Loading…
Latest reads
Every post, summarised
Grounded summaries of engineering blogs, newest first. Each links to the original.
- 4,671 posts
- 20 companies
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
Latest posts
Integrating Claude Code with Auth0 APIs
This tutorial explains how to integrate Claude Code with Auth0’s Management API using the Auth0 MCP Server, enabling AI-driven authentication configuration management from the terminal. The setup uses the OAuth 2.0 Device Authorization Flow to obtain scoped access tokens stored in the OS keychain, avoiding static secrets that risk exposure in source control. Developers initialize the MCP server with explicit scopes like read:clients or update:actions, then register it with Claude Code via the claude mcp add command, with credentials never stored in plain text. A CLAUDE.md file provides project context and constraints, guiding the agent to respect environment-specific configurations and security policies such as never modifying production callback URLs without confirmation. The tutorial demonstrates refactoring JWT validation code in a single terminal session, with human-in-the-loop confirmation ensuring safe agentic changes and all auth-critical paths undergoing manual PR review.
The making of the Yelp Assistant UI
Yelp Assistant is the first large-scale consumer LLM product at Yelp, connecting users with local professionals via chat. To iterate quickly across iOS, Android, and web while avoiding long app-store update cycles, the team chose a hybrid architecture: the top bar and input bar are native, while the chat content area is server-driven using Yelp’s CHAOS framework. The view and data are separated — the client fetches a CHAOS view template once, then conversation data hydrates that template on each turn. The backend sends MessageEvent objects, quick replies, and client actions (such as DisableChat) to control the UI. This hybrid approach enabled rapid iteration, but required building new CHAOS components for text input and network requests, and posed cross-platform consistency challenges.
Introducing IdentiFlows.dev: A Visual Reference for Identity Flows
Developers often struggle to quickly find accurate sequence diagrams for OAuth 2.0 and OpenID Connect flows. IdentiFlows.dev is a new website that provides a visual reference for common identity protocols, organized into four categories. Each flow page includes an introduction explaining the flow’s purpose, a sequence diagram showing the full exchange, and a step-by-step breakdown of each stage, with an interactive mode to walk through the steps. The initial release covers eight flows, including OAuth 2.0 Authorization Code with PKCE, the recommended login flow for most modern apps, SAML 2.0 SP-Initiated SSO, DPoP-Bound Access Token requests, and OpenID Connect Discovery. The project is an early release with plans to add more flows like Client Credentials, Device Authorization, and WebAuthn, and the author invites feedback to guide development.
The Intelligent Data Platform for the AI Era
MongoDB announced the largest platform expansion yet, including MongoDB 9.0, Atlas Infinite, and Atlas Agent Engine. MongoDB 9.0 delivers up to 2x higher throughput, 35% faster reads, and 30% faster updates over version 8.0. Atlas Infinite is a new deployment option that allows storage and compute to scale independently, with time to scale more than 96% faster and each shard holding 10x more storage. Atlas Agent Engine provides built-in memory, retrieval, identity, and governance for production agents, operating on the same data platform as operational data. The company argues that as AI moves from generating answers to taking actions, the data layer must provide live, fresh, and retrievable operational data rather than separate AI infrastructure.
Your Database Just Became Your AI Control Point
Agentic AI transforms the database from a passive record store into an active control point that serves real-time context and enforces policy at the moment an agent acts. According to IDC’s April 2026 survey of 715 organizations, 30% already run agentic AI in production and another 50% have established spending plans. Real-time data delivery is now baseline: 64.6% of enterprise use cases require real-time or near-real-time delivery, and 91.2% of streaming scenarios need end-to-end latency under 200 milliseconds. IDC identifies four requirements for agentic AI: a real-time operational foundation for unstructured data, elastic scale for bursty agent traffic, accurate always-current retrieval via integrated vector databases, and a governed runtime with durable memory and identity-bound policy. By 2029, IDC predicts 60% of enterprise data platforms will unify transactional and analytical workloads in hybrid processing architectures to power agentic AI.
Manufacturing data and AI: Connecting the product value chain
Manufacturing data is fragmented across systems such as MES, SCADA, QMS, ERP, and PLM, so answering a cross-stage question like whether a scrap spike stems from a supplier lot, machine setup, or process drift requires manual tickets, exports, and reconciliation. Databricks connects the product value chain by pooling or federating these sources using Lakehouse Federation and zero-copy Open Sharing, with Unity Catalog governing both mirrored and federated data under a single permission model and lineage. Lakeflow orchestrates the refinement of raw data into trusted Bronze, Silver, and Gold tables, while Genie and Agent Bricks enable natural-language queries grounded in governed business semantics. A quality engineer can thus ask about a defect’s history across plants and get answers that join MES, supplier, and QMS records without migrating every system. The result is that traceability becomes a query, transforming weeks of manual investigation into minutes of analysis across the entire value chain.
Lakebase Search: State-of-the-art full text and vector search for Postgres
Traditional OLTP systems like Postgres face challenges meeting the low-latency, high-accuracy search demands of AI agents, often requiring standalone search engines with ETL pipelines. Lakebase introduces lakebase_vector and lakebase_text, two extensions that bring scalable approximate nearest neighbor search and BM25 full-text search directly into Lakebase Postgres. lakebase_vector overcomes pgvector’s memory-bound HNSW index by using hierarchical IVF clustering and binary quantization, achieving 71-millisecond P99 latency at 97 percent recall on the 100-million-vector LAION benchmark. The extensions decouple storage from compute for pay-per-use scaling, enable fast parallel index builds, and support native hybrid search that fuses vector similarity with keyword relevance in a single SQL query. Early adopters like Conexiom reduced infrastructure costs by three times while achieving five times higher throughput, and the service is generally available on AWS and Azure.
How Databricks rolls out frontier models to 14,000 employees on Day 1
Databricks provides employees Day 1 access to new AI models using Unity Gateway for governance and cost management. The company uses a three-step pipeline: immediate experimental availability, per-user budget constraints, and data-driven promotion decisions. Unity Gateway CLI distributes model configurations to employee laptops, tagging new models as experimental. Per-user budgets include a monthly maximum, daily runaway limit, quality frontier budget, and experimental budget to control costs. Within three days of the Opus 5.5 and GPT-6 Sol launches, Databricks gathered benchmark, user feedback, and cost data to confirm they were on the efficiency frontier and promoted them, with Opus 5.5 showing a 29% cost reduction per session over Opus 4.8 and GPT-6 Sol showing a 48% reduction over GPT-5.6 Sol.
Engineering Multi-Agent AI Teams That Build and Test Themselves
Salesforce Marketing Cloud’s AI Delivery Lead Sohini Arya and team built Agent Designer, a system that automates the creation and testing of multi-agent AI teams. The orchestrator cannot write agent files itself; instead, seven specialist agents analyze architecture, failure modes, and compliance before the orchestrator synthesizes findings for human review. The system transforms an engineer’s request into a tested multi-agent team in 15-30 minutes, down from 2-4 hours manually, and helped approximately 200 engineers adopt the manager-of-agents model. Governance challenges included enforcing write-scope boundaries via prompts, reconciling independently produced design conflicts, and applying topology-specific budget formulas that account for flat, swarm, or team structures. Verifier agents return structured PASS/WARN/FAIL blocks, and self-healing is limited to two fix attempts and a $3 ceiling to prevent runaway repairs, with the orchestrator’s 500-turn budget divided across analysis, approval, generation, testing, and retries.
Why model versioning is not enough for production AI
Model versioning alone is insufficient for production AI because changes in inputs, preprocessing, prompts, retrieval, tool contracts, and serving settings can alter behavior independently. The post proposes a versioned release manifest that captures all interdependent components, including model revision, prompt revision, retrieval index revision, embedding revision, pipeline revision, and runtime revision covering token limits, batching, and timeouts. This manifest enables atomic release testing and rollback, with an evaluation gate that tests the full retrieval-generation path using deterministic checks and semantic rubrics on a versioned dataset. Practical steps include canary deployments that compare candidate and control signals, with explicit stop conditions and rollback procedures that restore compatible dependencies rather than just older model weights. The post concludes that the smallest useful implementation is a manifest, evaluation job, load test, release-aware traces, and a rehearsed rollback, and asks whether the on-call engineer can identify and restore a compatible known-good version.
How to roll out Genie One: A step-by-step enterprise playbook
Genie One is an AI coworker that lets business users ask questions in plain language and receive cited answers grounded in governed data. The post outlines a phased rollout playbook: Phase 0 foundations include choosing one team, defining metric views for 10-20 numbers, creating Pages for concepts, and establishing governance with Unity Catalog. Phase 1 runs a 4-6 week pilot with 5-10 users, scoring Genie against 25-50 real questions and logging every incorrect answer to fix root causes, with exit criteria including accuracy holding steady and users reaching for Genie over their old workflow. Phase 2 expands to 2-3 adjacent domains with shared ontology definitions and a lightweight review workflow. Phase 3 aims for organization-wide adoption with federated ownership, MCP server governance, and embedding Genie in Slack, Teams, Excel, or Google Sheets, with a realistic timeline of 1-2 weeks for foundations to 1-2 quarters for full rollout.
Is Your “Human-in-the-Loop” Actually Slowing You Down? Here’s What We Learned
Human-in-the-loop (HITL) frameworks integrate human judgment into automated AI workflows to improve reliability and safety, but without careful design they can create bottlenecks. The post reveals key trade-offs: more automation increases speed and reduces cost but raises risk, while more human oversight boosts accuracy and safety but adds latency and cost. Common failure modes include over-reviewing trivial cases, lacking clear trigger logic, and expecting human reviewers to scale linearly. The proposed solution is a tiered HITL system applying the Pareto principle: automation handles 85% of high-confidence decisions using confidence calibration, with three tiers addressing the remaining cases under 3 ms latency, asynchronous review within hours, and real-time oversight with 30-60 second windows. Success requires careful routing, reviewer tooling, and closing feedback loops across seconds, days, weeks, and quarterly cycles to continuously improve model performance and reduce human involvement over time.
Can your Postgres survive a bad query?
Postgres memory tuning is difficult because work_mem applies per operation, not per query, and parallel workers multiply memory usage without a query-wide cap. Some executor structures, such as those in recursive queries, have no disk-backed fallback and can grow until the backend runs out of memory. A benchmark using a recursive query that exhausts memory tested ClickHouse Managed Postgres, Cloud SQL, PlanetScale, and Amazon RDS on 16 GiB instances. ClickHouse Managed Postgres enforces a memory ceiling at 57% of host RAM, rejecting queries that exceed it, while RDS allows backends to consume nearly all RAM until the host crashes. Cloud SQL and PlanetScale enforce similar caps, but ClickHouse Managed Postgres’s ceiling is lower, a deliberate tradeoff that returns clean query failures instead of forcing every client through crash recovery.
The 24 most commonly misunderstood marketing data terms
Marketing and data engineering teams often interpret the same data terms differently, leading to campaign rework and misaligned expectations. The guide covers 24 commonly misunderstood terms organized around five practical questions about campaign execution, including where data comes from, who a customer record represents, and what makes an audience ready. For example, ‘customer’ might mean a person to marketing but an entity at a specific grain like account or household to engineering, while ‘real time’ can mean soon enough not to feel broken to marketing but sub-second streaming with always-on architecture to engineering. The post provides paired definitions for each term as conversation starters to surface assumptions and agree on meanings before work begins. It also recommends documenting business definitions in Unity Catalog, connecting agreed meanings to datasets or workflows, and using them for AI agents to produce consistent results.
Next.js applications, powered by Vite: introducing Vinext 1.0
Vinext 1.0 graduates from an AI experiment to a production-ready framework that lets Next.js applications run on Vite, supporting both Pages and App Router. It achieves over 99% test compatibility for customer-requested features, excluding cache components, and includes support for pre-rendering, cache warming, and observability with OpenTelemetry and Sentry. Cache warming moves page prerendering from the build machine to Cloudflare’s network by deploying a new Worker version to zero traffic and requesting pages to populate the cache before promotion. The framework supports React Server Components, Server Actions, API routes, middleware, and shared caching functions across both routers, with first-class runtime support for Cloudflare Workers and bindings like image optimization and hyperdrive. Migration requires only two commands, and the project is open source with a nightly Next.js end-to-end test suite and community-driven issue tracking, also supporting output: ‘export’ for fully static sites and background revalidation.
Introducing cf: the agentic CLI for the entire Cloudflare API
Cloudflare introduced cf, a new CLI that provides access to the entire Cloudflare API surface of over 3,000 operations, far exceeding Wrangler’s roughly 280 commands. Cf is designed primarily for AI coding agents, which account for nearly half of Wrangler usage, and defaults to JSON output with condensed formatting to minimize token consumption. The CLI includes a natural language search command, cf cli search, that helps agents discover the right operation without bloating context, and uses TypeScript-based configuration via cloudflare.config.ts for type safety. Cf is generated by Forge, Cloudflare’s internal pipeline that produces CLI commands from OpenAPI schema annotations, enabling consistent and comprehensive API coverage. The open beta is available globally via npm, and cf adopts Vite as its default local development server, with a migration path from Wrangler and maintenance support for 18 months after beta ends.
How fast is the web? Explore billions of real-user measurements with BEACON
Cloudflare has released the BEACON dataset, an anonymized collection of billions of real-user performance measurements from 10,000 of the largest websites. The dataset reports Core Web Vitals (LCP, CLS, INP) with full histograms, enabling analysis of any percentile including the long tail. It reveals that WebKit browsers underperform Blink in 46 countries, and that resource-discovery and render-blocking contribute more to slow LCP than download time. Soft navigations in single-page applications load two to three times faster than hard navigations, but landing pages remain heavy. The dataset is publicly available on Google BigQuery and includes differential privacy via the RAPPOR algorithm to protect individual user data.
Air Teams: Bring Your Best Agentic Workflows to the Whole Team – and Automate Repeatable Work
JetBrains introduced Air Teams, a shared workspace for agentic development that extends beyond individual developers to teams and organizations. The system comprises four components: Automations for recurring tasks triggered by events or schedules, shared cloud environments with preconfigured tools and credentials, cloud tasks that run in parallel without tying up local machines, and team projects that centralize members, environments, and credits. Automations include templates for code review, bug fixes, and dependency updates, and they run under a project’s service account so they persist after a creator leaves. Shared environments are defined in a repository’s .air/cloud/startup.sh script and can be snapshotted for reuse, with secrets shared without exposing values. Air Teams is available to business customers, with plans to expand to individual customers later.
Upgrading to Red Hat JBoss Web Server 7: Key changes & impacts
Red Hat JBoss Web Server 7, based on Apache Tomcat 11 and Jakarta EE 11, requires Java 17 as the new minimum and drops support for Java 11. Jakarta EE 11 does not require another import rename but adds a RecordELResolver for records and an OptionalELResolver for optionals. Tomcat 11 changes parameter parsing to throw exceptions on malformed requests, reduces maxParameterCount to 1,000 from 10,000, removes FailedRequestFilter and HTTP/2 push, and retains quotes in cookie values per RFC 6265. The Java SecurityManager is removed, UBI 8 images are discontinued for UBI 9, Prometheus is no longer bundled, and Red Hat recommends container isolation as replacement. Teams must confirm Java 17 readiness, check for SecurityManager usage, update their base image pipeline, and note that Spring Boot 4 on JWS 7 is unsupported.
How to rank fraud detection models using custom cost metrics
Accuracy is misleading for fraud detection because a model labeling everything legitimate can score 96% while catching zero fraud when fraud is only 4% of transactions. The post runs a hand-built Python grid search across algorithm, hyperparameters, feature set, class weight, and decision threshold, ranking by a custom cost metric balancing fraud savings against a US 6 analyst cost per flag. On a 480-transaction test split from a 2,400-transaction dataset with 4% fraud, the winning logistic regression caught 15 of 19 frauds, saving US 3,578 over a baseline that caught none. The winner’s evidence is passed to a local 1-billion-parameter LLM that drafts analyst notes, keeping the LLM downstream of the decision to avoid hallucination. The same workflow scales to Red Hat OpenShift AI by swapping local components for enterprise services like AutoML with AutoGluon and model serving via vLLM.
Top topics
- Architecture834
- Performance816
- LLMs660
- Machine Learning622
- Open Source563
- AI Agents553
- AI470
- Developer Experience448
Weekly digest
The best engineering reads, once a week.