It started well, though. It was an analytics tool with a developer-oriented mind; it was refreshing compared to the competition. But all good things do seem to come to an end, especially when it is a company. They went full weirdos in the last 2 years. AI just made everything worse.
PostHog was wonderful as an analytics solution, with its developer-first approach, good tools, and nice pricing.
At this point, I lost count of how many times in the last two decades I fell for it, although I'm more used to it now. Companies that grow into success and change. With the AI frenzy, PostHog also started going all-in, and not only that, but they seem to be exploring no-coding tools and whatnot. Supabase is another one that was cool, but now it is in the AI abyss.
Indeed, at this point, I'm the constant. Maybe I'm the problem here, and perhaps I should learn to accept the new AI overlords, give up and go full AI.
This one has yet to enter my head. I read Martin Fowler's related article* a few times, and it is sound... But I still can't see microservices being so "complex" to be so avoided early on. Especially in 2026, with the level of IaC and pipelines we have. I might be naïve, but I just don't get it.
Coupling is evil, as it is the main root of all the nastiest problems I have found in many projects I worked on. It is the main cause of "throw it all away and start greenfield" in my experience. I mean, it is important and needs to be done right, but it is yet evil.
To be honest, "It depends" is also exhausting. I have some defaults that I push until the breaking point, though I apply this more often on personal projects. DDD is an exception, as even professionally I will vomit DDD concepts as solutions for literally anything. "This could have been an aggregate", I will say like a mantra.
Given how chaotic the world of data analytics is (or logs... Or error tracking...), I imagine how this could be a better source of context without hallucinations...
I'm old generation and almost forgot for a while. GitHub was good even on their hands at the beginning, C# is amazing, TypeScript is amazing, wsl2 is game changer (which includes the change in Microsoft's position about linux), vscode is amazing, microsoft great increase in presence on opensource was nice (rushstack for example), etc...
But well, they still have the garbage side, which seems to be spreading again.
I disagree. Diagrams are a type of visual communication, and not everyone is good at translating things to visual. I open an excalidraw with clear concepts in my head, but nothing comes out of it. I try C4 or flow diagrams, and I spend an excessive amount of time refactoring them to end up mediocre anyway. Not just me, I know MANY developers that are amazing at explaining things but are mind-blocked when drawing simple circles and arrows.
Helping us navigate things we aren't good at has been one of the main selling points of AI.
Well, out of all the workflows I have seen, this one is rather nice, might give it a try.
I imagine if the context were being commited and kept up-to-date with CI would work for others to use as well.
However, I'm a little confused on the autocontext/globs narrowing part. Do you, the developer, provide them? Or you feed the full code map to flash + your prompt so it returns the globs based on your prompt?
Also, in general, is your map of a file relatively smaller than the file itself, even for very small files?