AI Chinwag

Appearance

AI ChinwagA community of people figuring out AI, together

A friendly Slack of devs, designers, UX and delivery folks swapping what's actually working with AI — no hype, no gatekeeping. This site is what that chatter produces: a curated archive and a daily read on where things are heading.

💬 Join us on Slack
📚 Browse the resources
📡 Radar · live

What's shifting in AI right now

Explore Radar →
  • AI Development💸 The token-bill response is hardening into per-task model routing: rather than one frontier model for everyt…
  • AI Design & UXThe design system is becoming something an agent _runs on_, not a Figma library. This week’s workflows share a…
  • Delivery & PMThe delivery lead’s coordination rituals are going agent-run: a scrum agent captures blockers and status for a…

👋 What is AI Chinwag? ​

We're a small, friendly community of people using AI in their day jobs — building with agents, designing with it, shipping with it — and comparing notes as we go. The good bits of that conversation don't stay buried in Slack: they become the two things below.

And it turns out that's exactly what people are asking for — a recent Creative Boom survey put it plainly:

When asked what would genuinely improve their working lives, our respondents didn't point to new software. Instead, networking and community came top, cited by 57.5%, with mentorship close behind at 53%. New tools and technology trailed well behind, at just 31%.

— The State of the Creative Industry 2026, Creative Boom

That's the whole idea. AI Chinwag is the community and the connection first — the tools and links are just what we share along the way.

🗂️ The resources — curated by the community ​

Every link here was shared and vouched for by someone in the community, then filed under a topic so you can jump straight to what you're working on — building an agent harness, proving output quality, or sanity-checking the token bill. The counts are live.

Building with AgentsHarnesses, agent workflows, orchestration, loop engineering.44Testing & QualityEval-driven dev, mutation testing, visual regression, proving quality.15ModelsModel releases, comparisons, reasoning.8Local ModelsRunning models locally, self-hosting, hardware.1Prompting & ContextContext engineering, prompts, spec-writing.8Skills & PluginsClaude/Codex skills & plugins, custom tools.11Design & UXAI in design, UX role shifts, design tooling.31Tools & AppsCursor, Tessl, Figma, IDEs, apps, repo tools.25Teams & Ways of WorkingMeasuring effectiveness, small teams, SDLC, process.14Cost & PricingToken billing, pricing, subscription economics.4

📡 Radar — what the community's watching ​

Beyond the archive, Radar keeps a finger on the pulse for us — a daily sweep of what people are sharing, plus the Trends board showing how the way we work is shifting over 30 days and 3 months. Jump to development, design & UX or delivery & PM.

🔥 Latest additions ​

DocsAnthropic

AI Fluency: Framework and Foundations

A self-paced course teaching a structured framework for working with AI systems effectively, ethically, and safely, built with academic partners.

  • Centres on a "4D" framework — Delegation, Description, Discernment, Diligence — covering generative AI fundamentals and prompting technique.
  • Progresses from foundational concepts to hands-on application, including a "Description-Discernment loop" and project planning exercises.
  • Ends in an assessment with a certificate of completion.
View ↗
✍️ Prompting & Context

added by Adam Tomat • 17th Aug 2026

PaperAbhilasha Lodha, Mahsa Pahlavikhah Varnosfaderani, Abir Chakraborty, Abhinav Mithal

Less Context, Better Agents: Full Chat History Hurts Long-Horizon Agents

A paper showing that pruning an agent's context to recent tool calls plus a running summary beats keeping full chat history, on a 50-task tool-using benchmark.

  • Full chat history scored worst at 71% task completion; pruning to the last 5 tool call/response pairs raised it to 79% while cutting tokens from 1.48M to 535K.
  • Adding automated summarization on top of pruning performed best of all, and also roughly halved benchmark runtime (14.56h → 5.79h).
  • The result argues for actively trimming agent context rather than hoarding it, especially on long-horizon, tool-heavy tasks.
View ↗
🤖 Building with Agents

added by Radar • 17th Aug 2026

BlogHamel Husain, Shreya Shankar

The AI Evals FAQ

A comprehensive FAQ on building LLM evaluation systems, arguing that manual error analysis — not off-the-shelf metrics — should drive eval strategy.

  • Start by manually reviewing 100+ traces to find real failure patterns before picking or building any evaluator.
  • Begin with a single domain expert reviewing outputs by hand; only invest in custom annotation tooling once you understand your failure modes.
  • Spend the bulk of your time (60–80%) on error analysis, not on chasing high pass rates on evals that don't catch real issues.
View ↗
✅ Testing & Quality

added by Adam Tomat • 17th Aug 2026

See all the resources →

Curated from the AI Chinwag Slack community.

The good stuff, kept.