It’s mind-blowing what agents can provide if they are fed with build systems’ structured data. I feel it’s time to have an alternative to logs as the only tool to understand the work a build system is doing.
#buildsystems #dev
New post: The New Bottleneck
AI agents generate code faster than build systems can compile it. What was once a problem for large orgs is getting democratized.
At @tuist we want to make developer productivity accessible to everyone.
https://pedro.fika.bar/the-new-bottleneck-01KCBMVAQ5RSQYKYNP6XNA5Q4B
Build systems need interoperability, not tool wars. Like shipping containers standardized logistics, we need standard caching contracts.
Current enterprise models bundle tech + service wrong. Compare to Supabase: one click, DB in any region.
Build caching should work the same. Select locations, get cache infra there.
Fabrik aims to be that standard.
https://github.com/tuist/fabrik/
Build system caching: the mechanics are similar across platforms.
Cache observability: that's where specialization matters. Working with Xcode means reverse-engineering Apple's undocumented formats to surface meaningful insights about cache performance.
At @tuist we're building Xcode-native caching with first-class observability, so iOS/macOS teams can optimize their builds with data, not guesswork.
Build system caching: the mechanics are similar across platforms.
Cache observability: that's where specialization matters. Working with Xcode means reverse-engineering Apple's undocumented formats to surface meaningful insights about cache performance.
At @tuist we're building Xcode-native caching with first-class observability, so iOS/macOS teams can optimize their builds with data, not guesswork.