---
title: "Latest reads"
description: "The engineering internet, summarised so you can actually read it."
---

# Latest reads
> The engineering internet, summarised so you can actually read it.

## Articles

### [Accelerate search queries with full-text search indexes on Databricks](https://yomu.fyi/post/accelerate-search-queries-with-full-text-search-indexes-on-databricks.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Yu Xu, Yingyi Bu, Ivan Vezilić
- Published: Jun 16, 2026
- Type: Announcement

Databricks introduces full-text search indexes in Beta on Databricks Runtime 18.2 to accelerate substring and keyword queries on large open-format tables without changing their layouts. The indexes tokenize text columns into a compact lookup structure mapping tokens to matching rows; at query time, the engine uses it to identify candidate files and skip most of the table. They are maintained asynchronously, require no query hints, preserve complete results when stale by scanning indexed and non-indexed data as needed, and support Unity Catalog managed Delta and Iceberg tables on serverless and classic compute. A Trust and Safety team reported a substring search running more than 100x faster on a petabyte-scale table, while Liquid clustering remains complementary because it optimizes column-value filters rather than text within fields.


### [Introducing CustomerLake: The Agentic CDP embedded in Databricks](https://yomu.fyi/post/introducing-customerlake-the-agentic-cdp-embedded-in-databricks.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Tasso Argyros, Justin DeBrabant, Michael Trapani, Dan Morris, Katy Yuan
- Published: Jun 16, 2026
- Type: Announcement

Databricks announces CustomerLake, an Agentic Customer Data Platform embedded natively in its lakehouse, bringing Customer 360, identity resolution, audience building, campaign automation, activation, and personalization alongside governed data and AI models. The announcement addresses fragmented identities, stale audiences, manual campaign workflows, and the duplication and governance burden created by separate martech systems. CustomerLake uses Unity Catalog and Lakehouse Federation to access customer data across Databricks, Snowflake, Google BigQuery, cloud object storage, operational databases, and other enterprise systems, while Profile Agents create business-ready profiles and Campaign Agents build audiences, recommend actions, activate channels, and optimize engagement. Its operating model is described as embedded, democratized, and autonomous, with Agentic Identity Resolution combining deterministic, probabilistic, and agentic workflows. CustomerLake is now available in Private Preview and launches with an open partner ecosystem.


### [What is customer segmentation?](https://yomu.fyi/post/what-is-customer-segmentation.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Databricks Staff
- Published: Jun 16, 2026
- Type: Explainer

Customer segmentation divides an existing customer base into smaller groups based on shared demographic, geographic, psychographic, behavioral, firmographic or value-based characteristics. Unlike market segmentation, it uses first-party data about existing customers, and a customer can belong to multiple segments at once. The guide distinguishes rule-based, survey-based, RFM, k-means, decision-tree and AI/ML-driven methods, with choices depending on data maturity and business goals. Effective implementation defines an objective, audits and unifies sources, resolves duplicate identities, selects a method, validates segments and measures outcomes. It also describes Databricks' CustomerLake capabilities, including governed Customer 360, agentic identity resolution, natural-language segmentation through Genie and bidirectional activation connectors; stated benefits include improved retention, conversion, customer lifetime value and marketing efficiency.


### [What is an open lakehouse? Open data standards, explained.](https://yomu.fyi/post/what-is-an-open-lakehouse-open-data-standards-explained.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Lisa Cao
- Published: Jun 16, 2026
- Type: Explainer

The piece defines an open lakehouse as a lakehouse whose storage, table format, processing engine, catalog, and ML and AI tooling use open standards and remain interchangeable. It contrasts this architecture with warehouses, lakes, and proprietary lakehouses, emphasizing low-cost object storage, ACID transactions, governance, schema guarantees, and the ability to change engines without rewriting data. Its reference stack combines open table formats such as Delta Lake and Apache Iceberg with Apache Parquet, Apache Spark, Unity Catalog, and MLflow, while allowing engines including DuckDB, Trino, and PyIceberg to work on the same data. The article also distinguishes open standards from open-source code, explains that a table format is only one layer of the stack, and states that the components can be self-hosted or consumed through a managed service.


### [Introducing Lakehouse//RT: Real-Time Performance on a Unified Lakehouse](https://yomu.fyi/post/introducing-lakehouse-rt-real-time-performance-on-a-unified-lakehouse.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Nong Li, Shoumik Palkar, Shant Hovsepian, Mostafa Mokhtar, Reynold Xin
- Published: Jun 16, 2026
- Type: Announcement

Databricks introduces Lakehouse//RT, a real-time data warehouse designed for operational analytics, BI, app serving, and observability workloads. It is powered by Reyden and is intended to deliver millisecond performance directly on lakehouse data without copying it into a separate serving layer. Preview participants saw up to 16x better performance, with response times as low as 10ms on smaller datasets, sub-100ms on larger ones, and sub-100ms latency at 12,000 queries per second on standard analytical benchmarks. Tests covering concurrency, dataset scale, and complex TPCDS queries report low latency where alternatives slowed or failed. Lakehouse//RT is in Beta for select read-only workloads, with incremental autoscaling and automatic baseline compute sizing.


### [13 Social Media Holidays to Celebrate This July](https://yomu.fyi/post/13-social-media-holidays-to-celebrate-this-july.md)
- Company: [Brandwatch](https://yomu.fyi/company/brandwatch.md)
- Author: Gemma Joyce
- Published: Jun 16, 2026
- Type: Explainer

This calendar presents 13 unofficial social media holidays for July 2026, giving brands occasions to engage audiences through patriotism, culture, kindness, humor, and food. It lists dates from International Joke Day on July 1 through International Day of Friendship on July 30, including two observances on July 18 and two joke-focused days. Each entry pairs the observance with a suggested engagement prompt, such as asking for favorite jokes, running polls, inviting emoji-only stories, collecting acts of kindness, or encouraging friends to share appreciation. The recommendations span educational, entertaining, inspirational, and community-oriented content, with examples tied to Independence Day, World Emoji Day, and National Ice Cream Day. It is positioned as a way to shape social content during the second half of 2026.


### [Case study: TriMet unlocks same-day paratransit with an innovative partnership](https://yomu.fyi/post/case-study-trimet-unlocks-same-day-paratransit-with-an-innovative-part.md)
- Company: [Uber](https://yomu.fyi/company/uber.md)
- Published: Jun 16, 2026
- Type: Problem & solution

TriMet’s 2024 partnership with Uber, TransDev, and Trapeze added same-day capability to its LIFT paratransit service in Metro Portland. Before LIFT+, trip management spanned multiple systems, while cancellations and manual coordination increased operational burden and weakened the rider experience. In April 2024, riders began booking through LIFT’s web platform or call center, after which TriMet’s integrated Trapeze PASS TripBroker used the Uber Plug-in to select dedicated vehicles or Uber based on rider profiles, requests, trip efficiency, and fleet capacity; it could also reschedule cancellations and delays. Since launch, TriMet has delivered more than 47,300 Uber trips and reduced average cost per trip from $78 to $54, while riders gained same-day flexibility and trip details including estimated arrival times.


### [How the University of Nebraska-Lincoln teams up with Uber to power late-night rides](https://yomu.fyi/post/how-the-university-of-nebraska-lincoln-teams-up-with-uber-to-power-lat.md)
- Company: [Uber](https://yomu.fyi/company/uber.md)
- Published: Jun 16, 2026
- Type: Explainer

At the University of Nebraska-Lincoln, undergraduate and graduate students needed reliable late-night transportation during the academic year. ASUN began powering rides with Uber in 2025, and the Husker Late Night Ride program launched in 2026 with safety as its primary goal. The program provides five $10 UberX vouchers per semester for rides between 9pm and 3am, seven days a week, with trips limited to a three-mile radius of UNL’s City and East campuses or Lincoln Airport. Students register online, receive vouchers at official UNL email addresses, and add them to Uber accounts for automatic application in eligible zones and times. Funded by Nebraska Athletics alcohol-sale revenue, the program has attracted over 6,000 students and delivered 15,236 Uber trips, with Friday accounting for 37.8% of rides; ASUN uses Uber for Business Dashboard to track spending, impact, parameters, and reports, while pay-per-performance supports cost efficiency.


### [Building Production-Scale Visual AI for Claims and Returns at Swiggy Instamart](https://yomu.fyi/post/building-production-scale-visual-ai-for-claims-and-returns-at-swiggy-i.md)
- Company: [Swiggy](https://yomu.fyi/company/swiggy.md)
- Author: Harshit Juneja
- Published: Jun 16, 2026
- Type: Problem & solution

Swiggy Instamart built two vision-language-model systems to handle more than 40,000 daily claims involving wrong, expired, poor-quality, damaged, and packaging-related items, while reducing reliance on human verification. The Image Validation Model (IVM) compares customer uploads with catalog images, routes each category through a specialised LangGraph StateGraph handler, and combines its structured verdict with fraud, confidence, and business-rule signals. Reverse Quality Check (RQC) compares delivery executives’ pickup photos with the original claim images, allowing matching returns to proceed and prompting retries when they do not. The architecture keeps visual assessment probabilistic but makes routing, aggregation, escalation, and financial decisions deterministic, with traces for category, outputs, confidence, cost, and latency. After more than six months in production, the systems are associated with a 6% customer-satisfaction increase, 20,000 fewer weekly Image Validation tickets, 98.5% RQC accuracy, and over 5,000 daily RQC tickets resolved without human intervention.


### [Uber Advertising Launches New Solutions to Connect Brands On & Off Uber](https://yomu.fyi/post/uber-advertising-launches-new-solutions-to-connect-brands-on-off-uber.md)
- Company: [Uber](https://yomu.fyi/company/uber.md)
- Published: Jun 16, 2026
- Type: Announcement

Uber Advertising has launched new solutions that connect brands with consumers across Uber rides, Uber Eats, and external channels. Offers on Uber include Destination Offers, Homescreen Ride Offers, Ride Offers on Journey, and planned Sponsored Upgrades, combining trip context with incentives and premium ride upgrades. On Uber Eats, Brand Takeover, Item Showcase, and Offers Spotlight provide prominent placements, shoppable item features, and targeted offer visibility during discovery and ordering. Offsite Ads extends campaigns to Meta and Google Shopping, using observed Uber behaviors to target audiences and connect external engagement with orders and store visits; the program is being piloted with selected advertisers. Reported findings include 81% openness to offers while riding, a 45% higher click-through rate for one Miller Lite campaign, and double-digit item attachment after Item Showcase views, with results varying by campaign.


### [Unlocking the Power of the TPU Stack: Introducing our new Developer Hub](https://yomu.fyi/post/unlocking-the-power-of-the-tpu-stack-introducing-our-new-developer-hub.md)
- Company: [Google](https://yomu.fyi/company/google.md)
- Author: Keelin McDonell
- Published: Jun 16, 2026
- Type: Announcement

Google Cloud has launched the TPU Developer Hub, an educational resource for model builders, optimizers, and developers working with TPU infrastructure and its supporting software stack. It organizes guidance across the developer lifecycle, covering pre-training, post-training, and inference, with material on hardware architecture, infrastructure consumption modes, and software capabilities including specialized compiler technology and XLA. Resources also address PyTorch migration and deployment, XProf-based tracing and observability, multi-chip parallelism, Pallas kernels, KV cache offloading, networking, and security for distributed workloads. The hub combines interactive Colabs, open-source code recipes, and deep-dive documentation, with content intended for both manual browsing and AI-assisted development. Its stated purpose is to help users select appropriate TPU infrastructure, diagnose bottlenecks, optimize model performance, and move from concepts toward production through practical, regularly updated guidance.


### [Enhance Security and Trust: New Session Metadata in Sign in with Google](https://yomu.fyi/post/enhance-security-and-trust-new-session-metadata-in-sign-in-with-google.md)
- Company: [Google](https://yomu.fyi/company/google.md)
- Author: Sergei Akulich, Brian Daugherty
- Published: Jun 16, 2026
- Type: Announcement

Sign in with Google now provides verified applications with two OpenID Connect session metadata claims, auth\_time and amr, in ID tokens. auth\_time records the last successful Google authentication time, distinguishing session freshness from token issuance, while amr is a JSON array identifying authentication methods such as pwd, mfa, hwk, swk, tel, and sms. The claims are available across Android, iOS, and web client and server applications, and can support audit logging, step-up authentication based on session age, and authorization rules requiring stronger methods for sensitive operations. Verified applications request them through the standard OIDC claims parameter without significantly changing an existing Sign in with Google authentication flow.


### [What a difference 10 years of open source makes](https://yomu.fyi/post/what-a-difference-10-years-of-open-source-makes.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jun 16, 2026
- Type: Explainer

ClickHouse marks ten years since becoming open source under the Apache 2.0 license, crediting more than 2,600 contributors and a decade of monthly releases for its growth into one of the world’s most widely deployed analytical databases. Recent engineering work spans joins, updates and deletes, JSON, full-text search, vector search, query acceleration, integrations, and developer tooling. Join-focused changes made the TPC-H SF100 workload 26× faster between versions 22.4 and 26.4, while statistics-based reordering reduced one six-table query from 3,903.7 seconds to 2.7 seconds. The project also publishes open benchmarks, including ClickBench, JSONBench, CostBench, and PostgresBench, with scripts, configurations, data, and results intended for independent validation. Its next phase continues to emphasize community contributions and broader ways to use ClickHouse.


### [From hobby to real world: integrating the hard stuff with agents](https://yomu.fyi/post/from-hobby-to-real-world-integrating-the-hard-stuff-with-agents.md)
- Company: [Stripe](https://yomu.fyi/company/stripe.md)
- Author: Anna Spysz
- Published: Jun 16, 2026
- Type: Tutorial

The post documents building a production-oriented leaderboard for open-source Stripe Projects applications, where users submit projects, share links, and vote. Unlike an earlier transcription app, this project required authentication, multiple database tables, hosting, and controls limiting unauthenticated users to one vote per project. The agent initially selected Supabase for database and auth and Vercel for hosting, then added Upstash Redis after an in-memory rate limiter proved unsuitable for serverless enforcement. Supabase's two-emails-per-hour authentication limit prompted a human-directed migration to Clerk, after the agent proposed an unsuitable Twilio solution. The resulting MVP used five resources across four providers, took more than a day, and received separate security and performance audits before review.


### [Databricks announces 2026 global partner awards](https://yomu.fyi/post/databricks-announces-2026-global-partner-awards.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Kori O'Brien, Stephen Orban
- Published: Jun 15, 2026
- Type: Announcement

Databricks announced its 2026 Partner Awards at Data + AI Summit, recognizing more than 65 achievements across its global partner network of over 8,000 organizations. The awards cover consulting and system integrators, independent software vendors, technical champions, learning and enablement, industry categories, and product-focused contributions. Accenture/Avanade received Global Partner of the Year for an eighth consecutive year, citing more than 15,000 trained practitioners, 9,500+ certified resources, and over 1,000 joint engagements. Other cited results include Capgemini’s Unity Catalog migration of 25,000 tables, 10,000 notebooks, and 2 petabytes across 150+ countries, while Kraken Technologies reduced data-processing costs eightfold and cut load times from three days to eight hours using Databricks and Delta Sharing. The announcement frames the winners’ work as supporting data and AI adoption through software, services, integrations, and consulting.


### [Announcing New OpenSharing and Marketplace capabilities for the AI era](https://yomu.fyi/post/announcing-new-opensharing-and-marketplace-capabilities-for-the-ai-era.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Akram Chetibi, Harish Gaur, Huey Han, Tia Chang, Mengxi Chen, Lin Zhou, DJ Sharkey
- Published: Jun 15, 2026
- Type: Announcement

Databricks announces new OpenSharing and Marketplace capabilities aimed at sharing data, AI assets, and partner applications without moving data. OpenSharing, a Linux Foundation project and evolution of Delta Sharing, adds vendor-neutral sharing for Agent Skills, AI models, unstructured data, Iceberg clients, Lakebase tables and change data feed, plus governed multi-cloud connectivity and agentic sharing through Genie Agent Sharing. The release also introduces identity resolution in Databricks Clean Rooms, where partners such as LiveRamp and Acxiom can work against protected first-party data without seeing raw customer records. Third-party apps are now available through Databricks Marketplace, allowing customers to deploy them in their workspaces without separate infrastructure, while providers gain distribution. Marketplace Commit Drawdown lets customers use existing Databricks universal commits to acquire Marketplace data and AI solutions.


### [Skip the learning curve: rethinking data migration for real outcomes](https://yomu.fyi/post/skip-the-learning-curve-rethinking-data-migration-for-real-outcomes.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Vijay Anala
- Published: Jun 15, 2026
- Type: Explainer

Data migrations are presented as high-risk, costly initiatives whose technical completion can delay adoption and strategic value, especially during infrequent warehouse transitions. The proposed alternative combines migration, modernization, and value creation in parallel, using experienced specialized partners and AI-enabled automation for code conversion, data-quality validation, and pipeline modernization. Rather than lifting and shifting legacy workloads, teams are urged to simplify architectures, retire unnecessary components, reduce technical debt, and align data with business needs while validating progress continuously. Progressive decommissioning reduces the “double-bubble” period in which old and new systems run together, helping costs fall as workloads move instead of waiting for final completion. The Migrate & Modernize Program connects organizations with partners, and the post reports faster cutovers, reduced migration costs, and complex workloads entering production ahead of schedule among early participants.


### [Salesforce signs definitive agreement to acquire Fin](https://yomu.fyi/post/salesforce-signs-definitive-agreement-to-acquire-fin.md)
- Company: [Intercom](https://yomu.fyi/company/intercom.md)
- Author: Eoghan McCabe
- Published: Jun 15, 2026
- Type: Announcement

Salesforce has signed a definitive agreement to acquire Fin for approximately $3.6 billion, with closing expected in the fourth quarter of Salesforce’s fiscal year 2027. Fin says it began as Intercom 15 years ago and, nearly four years ago, shifted toward modern LLMs to create and define the Customer Agents category. The companies plan to join forces, with Salesforce’s resources intended to help bring Fin to more consumers, while Fin says little will practically change for customers: the author will remain CEO and Des Traynor will continue running R&D. The announcement also points to recent work including the Apex model and the Operator internal agent, and describes the deal as a pivotal moment for Fin’s employees, founders, and leadership team.


### [How Wayfair cut ML model costs by 90% (twice!) with Cursor](https://yomu.fyi/post/how-wayfair-cut-ml-model-costs-by-90-twice-with-cursor.md)
- Company: [Cursor](https://yomu.fyi/company/cursor.md)
- Author: Cursor
- Published: Jun 15, 2026
- Type: Problem & solution

Wayfair’s Applied Research team used Cursor to automate and parallelize experiments for a product-tag validation model that checked attributes against images, descriptions, and reviews. The model was accurate but too expensive to run across the catalog. In December 2025, five researchers standardized the dataset and evaluation benchmark, then built and tested 110 distinct variants in four days; the winning architecture reduced inference costs by 94%, improved precision, and became the production baseline. In March 2026, the team benchmarked against that baseline, ran 140+ new experiments, and added genetic-algorithm searches, achieving another 90% cost reduction. The workflow used 20+ parallel agents, desktop and CLI surfaces, and cloud agents, leaving researchers to formulate hypotheses, interpret results, and choose candidates while Cursor implemented, ran, and measured variants.


### [How to Rent a Car in Grand Rapids, Michigan](https://yomu.fyi/post/how-to-rent-a-car-in-grand-rapids-michigan.md)
- Company: [Uber](https://yomu.fyi/company/uber.md)
- Published: Jun 15, 2026
- Type: Tutorial

Renting a car in Grand Rapids can provide flexible transportation around the city and for trips to nearby areas, including for travelers arriving at Gerald R. Ford International Airport. Uber Rent lets users compare vehicles from rental companies in the Uber app or online, with choices including economy cars and SUVs; available suppliers may include Budget, Hertz, and Avis, depending on location. The booking process covers entering pickup and drop-off details, reserving a vehicle, collecting it with a matching driver’s license and physical credit card, and returning it to the listed location. Renters must be at least 18, while a Young Driver Fee may apply under 25; personal insurance is not required, though supplier-provided options are available, and additional conditions may appear in the reservation terms.


[Newer posts](https://yomu.fyi/posts/page/112.md) · [Older posts](https://yomu.fyi/posts/page/114.md)