SMRTR ProgrammingMar 25, 2025lobste.rs

The role of developer skills in agentic coding

SMRTR summary

Thoughtworks is examining generative AI's impact on software delivery. They've developed a model for LLM-assisted coding, covering tasks like information retrieval, code generation, and transformation. Current interaction modes include chat, in-line assistance, and CLI.

Developers primarily use chat and in-line coding assistance, with some experimentation in larger code generation. Future focus may be on model improvements, size optimization, and organization-specific fine-tuning.

LLM-assisted coding has benefits but also limitations. Developers must still understand problems, write tests, and review generated code. Its usefulness varies based on tech stack, problem complexity, and developer experience. Coding assistants may sometimes reinforce bad practices

SMRTR provides this summary for quick context. The original article belongs to lobste.rs.

Read the original article
SMRTR Programming

Get the next batch of curated stories in your inbox.

This archive is built from SMRTR newsletter stories. Subscribe for hand-picked stories without the extra noise.

Related Stories

Browse Programming
ProgrammingAug 23, 2026

Rust Glancer

Rust-analyzer's memory bloat stems from treating all 6,666 dependencies equally — a tiered, IntelliJ-style backend could fix that.