Loading…
Latest reads
Every post, summarised
Grounded summaries of engineering blogs, newest first. Each links to the original.
- 4,738 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
Sentry JavaScript SDK v11 adds tracing for all JS platforms
Sentry JavaScript SDK v11 replaces OpenTelemetry-based tracing with a code transformer from the apm-js-collab packages that injects diagnostics channels into supported modules at load time. The transform wraps target functions in tracingChannel() calls, and Sentry integrations subscribe to those channels to create native spans directly from the events. This approach works both at runtime via Module.registerHooks and at build time through bundler plugins for Vite, esbuild, webpack, and others, with no risk of double-wrapping. The previous OpenTelemetry instrumentation required Node-specific loader hooks, could not instrument bundled dependencies, and had ordering issues with ESM imports, limiting support to Node.js. v11 extends automatic tracing to Cloudflare Workers, Deno, Bun, and bundled server deployments while offering three OpenTelemetry integration modes: Sentry only, OpenTelemetry-compatible, or custom OTel with OTLP export.
Multi-Tenant SaaS Apps with Auth0 Organizations — Part One
Auth0 Organizations provide an isolated workspace per tenant with dedicated members, identity provider connections, metadata, and role-based access control, addressing enterprise requirements like Acme Corp’s SSO ultimatum. Login flows differentiate between B2C users (no organization parameter) and B2B users (organization parameter scopes connections, branding, and token claims including org_id). Home Realm Discovery automatically matches users to their enterprise IdP by email domain when an organization is set, requiring the Identifier First authentication profile and eliminating password prompts. Authorization is per-organization via org_id claims in access tokens, with organization-level roles and permissions resolved through RBAC. The post also covers invitation tickets, tenant context switching where users obtain new tokens with a different org_id, recommends one Auth0 Organization per app tenant, describes Auth0 Actions for custom metadata, and introduces this as part one of a five-part series on B2B identity.
Biomedical Imaging's Real Bottleneck Is the Data, Not the Model
Medical imaging’s real bottleneck is not model architecture but data access. Hospitals, academia, medtech, and pharma all struggle to extract, de-identify, and query imaging data locked inside PACS and vendor archives. The approach proposed is a data lakehouse model on Databricks using a medallion architecture that ingests raw DICOM into Bronze, extracts de-identified metadata into Silver, and curates queryable Delta tables in Gold. This foundation unifies imaging with multi-omics, clinical-trial, and real-world data under a single governance model. The algorithms exist; the missing piece is a governed, queryable, and linkable data layer that makes the model the easy 10% again.
5 Steps to Create SimReady Assets for Robotics with Frontier AI Models
The post presents a five-step workflow for converting CAD assets into SimReady assets for robotics simulation using NVIDIA Omniverse tools and frontier AI models such as GPT-6 Astra. The workflow begins with importing STEP files into OpenUSD and matching visual appearance using reference images and videos from ABB Robotics. Physics properties including joints, collision geometry, mass, and friction are then configured, with the AI model interpreting manufacturer datasheets and URDF descriptions to set correct axes and limits. SimReady Foundation validation checks are run to ensure the asset meets simulation requirements, flagging issues for correction. A final pick-and-place task in NVIDIA Isaac Sim demonstrates successful simulation of the ABB YuMi robot, with both arms and grippers completing four cycles using cubes and a Sharpie marker.
How we migrated Clera’s 500 GB production database to ClickHouse Managed Postgres overnight
Clera, an AI talent agent representing 200,000+ candidates, migrated its 500 GB production Postgres database to ClickHouse Managed Postgres after outgrowing its previous provider. The previous database suffered from point lookups taking multiple seconds, 100% CPU for six straight days, and timeouts. Clera evaluated Supabase, Neon, PlanetScale, and ClickHouse, choosing ClickHouse for its local NVMe storage that avoids network-attached disk latency. In sysbench TPC-C testing, ClickHouse showed 16–42% higher throughput than the runner-up. The team migrated 500+ tables from the EU to the US overnight using ClickPipes and manual replay, with production traffic still hitting the database. After migration, CPU usage dropped to 10–20%, and no settings have been changed since the cutover.
Production-grade LLMs and agents: a field guide
Production-grade LLM agents require deterministic machinery around the probabilistic model, not a larger model. A maturity model defines six levels: determinism, evaluation, confidence, safety and governance, operability, and production-grade build, each assuming the one below. Teams should fix their weakest level in order, starting with determinism before evals, and contain the model to a single node in a fixed graph where it proposes actions a separate component applies after approval. A self-assessment checklist with items like evals as a CI gate and layered guardrails helps teams identify their current level, with fewer than half ticked indicating they are earlier than the demo suggests. The post links to a series of 25 deep-dive articles covering each rung of the maturity model.
Part 5: Operating an LLM system: observability, cost, routing, and the platform underneath
Level 5 covers operability for LLM systems: daily running, cost control, provider outage survival, and fast shutdown on misbehavior. A single gateway funnels all model traffic, measuring cost, routing requests, enabling failover, and hosting an instantaneous kill switch; a shared decision_id threads metrics, logs, traces, and the audit ledger. Observability supplements RED metrics with a decision layer capturing auto vs. human splits, guardrail blocks, judge disagreements, token spend per tenant and capability, and shadow-eval pass ratios via tiered structured logs. Cost controls attribute spending at the gateway per tenant and capability, with hard budget ceilings; optimization follows order: skip model, batch, cache, right-size, trim—each step profiled against quality. The platform spine includes stateless agent services, the gateway, an audit datastore, cache, secrets store, and object storage—scoped with least-privilege IAM and encryption, avoiding queues or vector stores until warranted.
Announcing the new Intercom Messenger - The Intercom Blog
Intercom announced a new messenger after a year-long development effort, its biggest project ever, taking ten times longer than typical releases, during which the team explored bots and machine learning. The messenger addresses the gap between consumer messaging trends and outdated business methods like email and live chat, which customers find impersonal. It offers three message formats (Chat, Notes, Posts), different delivery modes (interruptive or discreet), and lightweight replies with emoji reactions. The messenger supports customization like color and wallpaper, provides native experiences on web, iOS, and Android, and integrates with multiple channels like web, push, in-app, and social media. Built on the Intercom platform, it uses customer data to personalize messages for 13,000 businesses, with future plans for conversational UIs, bots, and AI to invent the future of business communication.
How we redesigned our blog, Inside Intercom - The Intercom Blog
Intercom redesigned its blog, Inside Intercom, because the volume of articles, podcasts, books, and videos had outpaced the old design’s ability to help readers discover and consume content. The new homepage employs a magazine-like hierarchy on a solid grid, featuring a lead article with secondary and tertiary items, while podcasts now have their own category with an inline player for seamless listening. Two new categories for Startups and Marketing, along with reading enhancements like estimated reading times and progress bars and the removal of the pop-up newsletter signup, improve the user experience. Typography pairs Tiempos serif with Akkurat Pro sans-serif to create an editorial feel, and custom illustrations by a diverse range of artists plus hand-painted author portraits by Jon Stich add a personal touch. The redesign, led by designer Maria Gonzalez and developer Daniel Husar, aims to make content discovery simpler and reflects Intercom’s design-led culture.
Ways to think about customer onboarding - The Intercom Blog
Customer onboarding is critical, yet most apps lose half their new users by the second day because it fails to deliver real value. Instead of focusing on feature adoption like creating a to-do list, onboarding should help users achieve their ultimate goal, such as delivering a project on time, turning them into superheroes. The Jobs-to-be-Done framework reveals the motivations driving sign-ups, enabling segmentation and tailored onboarding flows for groups with similar needs. The four forces model adds context by showing push and pull factors when users switch, maximizing forces for change like long email trails and minimizing resistance like switching anxiety. Finally, behavior-based messaging should be scheduled by time with a priority list rather than predefined sequences, avoiding the complexity of N-dimensional decision trees, as implemented in Intercom’s Smart Campaigns.
Content as a growth engine for your business - The Intercom Blog
Content marketing advice often focuses on promotion rather than the substance of the content itself, leaving businesses without a clear strategy. Drawing on his experience leading Intercom’s content team and two decades as a journalist, John Collins argues for evergreen content that addresses timeless themes like filing a good bug report rather than news events. Evergreen content decays slowly and provides a compound return, with analysis showing it grows traffic three times faster than news-focused blogs over a year. He applies a journalist’s 5W and 1H framework—who, what, where, when, why, how—to plan content strategy, emphasizing that specialists should write, topics should help customers solve problems, and content must be relevant to the business. Collins concludes with four key takeaways: stay committed for the long term, make content relevant, impose a high editorial bar, and embrace unexpected benefits like easier hiring.
Build your customer relationships around consistency and trust - The Intercom Blog
The Industrial Revolution scaled business reach but separated owners from customers, forcing service through mail or phone, a pattern that the internet initially continued. Now, consumerization of technology provides tools for human interaction at scale, enabling businesses to return to building personal relationships online. Amazon exemplifies this by minimizing customers’ total cost beyond price—time, mental energy, and uncertainty—through frictionless purchasing, easy returns, and proactive resolution of issues. To build trust, businesses must help customers succeed, invest in relationships by being generous and never transactional, and be available on customers’ preferred messaging apps like Messenger or WhatsApp without requiring repeated context. Consistent behavior over time builds brand reputation, and businesses should engage on review sites and social media to shape their message and earn trust, as perception eventually matches reality.
Inside Intercom New York startup panel - The Intercom Blog
At an Intercom World Tour panel in New York, founders from Augment, an AR platform; Greenhouse, recruiting software; and Hello Alfred, on-demand home services, shared lessons on building emerging technologies. Mike Boufford of Greenhouse described fixing bugs by chatting with customers and deploying fixes within minutes, demonstrating how product and support intertwine. Christian Bjelland of Hello Alfred explained that their service must advocate for customers when real-world problems like lost laundry occur, requiring quick reactions. On growth, Boufford advised proving product-market fit before raising money, while Bjelland stressed dense, sustainable growth in specific zip codes to maintain quality. Lindsay Boyajian of Augment noted that no AR platform existed before, so constant user feedback and rapid iteration are critical to building a useful product.
Design in the age of entanglement - The Intercom Blog
The post argues that product design has entered a fourth era, the age of entanglement, following three earlier phases: augmentation, the PC era, and the web era. The author claims that three overlapping technology trends that made innovation predictable are ending, creating uncertainty. Designers now face a complex web of interconnected ideas, technologies, and systems that no single person can fully grasp. The post advises designers to think in systems, build one thing at a time from the center outward, and observe and iterate based on real-world reactions. It suggests that this uncertainty may be a return to a more natural state, where technology takes its rightfully uncertain place in the world.
Get To Know Your Users - The Intercom Blog
Startups need a comprehensive view of their users, but customer information is often scattered across separate systems like databases, support apps, and emails, creating data silos. Intercom today launched a unified user profile screen that aggregates profile data (e.g., company role), business data (e.g., plan), activity data (e.g., login frequency), and communication data (e.g., past conversations) into one authoritative view. This screen answers key questions such as who the user is, how often they use the app, whether they have seen announcements, and if anyone has spoken with them recently. When customer data is siloed, companies may email inactive users or ignore influential signups, costing them customers and leaving them clueless about churn. By simplifying user research by an order of magnitude, the tool enables proactive, targeted communication that improves loyalty and changes how products are designed.
Inside Intercom London startup panel - The Intercom Blog
At the first Inside Intercom World Tour stop in London, startup founders Bridget Harris, Will Swannell, and James Blackwell shared insights on building great products. On pricing, James advised setting it at the high end you think you can get away with, while Bridget regretted offering a free product that forced later price increases; Will argued simplicity is key, with the cost answer in two words. For hiring, Bridget uses a process of elimination to avoid bad hires and recommends remote options; Will noted HR’s importance is often underestimated. On funding, Will said crowdfunding builds a community of buyers, while Bridget warned VC funding changes priorities. The panel also stressed measuring customer acquisition through each interaction to profit, treating every interaction as a sales opportunity, and ensuring the product solves a real problem; James added that great products create moments of delight when things just work.
How users understand new products - The Intercom Blog
Users form different mental models of the same product based on their experiences and assumptions, as illustrated by the “draw toast” exercise where people depict the process in diverse ways. When a product’s system model—how it actually works—aligns with a user’s mental model, the interface feels intuitive; a mismatch causes confusion. To discover these mental models, product teams should talk to users, ask them to draw how they believe the product works, and listen to the nouns and verbs they use. Designers can bridge gaps by employing standard interface patterns (e.g., buttons that behave like buttons), using customers’ own language, and leveraging metaphors like Apple Wallet’s ticket-like appearance or Nest’s dial-based thermostat interface. When a product cannot mirror existing mental models, teams can educate users through tutorials and walkthroughs, similar to video game onboarding, helping customers learn new concepts and quickly grasp the product’s value.
How A/B Testing Can Help Your Team Prioritize, Not Just Validate
A/B testing can help teams prioritize projects by forcing them to estimate effect sizes and required sample sizes before committing resources. The most common mistake is calling tests too early based on statistical significance without sufficient statistical power, leading to unreliable results. To run a test correctly, teams should estimate the expected KPI improvement, use a sample size calculator to determine test duration, and decide if the wait is worthwhile. For example, a complete signup flow redesign expected to boost conversion from 10% to 15% requires 24 times less data than a color change expected to lift it to 11%, and a non-significant result can still be useful. Startups with limited traffic should focus on large, high-potential changes to avoid working towards a local maximum, while larger companies with high traffic can profitably test small optimizations.
Content strategy: How words build a product - The Intercom Blog
Words in a product evolve like the Ship of Theseus, with each change affecting identity; the term “page” long constrained web design, showing how working titles gain unintended baggage. Intercom’s content strategy transforms this organic dialect into an intentional one, addressing team silos and UI inconsistencies. They categorized every visible word—objects, views, people, actions—into a taxonomy and glossary that gave consistency and control over how each part is described and how parts relate. They interviewed all teams to map natural vocabularies, discovering job-related differences and treating them as assets. Finally, they derived three UI writing principles: speak naturally, speak directly to the user but never for the user, and guide users through complex ideas, recommending that content strategy start after discovering the organic voice already present.
Give Your Customers Best Practices, Not Just FAQs - The Intercom Blog
Most companies invest heavily in acquisition content but neglect post-signup support, leaving customers with poorly written FAQs and contributing to high churn rates. Intercom advocates replacing isolated how-tos with best practice content that teaches customers how to get the most from a product. The approach includes giving the entire lesson, using real-life customer examples, writing about how the team uses the product, mirroring customer language, and revisiting content often. Practical techniques include breaking tasks into quick wins, using bullet points and screenshots, embedding videos, and leveraging a bot for automated answers. The goal is to empower customers to self-serve, reducing support contacts and building long-term loyalty.
Top topics
- Architecture834
- Performance817
- LLMs673
- Machine Learning622
- Open Source568
- AI Agents564
- AI483
- Developer Experience448
Weekly digest
The best engineering reads, once a week.