This article was generated with the assistance of AI to help our volunteer team share valuable insights efficiently. While we've reviewed the content for accuracy, please let us know if you spot any errors or have additional insights to contribute!
Missed the talk? No stress!
This session wrapped up the trilogy with a deep dive into the realities, trade-offs, and exciting possibilities of building software that prioritises local data, offline capability, and resilience over constant cloud dependency.
Starting with a series of real-world cloud outage frustrations, from Slack going offline to AWS-related service failures, the talk explored a simple but increasingly relevant question: have we become too dependent on the cloud?
From there, Schalk unpacked the growing complexity of modern front-end development and argued that many of today’s challenges stem from trying to make cloud-based software feel local and instant. The session explored how Local First architecture flips that model around by treating the device as the primary source of interaction, while syncing happens in the background.
The session also included a live demo of an ambitious locally powered project management-style app built around CRDTs, Automerge, and sync-based architecture. What started as a “simple todo app” evolved into a surprisingly capable system with nested teams, projects, tasks, comments, attachments, search, and authentication-like behaviour; all without relying on a traditional central database.
Some of the topics covered included:
Why front-end complexity keeps increasing, even as tooling improves
The hidden costs and fragility of cloud-centric SaaS models
Why intermittent internet connectivity is a harder problem than no internet at all
How Local First architectures shift complexity from networking into replication and syncing
Using CRDTs and sync engines like Automerge to manage distributed state
Why Local First can simplify data fetching and state management compared to traditional client/server patterns
How modern Local First tooling begins to fulfil the original promise of GraphQL
Key Takeaways:
Local First isn’t about removing servers entirely. It’s about using them differently. Syncing and permissions remain important, but UI interactions and data manipulation happen locally first.
Much of modern front-end complexity comes from hiding latency. Optimistic updates, caching, invalidation, loading states, and retries all exist because the data lives somewhere else.
Replication is difficult, but specialised tooling is making it far more approachable. Libraries like Automerge and emerging sync platforms abstract away much of the heavy lifting.
Offline-capable software matters far beyond low-connectivity regions. Industries like mining, marine research, aviation, and field operations all benefit from resilient local-first systems.
The Local First ecosystem is still evolving. The space is messy, experimental, and rapidly changing — but that’s also what makes it exciting.
One of the most exciting moments of the session was the live demo of instant search and querying across hundreds of locally stored teams, members, projects, and tasks. All handled declaratively with minimal boilerplate and without traditional server roundtrips.
If you’re interested in the future of front-end architecture, offline-first applications, sync engines, or distributed state management, this session is well worth watching in full.
Watch the full session
👉 Watch the full talk on FEDSA’s YouTube channel to see concepts in action, including demos and deeper discussions around trade-offs and real-world use cases.
Join our Discord for discussion and updates, and subscribe to our monthly newsletter to stay in the loop.
Keen to contribute to the FEDSA community? Sign up here.
Otherwise, join our free Discord community to chat to other members!