AI Engineering
Practical ideas and tools for building with AI.
About this list
This reading list brings together items from the sources selected for AI Engineering. Open each item at its original source for the complete article, episode or video. This is a selection of current coverage, not a complete archive.
Latest stories
Newest first
Browser Use vs Firecrawl: Use Cases, Pricing & Performance
Compare Browser Use vs Firecrawl on use cases, pricing, authentication, and measured performance. Learn which fits your workflow and when to use both.
Apertus 1.5 sets the bar for open, reproducible evaluation reporting
Apertus 1.5 is the first model release to submit all of its evaluation results to Evaluation Cards and Hugging Face Community Evals, with enough detail for anyone to reproduce them.
V0.31.0 - gpt-6-luna as the default OpenAI model
What's Changed Publish zepai/graphiti on v*.. tags like MCP by @mehulp93 in #1862 Add pevans to the CLA allowlist [ZEPAI-3570] by @pevans in #1883 Update root uv.lock to fix Dependabot alerts [ZEPAI-3570] by @pevans in
Release v0.22.1 · NousResearch/hermes-agent
Hermes Agent v0.22.1 (rc.5-v0.22.1) Release Date: October 14, 2026 This is a patch release, with a handful of fixes for the Windows Store, plugin updates, and more. ✨ Features Agent): add nemotron
Release rc.5-v0.22.1 · NousResearch/hermes-agent
{“attempt”:5,“autopublish”:false,“claimEpoch”:1792016072,“commit”:“29
Gh-aw/work-queue/6864335f70d503a8bc846426012bd22b0a9fb0f8: Add trusted dispatch and fix work-queue history amplification (#68827)
Add trusted work-queue dispatcher command Activate verified queue dispatchers through the GitHub CLI while preserving scheduler-selected immutable Claims. Require exact native run correlation and revision checks, and propagate dispatcher wait failures. Co-authored-by: Copilot App 223556219+Copilot@users.noreply.github.com Add end-to-end work queue performance suite Exercise trusted controls, publication transports, deployment refresh, native bindings and idempotent recovery with deterministic API budgets.
Release v1.6.0 · strands-agents/evals
v1.6.0 Auto-drafted from commits in v1.5.0..v1.6.0, grouped by conventional-commit type. Edit on the release page after publish if you want a polished writeup; the canonical release notes live on the website. 🚀 Features feat(redteam): reframe attacker prompts to survive aligned attacker models (#298) (ac52c39) feat: add StructuredOutputSimilarity evaluator (stickler extra) (#400) (
Tree Appreciator
Tool: Tree Appreciator Another game experiment with Claude. This one used the default Claude iPhone app (chat mode, not Claude Code mode) where I prompted Claude Opus 5.5 Max with: Build a three js game called "Tree appreciator" - it shows you two algorithmically generated trees and the game mechanics is about appreciating the differences between them. Think hard about how to make sure it is both fun and beautiful. Somewhat embarrassingly given my recent talk, I've found that telling Opus 5.5 "make sure it is fun" works uncomfortably well.
Release v0.91.14 · github/gh-aw
What's Changed Fix CI parser dependencies and linter metadata by @pelikhan in #68807 Require explicit multi: true for multi-repository checkout by @pelikhan in #68816 Harden OTLP credential redaction and eval sandbox isolation by @pelikhan in
Benchflow 0.7.9
What's Changed Start BenchFlow 0.7.9 development by @bingran-you in #1129 feat(agents): native Google Antigravity CLI agent (antigravity / agy) by @bingran-you in #1132 fix(codex): hand Codex the bare model slug and a codex that knows it (
Release v1.0.22 · github/copilot-sdk
Other changes improvement: [All SDKs] update bundled Copilot CLI dependency to 1.0.98-2 (no SDK API changes) Full Changelog: v1.0.21...v1.0.22 Warning Firewall blocked 1 domain The following domain was blocked by the firewall during workflow execution: github.com To allow these domains, add them to the network.allowed list in your workflow frontmatter: See
Release v1.23.0-r1 · OpenHands/OpenHands
fix(security): rebuild agent-canvas 1.23.0 with patched dependencies
Release langgraph-checkpoint-postgres==3.2.0 · langchain-ai/langgraph
Changes since checkpointpostgres==3.1.3 release(checkpoint-postgres): 3.2.0 (#9313) feat(checkpoint-postgres): add prune to PostgresSaver and AsyncPostgresSaver (#9290)
Release v4.3.1 · seojoonkim/memkraft
MemKraft 4.3.1 Shared owner-ledger patch. No migration required. MemKraft(base_dir=...) resolves the owner ledger from MEMKRAFT_OWNER_DIR when it is set, the same directory the Hermes provider already uses. Before this, a script that recorded an owner saying with a plain store wrote it to /owner/sayings.jsonl, where owner_saying_search in Hermes never saw it. An explicitly assigned owner_dir still takes precedence; without the env var the per-store default is unchanged.
V0.91.13: Add allowlisted dynamic repository checkout lists (#68593)
Add allowlisted runtime repository checkout lists Co-authored-by: Copilot App 223556219+Copilot@users.noreply.github.com Enforce runtime checkout allowlists and report rejected repositories Require allowed-repos for repository expressions, keep wildcards exclusive to scoped allowlists, and propagate hard checkout rejections into agent failure issues. Fail compilation instead of silently dropping invalid main or imported checkout configurations. Co-authored-by: Copilot App 223556219+Copilot@users.noreply.github.com
Release v2.1.300 · anthropics/claude-code
What’s changed Added virtualText to a mod’s prompt decorations: text drawn in the prompt box that never enters the draft or the submitted prompt Added cache_creation_5m_tokens and cache_creation_1
Uber’s AI infra, from the inside
Uber Distinguished Engineer Uday Kiran Medisetty shares how Uber builds internal developer tools, integrates AI agents into engineering workflows, and measures their impact on productivity.
AI is mass decompiling games. That’s both awesome and scary
AI is mass decompiling games. That's both awesome and scary. Indie game developer Popcar talks about the recent wave of AI-powered decompilation of existing games, enabling weird mashups between existing games, and producing a boom in game preservation projects to get old games working on new hardware (often in a browser using WebAssembly). I like how this piece captures both the positives and the negatives of this new development: Most people already realized the scary part by now.
Showing 20 of 186 articles
Know a source we’re missing? Tell us about it