Show HN: Lantern – desktop app that maps your repo and reviews AI-written PRs
A desktop app that maps out your repository and reviews pull requests written by AI.
Why this: Covers review · Published within 6 hours
16 items · All topics →
A desktop app that maps out your repository and reviews pull requests written by AI.
Why this: Covers review · Published within 6 hours
This Hacker News thread asks developers to place themselves on a scale from autocomplete-only use to fully hands-off agents, and share how far they have actually gone with AI coding.
Why this: Covers LLM and Agent · Published within 24 hours
Audible's new AI-made audio previews sum up a book's themes and what readers praised, so listeners can tell whether they'll like it before sampling it.
Why this: Covers review · Published within 24 hours
This dataset uses real-world privacy fixes to test whether coding agents catch the privacy defects that regulations like the GDPR require software to avoid.
Why this: Covers Agent and benchmark · Relevant to this topic
Model annealing borrows the idea of review and consolidation, letting a small labeled dataset be reused through training instead of merely memorized, so the model learns more from less data.
Why this: Covers model and review · Relevant to this topic
This study shows that when AI judges grade clinical AI output, the evaluation itself changes clinical workflows and becomes an intervention.
Why this: Covers model and review · Relevant to this topic
OpenFold3-preview can now be deployed from SageMaker JumpStart inside your own AWS account, forecasting the 3D structures of protein, DNA, RNA, and ligand complexes, with async inference and multi-GPU support.
Why this: Official source · Covers inference and deploy
This release adds the Ripple model to the MLX runner — model, renderer, parser, and import registration — plus fused MLP activation, value blending, and attention gating.
Why this: Covers model and review · Relevant to this topic
Merge puts GPT-6 Astra, ChatGPT, and Codex to work on code review and marketing support — and even lets AI help plan executive dinners and prep sales updates.
Why this: Official source · Covers GPT and Tools
The post sets out three principles for human-in-the-loop work and argues that keeping people in the review loop is essential once agents run in production.
Why this: Official source · Covers Agent and Workflow
OpenAI is buying Pierre Computer Company to give Codex stronger tooling for writing, reading, and reviewing code.
Why this: Official source · Covers Tools and review
Qodo feeds company knowledge and coding standards into Codex, so developers can spot cross-repo P0 issues before they even open a pull request.
Why this: Official source · Covers review
Ropes & Gray is building AI for transactional diligence together with OpenAI — its lawyers check the grouped documents, summaries, and reports first, before advising clients.
Why this: Official source · Covers review
At Greptile, GPT-6 Astra tests user workflows, Daybreak hardens security, and Codex cuts down on rework.
Why this: Official source · Covers GPT and Workflow
CodeRabbit is using GPT-6.1 Sol to look into security risks in complex codebases, with its security agent helping developers go over findings and suggested fixes.
Why this: Official source · Covers GPT and Agent
Cloudflare puts OpenAI Daybreak to work investigating vulnerabilities and drafting fixes for review, while developers build faster and write better code with Codex.
Why this: Official source · Covers review