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
How to Track Opt-Outs in PHP
This tutorial demonstrates building a PHP application that records customer opt-out preferences from SMS replies using Twilio’s Incoming Message Webhook and Event Streams, which require PHP 8.5, Composer, and SQLite. When the SMS body is an opt-out keyword such as STOP, START, or HELP, Twilio includes an OptOutType parameter in the webhook request, which the application checks using strcasecmp and array_first. The application, based on the Twilio Slim Base project, stores the opt-out status and account SID in an SQLite database via PHP-DB and laminas-servicemanager, using a TableGateway for the user_optouts table. The handleWebhook() function validates the OptOutType against class constants (OPT_OUT_TYPE_START, OPT_OUT_TYPE_STOP, OPT_OUT_TYPE_HELP) and inserts a record before returning a TwiML response. Twilio already blocks further messages for standard opt-outs, but local recording enables advanced actions like syncing a CRM; the tutorial covers bootstrapping, configuring Twilio messaging services with Advanced Opt-Out, using ngrok, and verifying with an SQL query.
How to Get Started With WebRTC: Intro to Browser APIs
WebRTC provides browser APIs for real-time voice, video, and data sharing without plugins, with mandatory encryption via DTLS-SRTP and DTLS; the standard became a W3C Recommendation in January 2021. The core APIs are getUserMedia for capturing audio/video from user devices, RTCPeerConnection for peer-to-peer media exchange via offer/answer and ICE candidates, and RTCDataChannel for arbitrary data transfer. WebRTC is supported by all major browsers and used by services like Google Meet, Discord (which uses media servers for group calls to avoid expensive peer-to-peer), and AI voice agents. Twilio offers Video and Voice JavaScript SDKs that handle signaling and media relay, enabling calls to phone numbers and SIP endpoints. The post includes a camera preview example using getUserMedia that requires a secure origin and can be served with Python, then directs readers to Twilio’s quickstart tutorials for building video and voice applications.
Limits of Confidence in Diffusion
The post analyzes limitations of confidence-based ranking in discrete diffusion, where sequences are generated by writing multiple token positions per step from per-position distributions. It shows that a step matches the training distribution only when the written positions are conditionally independent given fixed tokens, and no product of per-position distributions can match a dependent group. Additionally, per-position distributions do not determine group dependence: two joint distributions can have identical per-position marginals but differ in value combinations. On the synthetic ScanAndAdd task, every group of two or more undetermined positions written by a confidence ranking is dependent, and the generated distribution is 29 times the sampling-noise floor in total variation, with per-sample metrics at 1.0. The post also includes updates on two related readings: on-policy distillation for reasoning models and self-supervised learning with Transformers.
Architecting Secure Identity: Extending Auth0 PrivateLink with Intelligent Gateways
Enterprises centralize authentication into a single Auth0 environment but must comply with frameworks like PCI-DSS, HIPAA, and SOC 2 that restrict data to private channels. Auth0 Outbound PrivateLink provides a secure unidirectional connection between Auth0 and AWS infrastructure, avoiding the public internet. However, a single PrivateLink connection per environment cannot directly reach multiple isolated VPCs for staging and production. The Intelligent Gateway Pattern solves this by deploying an API gateway or reverse proxy at the PrivateLink termination point to fan out traffic. Auth0 Actions append context such as custom headers like X-Target-Env or specific paths, and the gateway inspects these to route requests to the appropriate isolated VPC, decoupling secure transport from internal routing. This pattern unblocks identity consolidation by allowing a single Auth0 tenant to serve both staging and production workloads while maintaining strict VPC isolation.
Language Discrimination Improves Linguistic Learning in Multilingual Speech Models
Multilingual self-supervised speech models underperform monolingual models under a matched total pretraining data budget, but strengthening language discrimination during pretraining reduces this gap while preserving cross-language sharing. In a controlled English/French HuBERT setting, two interventions—an auxiliary language classifier and per-language k-means targets—improve language discrimination. Continuous-feature phone discrimination error (phone-ABX) decreased from 11.6% in the bilingual baseline to 10.4%, approaching the monolingual 10.8%. Lexical performance (sWUGGY) rose from 52.1% to 56.7% (monolingual: 58.5%), and prosodic performance (ProsAudit lexical subtask) from 68.9% to 72.9% (monolingual: 72.6%). The strongest gains occur when language discrimination is introduced in the first HuBERT iteration; later or repeated interventions yield smaller improvements and increase language-wise segregation, supporting a causal role for language discrimination in reducing the cost of multilingual learning.
Seer, the Sentry MCP and CLI, or your own coding agent: where each one fits
Seer is Sentry’s agent debugger that reasons across errors, traces, logs, profiles, and linked repositories to identify root causes, not just surface-level stack traces. The Sentry MCP, set up via the agent plugin with a single command, bridges Sentry’s data into coding agents like Claude Code and Cursor, enabling them to invoke Seer’s full analysis. The new Sentry CLI provides natural-language access to Seer’s diagnosis and issue management directly from the terminal, using selectors like @latest and @most_frequent. These tools are complementary: Seer provides deep diagnosis, while the MCP and CLI deliver that context into the environments developers already use. Teams can automate the entire workflow so Seer diagnoses, an agent fixes, and a draft PR awaits review without human involvement until approval, with Seer triggering on issues meeting event count and fixability thresholds.
Creating a whitelabeled connected account dashboard with embedded components
This tutorial explains how to build a whitelabeled connected account dashboard using Stripe Connect embedded components, eliminating the need to manually implement pagination, permissions, and state management. The example uses FurEver, a pet grooming SaaS platform, and integrates the ConnectPayments, ConnectPayouts, and ConnectNotificationBanner components for payments, refunds, dispute management, balance and payout views, and account enablement alerts. The server-side requires creating an account session with enabled components, while the frontend initializes ConnectComponentsProvider with the fetched client secret and publishable key. Platform user roles can be enforced by setting features such as refund_management based on admin status in the account session, automatically adjusting component behavior. The dashboard’s appearance matches the platform using the appearance API with variables for colors, fonts, and locale, and it receives automatic updates for new payment methods and compliance.
Lakebase Postgres branch-based restores for fast recovery at scale
Traditional Postgres point-in-time recovery is slow because compute and storage are coupled, requiring provisioning a new instance, restoring a snapshot from object storage to disk, and replaying WAL, leading to hours of downtime for large databases. In a survey of developers running 1TB+ production Postgres, 59% had a critical production failure and 30% were down for over three hours. Lakebase Postgres decouples compute and storage using safekeepers, a pageserver, and object storage, storing database history as an immutable timeline. A restore becomes a branch at a point in history, a metadata operation that does not copy data, making it near-instant regardless of database size. This eliminates data copies and WAL replay, so restoring a 100 TB database is as fast as restoring 10 GB, and agents can use restores as ordinary tool calls for versioning and undo.
Build Applications on NVIDIA BlueField Faster with NVIDIA DOCA Agent Skills
General-purpose AI agents lack domain-specific knowledge for NVIDIA DOCA development, often inventing APIs or skipping hardware checks, leading to correction cycles. DOCA AI agent skills provide a structured, machine-readable specification in a SKILL.md file containing verified API signatures, hardware capability requirements, and build constraints for components like Flow, GPUNetIO, and PCC. In a 65-prompt evaluation, agents without skills satisfied only 19% of graded checklist items, while agents with skills satisfied 100% across all prompts. The skills reduce correction cycles; in a side-by-side demo, the with-skills agent used 73% less handwritten code and 46% fewer hardware commands. The skills are available on GitHub and enable developers to build faster, ship more stable code, and deploy with fewer unknowns on BlueField infrastructure.
Build Local AI Apps with C++ and NVIDIA TensorRT RTX Samples
DIN Deploy is an open-source collection of C++ samples combining ONNX Runtime with NVIDIA TensorRT RTX for local AI inference on Windows and Linux. It separates model conversion from deployment: Python exports models to ONNX, then native C++ runs inference via ONNX Runtime session and tensor APIs. Supported workloads include automatic speech recognition (Whisper offline, Parakeet TDT streaming), interactive segmentation (SAM 2.1), and prompt-driven image generation (FLUX.2-klein-4B). The FLUX.2 sample includes Vulkan and DirectX graphics interop and demonstrates post-training quantization with NVIDIA Model Optimizer, producing quantized ONNX models as drop-in replacements without code changes. On DGX Spark, GPU acceleration provides speedups: Whisper large-v3-turbo achieves 58.5x real-time factor, Parakeet TDT 206.41x, and SAM 2.1 runs at 38.3 FPS on GPU versus 0.5 FPS on CPU.
Accelerating airline retailing innovation: how Datalex modernized with AWS Experience-Based Acceleration and agentic AI
Datalex, a leader in airline ecommerce, needed to modernize its mature EJB2/Java 8 system for Modern Airline Retailing while maintaining continuous airline operations. Over three days, 16 Datalex engineers and 6 AWS specialists proved migration feasibility by extracting a Reservation service into a Spring Boot microservice on Java 21. The team used agentic AI tools—Kiro and AWS Transform Custom—to automate code changes, and established repeatable patterns for DevSecOps, observability, and a multi-agent AI prototype via Amazon Bedrock AgentCore. A business service proxy using the Strangler Fig pattern enabled incremental migration with backward compatibility. The board approved significant investment, the prototype team expanded into a full Agile team, and the EBA shortened discovery from an estimated 8–12 weeks to three days.
Introducing Clef: our open-source decision models, and new RL fine-tuning platform
Cloudflare released two open-source decision models, Clef and Clef-flash, hosted on Workers AI under an Apache 2.0 license and fully compatible with the Jev API. These models produce bounded structured outputs for agentic workflows, classifying inputs like customer support messages or website domains with typed probabilities and confidence scores. Clef includes a vision encoder and a 64k context window, distinguishing it from text-only decision models such as Jev, and currently leads the Jev Decision Index benchmark. The models demonstrated lower latency than alternatives, with Clef classifying a website in 2.2 seconds versus 4.7 seconds for Cloudflare’s fastest general LLM, while Clef-flash offers even faster inference at 38.8 ms median latency. Cloudflare also debuted a reinforcement learning fine-tuning platform that lets customers adapt Clef to their specific use cases using their own data.
Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs
Olmo-core 3 is an open training infrastructure upgrade for large mixture-of-experts (MoE) models, designed to scale to trillion parameters while preserving efficiency. It switches from FSDP to a DDP-based system that keeps experts resident on GPUs and routes data to them, avoiding weight gathering overhead. The system combines expert, pipeline, and distributed optimizer parallelism with optimizations like rowwise expert parallelism, GPU-resident routing, and grouped GEMM. In a benchmark, increasing experts from 8 to 128 while routing each token to four experts kept active parameters at ~3.2B, total parameters grew from 4.6B to 47B with throughput falling less than 5%. The infrastructure supports MXFP8 lower-precision format, achieving about 21% higher throughput than BF16, and has been tested on a 1.2-trillion-parameter model.
Connecting customer context to measurable ROI with agentic marketing
Customer journeys have become non-linear while identity signals decay, and marketing value is shifting from reach frequency to a prediction economy that rewards anticipating needs. AI agents can recommend or take the next-best action for each customer within human-defined guardrails, with each action measured incrementally against what would have happened without it. Identity resolution must be continuous, connecting known and pseudonymous signals across channels, to provide reliable context for agent decisions. The context required includes customer behavior, business objectives, decision history, and control guardrails, with a single transaction carrying up to 50 metadata parameters. A practical starting point is focusing on one meaningful decision, such as reactivating lapsed customers, and expanding as trust and results build.
A look back before we look forward: A Developer Survey retrospective
A retrospective analysis of the Stack Overflow Developer Survey compares 2024 and 2025 data across AI adoption, work settings, and demographic shifts. AI tool usage among developers rose from 62% to 79% year-over-year, while agent adoption jumped from 31% in 2025 to 59% in an April 2026 pulse survey, with ChatGPT and GitHub Copilot as primary gateways. Enthusiasm for AI tools cooled: favorable sentiment fell from 72.0% to 59.7% overall and from 72.2% to 52.8% among learners, and trust in accuracy declined from 49.1% to 37.4% for students, while skepticism among learners quadrupled from 6.4% to 26.3%. Fully remote work dropped from 38% to 32.4%, but 82% of respondents had some remote flexibility. The article identifies growing reliance on AI alongside skepticism and frames unanswered questions about learning, trust, and job satisfaction for the upcoming 2026 survey.
One year later: Sovereign AI and the fight for choice
A year after arguing that AI sovereignty means choice, Cloudflare adds two European open models—EuroLLM and Apertus—to Workers AI, reflecting a year where access to frontier models became constrained and attacks using AI increased. EuroLLM supports 35 languages including all 24 EU official languages, many underserved by existing open models, and was trained on the MareNostrum 5 supercomputer. Apertus, Switzerland’s first fully open multilingual model, was trained on 15 trillion tokens across 1,500+ languages on the Alps supercomputer (10,000+ GH200 GPUs) and outperforms leading models on rare and regional languages. Cloudflare also launches hands-on workshops for government cyber agencies and critical infrastructure operators to build model-agnostic AI security defenses, starting in Singapore in October. The company open-sourced its own security harness that coordinates multiple AI models for vulnerability hunting, ensuring defenses remain operational even if access to one model is lost.
Master your skills: Building skills you can trust
Skills in AI agent projects silently misfire when their trigger descriptions are passive or too general, causing the agent to skip using them. The authoring tool skill-creator, from Anthropic’s skills collection, provides an eval loop that generates positive and negative test prompts and iterates on description accuracy using a training and held-out test set to avoid overfitting. These tests are ephemeral, so the open-source framework promptfoo extends them into permanent configuration files that can be committed alongside skills. Promptfoo’s agent SDK providers test through the actual agent runtime and include a skill-used assertion to confirm proper routing, and its YAML-based configuration fits naturally into version control. The final step is wiring promptfoo evals into CI, using GitHub Actions with path filters to run on pull requests that modify skills, catching regressions before they ship and costing roughly $0.10-0.15 per eval.
GPU virtualization at scale: AMD MI300X SR-IOV on Red Hat OpenStack Services on OpenShift
Enterprise AI environments are moving from dedicated GPU servers to shared accelerator infrastructure, requiring higher GPU utilization with strong isolation. AMD Instinct MI300X GPUs support hardware-assisted SR-IOV in CPX mode, each exposing eight virtual functions with separate PCI and VRAM allocation. When integrated with Red Hat OpenStack Services on OpenShift, the control plane runs on OpenShift while RHEL compute nodes use EDPM, with Nova scheduling VFs via VFIO PCI passthrough. Validation on an 8-GPU system created 64 VFs, demonstrated 64 VMs with one VF each, and tested up to 32 VFs per VM. BabelStream benchmarks showed approximately 96-99% of bare-metal memory bandwidth, while the testing also identified limitations in GPU peer-to-peer DMA and runtime VF lifecycle.
Cloudflare OS: your company’s agent workspace, managed for you
Cloudflare OS is an open-source agent workspace that connects to company data and systems. The fully managed version lets organizations launch an agent workspace in a few clicks from the Cloudflare dashboard, with control over access, skills, and connected systems. Recent updates add GitHub repository integration, allowing agents to explore codebases, fix bugs, and open pull requests. Google Workspace connectivity now supports reading Gmail, creating drafts, sending emails, and accessing Google Drive. Users can export work to formats like Excel, CSV, PDF, Markdown, or HTML, with Word and PowerPoint export coming soon. The managed service is available via waitlist, while the open-source version can be deployed into your own Cloudflare account.
Announcing Cloudflare K2: serverless event streams
Cloudflare launched K2, a serverless event streaming service in public beta, to decouple producers and consumers in distributed systems. K2 stores events as an ordered log, allowing independent consumers to read at their own pace without losing data. Under the hood, it builds a partitioned durable log on top of R2 object storage, using atomic operations for ordering. Produce latency at the 99th percentile is about 1 second due to batching writes to R2 segments. K2 is available now for Workers Paid accounts with a 10GB storage limit and 30 MB/s produce per stream, with no billing during beta.
Top topics
- Architecture834
- Performance816
- LLMs660
- Machine Learning622
- Open Source563
- AI Agents553
- AI470
- Developer Experience448
Weekly digest
The best engineering reads, once a week.