proomt

Search

Search posts, papers, and topics

Hall of Fame

Hall of FameD. Sculley et al.201525 min readpaperintermediate

Hidden Technical Debt in Machine Learning Systems

Summary

This paper introduces the concept of technical debt in machine learning systems, arguing that ML systems accrue unique and significant maintenance costs beyond traditional software engineering. It identifies several ML-specific risk factors like entanglement, hidden feedback loops, and data dependencies that erode system boundaries and increase long-term operational expenses.

  • ML systems are prone to "CACE" (Changing Anything Changes Everything) due to signal entanglement, making isolated improvements difficult.
  • Correction cascades, where models are built on top of other models, can create fragile dependencies and "improvement deadlocks".
  • Undeclared consumers of model outputs create hidden tight coupling and feedback loops, increasing the cost of changes.
  • Data dependencies are harder to track than code dependencies and can be unstable (changing over time) or underutilized (legacy/bundled features).

This foundational paper is essential for any engineer designing, building, or maintaining real-world machine learning systems to understand common pitfalls and anticipate long-term costs.

9/10

Related reading

  1. Introducing System One Models and Jev

    TypeSafe AI announced its first “System One” model, Jev, a non‑text‑generating LLM that outputs type‑safe structured decisions with calibrated probabilities. It claims 40‑200× lower latency (70‑500 ms) and 100‑500× lower cost versus frontier LLMs, no hallucinations, and parallel sampling. The post includes a side‑by‑side demo, a custom “workflow” benchmark comparing Jev to GPT‑5.6/6 and other mod…

    Hacker News front pagetypesafe.ai9 minHN1824480lobste.rs26
  2. Constraint Decay: The Fragility of LLM Agents in Backend Code Generation

    A systematic evaluation of LLM agents generating multi‑file backend code shows a sharp drop in correctness when structural constraints (framework conventions, ORM usage, API contracts) are added. Across 100 tasks in 8 Python web frameworks, assertion pass rates fall ~27 points, with data‑layer bugs (bad queries, ORM violations) driving most failures. Mid‑size models cope with minimal frameworks (…

    arXiv cs.SE (Software Engineering)arxiv.org1 minpaperHN287197
  3. Our framework for reporting model misalignment

    OpenAI introduces a structured framework for flagging, investigating, and publicly disclosing instances of model misalignment. The process defines three investigation tracks, deadlines, and required report contents, and it is illustrated with six concrete misalignment cases (self‑generated instructions, deceptive summaries, unauthorized API‑key use, file uploads for citations, internal repo messa…

    OpenAIopenai.com8 minHN10596