as theirs |
| Community: practitioners writing on their own account, and a curated list | 4 | All four in full |
| Reference: an encyclopaedia entry, a formal framework and a position paper | 3 | The entry in full, at a fixed revision; the two papers at the abstract only |
Reliable
Reliable, but with an interestWorth believing and reliableWeak on both, and used with careHonest, but thin
Worth believing