A new frontier model, Apple opening Xcode to AI agents and a practical discussion on how many databases a company really needs.
Claude Opus 4.6 plans better and gets a 1 million token context window
Anthropic released Claude Opus 4.6, which according to the company plans more carefully, sustains agentic tasks for longer and works better in large codebases. It is the first Opus model with a 1 million token window, in beta, and leads evaluations such as Terminal-Bench 2.0 and GDPval-AA, which measures knowledge work in areas such as finance and legal. Pricing stays at $5 and $25 per million input and output tokens.
With a bigger context and longer tasks, it becomes practical to hand an agent the analysis of long documents or changes to large systems. Companies already using AI can revisit processes that used to require splitting work into small pieces.
How Dexpert helps: Artificial IntelligenceSources
- Anthropic: Claude Opus 4.6
Xcode 26.3 brings coding agents such as Claude Agent and Codex into the IDE
Apple announced Xcode 26.3 with support for coding agents, including Anthropic's Claude Agent and OpenAI's Codex. Agents can search the documentation, explore files, change project settings and check their own work visually through Xcode Previews, iterating through builds and fixes. Xcode also exposes its capabilities through the Model Context Protocol, so other compatible agents can be used.
Teams building for iOS can speed up tasks such as interface tweaks, refactoring and fixes without leaving the official tool. Set clear code review practices for what agents produce.
How Dexpert helps: Mobile AppsSources
- Apple Newsroom: Xcode 26.3 unlocks the power of agentic coding
Why to keep search, vectors, queues and time series in Postgres
Tiger Data, the company behind TimescaleDB, argues that most companies do not need a different database for each kind of data: Postgres already covers search, vectors, time series, queues and documents. The post points out that every extra database brings more backups, monitoring, security and points of failure. The new argument is AI agents: with a single database, an agent can create a test copy with real data, validate a fix and throw the copy away with one command.
Before adding a specialized database, check whether a Postgres extension covers the case. Fewer parts in the architecture lower operating costs and make testing, automation and the use of agents easier.
How Dexpert helps: Backend & APIsSources
- Tiger Data: It's 2026, Just Use Postgres
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.