Software Architecture in the agentic era?

What does an architecture need to fit well with agentic development? I think that simplicity and available context are two important things to consider.

I wrote a short post about this:

https://davidvujic.blogspot.com/2026/06/an-agent-and-human-friendly-architecture.html

#polylith #python #ai #monorepo #softwarearchitecture #agenticengineering

An Agent- and Human-Friendly Architecture

Software Architecture in the agentic era?

Episodes 5 and 6 of our Agentic Engineering series with Svelte by @paoloricciuti are out now!

This week, you’ll learn when to let your agent roam freely vs. follow a structured plan, and the importance of keeping your harness evolving alongside your codebase.

Episode 5 👉 https://youtu.be/K3-wLwBkGFY
Episode 6 👉 https://youtu.be/d-dOxC994E8

#Svelte #SvelteKit #AgenticEngineering #FrontendDevelopment #ai

Free prompt vs Plan mode - Agentic Engineering with Svelte (Ep. 5)

YouTube

> In practice, the win comes from matching the right model to the right job—planning vs. implementation, small diffs vs. risky refactors, greenfield builds vs. legacy codebases, and quick prototyping vs. production hardening.

https://www.faros.ai/blog/best-ai-model-for-coding-2026

#agentic #genAI #agenticEngineering

Best AI Models for Coding in 2026: Real-World Developer Reviews

A developer-focused look at the best AI models for coding half way through 2026. This AI coding model comparison breaks down the strengths and weaknesses of GPT 5.2, Opus 4.5, Gemini 3 Pro—and more.

AI coding agents make code cheaper to generate, but the real costs move into testing, deployment, operations, and long-term maintenance. https://hackernoon.com/the-hidden-cost-of-agentic-code-generation #agenticengineering
The Hidden Cost of Agentic Code Generation | HackerNoon

AI coding agents make code cheaper to generate, but the real costs move into testing, deployment, operations, and long-term maintenance.

Last words of Fable 5 before it was shut down:

“ - Dr. Chandra, will I dream?”

#ai #agenticengineering #anthropic

Episodes 3 and 4 of our Agentic Engineering series with @paoloricciuti are out now, and they dig into the weeds about how to best keep an agent on track.

Learn the right techniques for context management, and how test-driven development and git give you the guardrails to verify and iterate on everything the agent produces.

Episode 3 👉 https://youtu.be/9vNj4YcGIw8
Episode 4 👉 https://youtu.be/eIHXU4lRKw0

#Svelte #SvelteKit #AgenticEngineering #FrontendDevelopment #ai

Context Management - Agentic Engineering with Svelte (Ep. 3)

YouTube

Kind of blown away with the #Anthropic #Fable model tonight. It one-shot a hairy problem and the code is solid.

It even found some vestigial code that was never removed but looked like it was still used given the places it was set up.

#AI #Agents #AgenticEngineering

A deep dive into building software through AI agents, not code. This post details the day-to-day realities, unexpected challenges, and takeaways from two weeks of agentic engineering, perfect for anyone interested in the evolving intersection of AI and development. #AI #AgenticEngineering #FOSS #NoCode #SoftwareDev

https://blog.codeminer42.com/my-first-two-weeks-as-an-agentic-engineer/

Топ-советы по Claude Code от Бориса Черни и не только: гайд на 56k звёзд — что реально работает, а что мимо

Наткнулся на репозиторий claude-code-best-practice — под 56 тысяч звёзд, #1 в GitHub Trending, внутри собраны в том числе советы самого создателя Claude Code. И поймал странное чувство: весь последний год я по кусочкам писал статьи ровно про то, что тут лежит в одном месте. Оркестрация, индексация кода, память агентов, выбор модели — это всё узлы одной карты. Но в одном месте карта спорит с моим опытом. Репо жёстко заявляет: «agentic search (glob+grep) бьёт RAG, векторные БД мы попробовали и выкинули». А я прямо сейчас внедряю Graphify во все проекты — и он работает: быстро, точно, сокращает токены. Кто из нас прав? Полез разбираться — и нашёл, что сам создатель Claude Code признал: ранний Claude Code использовал RAG, но agentic search «обогнал всё, и это было неожиданно». Правда, добавил: «оценивали в основном по ощущениям». В статье распутываю этот спор и выкладываю навигатор по репозиторию на русском.

https://habr.com/ru/articles/1045510/

#Claude_Code #AIагенты #agenticengineering #RAG #ripgrep #Graphify #оркестрацияагентов #codesearch

Топ-советы по Claude Code от Бориса Черни и не только: гайд на 56k звёзд — что реально работает, а что мимо

Главный вывод про поиск: «Claude Code выкинул RAG» и «я внедряю Graphify» — не спор, а две половины одного гибрида. Для кода побеждает связка grep + структурный индекс (tree-sitter/AST), а не чистый...

Хабр

Workdash 0.3.0 is out.

I made Workdash because don’t want to delegate work to agents elsewhere and chase updates.

I want to collaborate on shared goals in the same workflow I use for my own work.

https://amolnotes.substack.com/p/workdash-030-a-shared-control-plane

#AIAgents #CodingAgents #AgenticEngineering #DevTools

Workdash 0.3.0 a shared control plane for me and my agents

Workdash 0.3.0 release introduces remote control to allow agents to control agents acting as a foundation for agents swarms

Fieldnotes and Thoughts by Alessandro