AI referrals up 38% in 30 days. Google Organic down 12%. That changed what we planned to publish next. I run Angular Space, where traffic now comes from a mix of Google, AI search, social and direct referrals. We already had Plausible for analytics and Search Console for search performance. Another dashboard was the last thing I wanted. At first, we used Viktor to pull the numbers together. Useful, but still just reporting. The shift came when we gave him ownership of the weekly acquisition review. Now Viktor compares Google Organic with AI referrals, looks at which pages are gaining or losing traffic, spots meaningful changes and brings the team the few things actually worth acting on. This week, AI referrals were up 38% while Google Organic was down 12% over the same 30-day period. More importantly, he showed us which content was gaining visibility through AI while slipping in traditional search. That changed our editorial priorities before the next planning session. We still decide what gets updated, published or left alone. Most teams do not need another analytics dashboard. They need someone to own the interpretation. A copilot helps you work. An AI employee works when you don't. Hire @viktor_com for your team. $100 in credits included, no card. Full link in first comment. #AIemployee #Analytics #ContentOps Paid Partnership

Angular gives you plenty of ways to render dynamic content But sometimes the “official” way is not the nicest one New article drop at Angular Space by @Waterplea 👇 Polymorpheus is one of those tiny libraries that solves a very specific problem so well that, once you start using it, you wonder why #Angular doesn’t work like this by default Alex Inkin explains how it works, where Angular’s built-in options start getting awkward, and why Taiga UI uses Polymorpheus everywhere url in comment below 👇

Worst excuse for MFEs I heard was “we need separate deployments” You can absolutely have separate deployments without MFEs…

If you haven't noticed 4 new articles landed on Angular Space recently :) - Lessons from speaking at conferences by Jarosław Żołnowski - The best API integration platforms for SaaS apps and agents by Emmanuel Oyibo - Whether AI made developer knowledge less valuable by Paweł Ostromecki - Handling global errors in modern zoneless Angular by Jarosław Żołnowski Different topics, same goal: practical insights you can actually use Which one are you reading first?

The idea that you need Python to get into AI is keeping frontend developers on the sidelines…. You don’t need Python to build on top of LLMs You already know JavaScript/TypeScript, APIs, async flows, state, streaming and product UX Add model APIs, tool calling, structured outputs, RAG and evals You’re much closer than you think

Is this what work looks like next? No app to find. Hold one physical button, speak, and Viktor gets the work moving in your team's Slack or Microsoft Teams - all eyes-free. Maybe the chair was never part of the job. Just part of the interface. One button. @viktor_com does the rest. $100 in credits included. Paid Partnership

People keep saying Angular is dead. Meanwhile, it’s being used by: Google Cloud Gemini Deutsche Bank Citibank Capital One NatWest Delta Air Lines BMW UPS Epic Games Battle net Bitwarden ClickUp IBM Adobe HP Amtrak JetBlue Ryanair Spirit Airlines United Airlines Aer Lingus Crunchbase Otter AI H&R Block Franklin Templeton Pearson VMware Universal Orlando US Homeland Security European Bioinformatics Institute + many many more Angular isn’t dead. It’s just busy running things outside your Social Media bubble

No idea how to get there? AI Assisted Engineering is a craft. I’m teaching companies how to setup systems that can help you control and maximize output quality while saving on tokens. Check it out: aicontextengineering.eu/

Most companies have better governance for CSS changes than for the instructions controlling their AI coding agents. I would never ship production code with: - no owner - no review - no validation - no staged rollout - no rollback Yet this is exactly how most companies ship AI instructions... One CLAUDE.md here. Cursor rules somewhere else. Another AI tool running on completely outdated conext. Prompts copied across dozens of repositories. Authored by everyone. Owned by no one... Sounds familiar? Nobody can confidently answer: - Which version reached the agent? - Who approved it? - Did it improve the output? - Can we roll it back safely? We call this AI adoption. It is actually uncontrolled production configuration. That is why, together with @JarekZolnowski, I founded => ACE - AI Context Engineering <= We built the system we believed engineering teams were missing. A context supply chain for AI coding agents: - Author once. - Validate before release. - Roll out in stages. - Roll back safely. - Measure exactly what reached the agent. Now we are opening registration for the first live ACE workshop! - One hands-on day. - Complete source code included. - The system and know-how transferred to your team. Website with details in comments👇 Special thanks to @gregorojstersek , Dawid Perdek and Przemysław Sokołowski, Lukasz Boduch, @ostromeckyp, @pawelkubiakdev for reviewing the solution, challenging our assumptions and helping us shape it into something engineering teams can actually use. #AI context is production infrastructure. It is time we treated it that way.

Dario tried AngularJS back in the day and on that day he decided that he hates software engineers

Angular Space is picking up new wind! We’re excited to welcome 👉 @coderabbitai 👈 as our NEW sponsor and continue growing the community together 🥳🎉 I have been using CodeRabbit for AI reviews and I Am honestly very impressed how useful it is with accurate hints Perfect match for Angular Space!

If you need tricks to get what you want, the model probably doesn’t really understand you This felt different :) Angular architecture visualized Made with Uni-1 by @LumaLabsAI

Feels less like prompting and more like describing a system That’s new

Recorded this 👇 Had an idea for a presentation “how AI influenced modern coding” Usually that means building slides from scratch And all AI solutions I tried so far sucked This time I used @GammaApp and just described it. Got a few options, picked one and it was basically ready No templates, no real “designing” And honestly this looks insane! I’ve never made a presentation that looked THIS good design-wise Im stunned how good the actual outcome is...

2026 reality check ✅ We still force candidates to hand-write perfect syntax on a blank screen under timer pressure… while actual engineering is now prompt => critique => iterate => ship Engineering leaders keep telling me that AI has broken our ability to assess real skills with old methods Even before, old methods have been checking the wrong things - but today they are completely useless If senior devs are spending their day directing agents instead of typing boilerplate, why are we still auditioning people like it’s 2012?

Even durning the AI revolution where people are moving to LLMs for knowledge Angular Space keeps growing year by year! Validated long form content is still desirable 💪💪💪