A mobile architecture decision that AI changed, a managed way to put agents into production and a new path to scale PostgreSQL.

Shopify swaps React Native for Swift and Kotlin because AI made native code cheaper
Shopify, which adopted React Native in 2020 to write each feature only once, announced it will go back to building its apps in Swift and Kotlin. The company says the bet worked very well and React Native remains an excellent framework, but coding models improved so much that building the same feature on both native platforms no longer costs what it used to. So it reevaluated the decision from first principles, since one of its assumptions had changed.
The choice between native and cross-platform apps needs to be revisited in light of what AI changed in development costs. Companies planning a new app or a rewrite can redo that math before deciding.
How Dexpert helps: Mobile AppsSources
- Shopify Engineering: Native is now the future of mobile at Shopify

OpenAI's Agents API gives applications managed access to the Codex harness
OpenAI launched the Agents API, which gives applications access to the Codex harness as a managed service. OpenAI handles sessions, orchestration, context compaction and recovery, while the application provides the tools and chooses where the agent runs, including OpenAI-hosted sandboxes where it runs code, edits files and connects to MCP servers. Examples include data analyst agents with read-only SQL, Slack bots and document reviewers.
Putting agents into production requires less in-house infrastructure, which shortens the path from prototype to product. The work shifts to the tools, data and permissions the agent will have.
How Dexpert helps: Artificial IntelligenceSources
- OpenAI Developers: Agents API

PlanetScale launches Neki, sharded Postgres without changing databases
PlanetScale put Neki in preview; it spreads a Postgres database across many machines while keeping real Postgres on every shard, with a primary and at least two replicas across three availability zones. The application connects to a router over the standard Postgres protocol, so drivers, ORMs and connection strings keep working, and schema changes, upgrades and resharding run without downtime. You can start unsharded and split the database only when it outgrows one machine.
Companies whose Postgres is nearing the limit of one machine get a way to grow without rewriting the application or giving up extensions. Plan the shard key before growth becomes urgent.
How Dexpert helps: Backend & APIsSources
- PlanetScale: Introducing Neki
Want to put one of these to work in your company?
Dexpert builds custom systems, apps, integrations and AI solutions. Tell us what you need and get an estimate, or take the free digital maturity assessment.