---
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

### [How we keep GPUs reliable across Databricks AI](https://yomu.fyi/post/how-we-keep-gpus-reliable-across-databricks-ai.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Steven Chen, Feng Wang, Bhavik Soni, Chengguang Yang, Albert Zhong, Naren Loganathan, Harsh Panchal, Jianwei Xie
- Published: Jul 1, 2026
- Type: Explainer

Databricks AI describes GPU reliability as a fleet-scale problem involving crashed jobs, silent slowdowns, and numerical corruption during distributed training. Its approach combines stress testing with demanding reinforcement learning, agentic coding, and document intelligence workloads, plus gpu-monitor, a multi-stage service that checks nodes during bootstrap, continuously while active, and periodically between workloads. The checks cover compute burn-in, peer connectivity, NCCL correctness and bandwidth, RDMA, ECC and HBM memory, PCIe integrity, throttling, thermal conditions, XID errors, and fabric health. A production incident showed why cumulative InfiniBand port downtime matters: a single outage exceeded the lower-level NCCL\_IB\_TIMEOUT, killing a collective before the longer PyTorch watchdog timeout. Nodes that fail checks are quarantined, while inter-node probes sweep message sizes from 8 bytes to 2 GiB and apply latency or BusBW criteria; the system is continuously expanded as new failure modes emerge.


### [Celebrating the Winners of the 2026 Built-On Databricks Startup Challenge](https://yomu.fyi/post/celebrating-the-winners-of-the-2026-built-on-databricks-startup-challe.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Andrew Ferguson, Joslyn O'Connell, Sully Clark
- Published: Jul 1, 2026
- Type: Announcement

The 2026 Built-On Databricks Startup Challenge recognized early-stage startups building B2B applications on Databricks, with winners presented at the 2026 Data + AI Summit. VisionHeight won the Grand Prize for an agentic threat-intelligence platform that maps adversary infrastructure across the Internet while it is being constructed, aiming to give defenders earlier warning. Linkup took second place with a production-grade Web Search API that independently crawls and indexes the open web at the fact level, delivering sourced results in about two seconds. Intelo placed third with five coordinated AI Agent Teams covering retail merchandising and planning, while Clarecast, Gemini Sports, and LakeFusion received Honorable Mentions. Judges assessed market potential, founding-team caliber, and innovative Databricks use; the announced Startup Program also offers qualifying startups up to $200,000 in combined Databricks and Neon credits.


### [Securing Heroku CLI Credentials with System Keychain Storage](https://yomu.fyi/post/securing-heroku-cli-credentials-with-system-keychain-storage.md)
- Company: [Heroku](https://yomu.fyi/company/heroku.md)
- Author: Katy Bowman
- Published: Jul 1, 2026
- Type: Announcement

Heroku CLI 11.8.0 changes credential storage by making the system keychain the default and retaining .netrc as a fallback. The transition began in 11.6.0, when the CLI stored tokens in both locations to preserve existing workflows. The credential manager uses macOS security, Linux Secret Service through secret-tool, or Windows PasswordVault, storing credentials under “heroku-cli” with the account email as the identifier. Authentication lookup prioritizes HEROKU\_API\_KEY, then HEROKU\_NETRC\_WRITE, keychain storage, and .netrc files; unsupported or failing keychain access falls back to .netrc. Version 11.8.0 also adds a Git credential helper, so commands such as git push heroku main continue without configuration changes, while CI/CD can keep using HEROKU\_API\_KEY and Linux users may need secret-tool.


### [Redeploying Claude Fable 5](https://yomu.fyi/post/redeploying-claude-fable-5.md)
- Company: [Anthropic](https://yomu.fyi/company/anthropic.md)
- Published: Jul 1, 2026
- Type: Announcement

Access to Claude Fable 5 and Mythos 5 was restored after US export controls had temporarily suspended both models for all users. The post says Fable 5 will return globally on July 1 across Claude Platform, Claude.ai, Claude Code, and Claude Cowork, while Mythos 5 access has resumed for some US organizations and broader partner access remains under discussion. Anthropic describes a new safety classifier intended to block the bypass reported by Amazon researchers in more than 99% of cases; blocked Fable 5 requests are sent to Opus 4.8, although the classifier increases false positives in coding and debugging. It also outlines defense-in-depth safeguards, a proposed shared industry framework for assessing jailbreaks, and expanded government collaboration on pre-release evaluation, information sharing, and joint research.


### [Claude Fable 5 and Claude Mythos 5](https://yomu.fyi/post/claude-fable-5-and-claude-mythos-5.md)
- Company: [Anthropic](https://yomu.fyi/company/anthropic.md)
- Published: Jul 1, 2026
- Type: Announcement

Claude Fable 5 and Claude Mythos 5 are presented as a joint launch of the same underlying model, with Fable 5 intended for general use and Mythos 5 initially reserved for selected cyberdefenders and infrastructure providers. Fable 5 uses conservative safeguards that route some sensitive queries to Claude Opus 4.8, while Mythos 5 lifts safeguards in selected areas through Project Glasswing and a planned trusted-access program. The post reports strong results in software engineering, knowledge work, vision, long-context tasks, cybersecurity, protein design, molecular biology, and genomics, including a reported 50-million-line Ruby migration completed in a day. Both models are priced at $10 per million input tokens and $50 per million output tokens, with Fable 5 available through the Claude API and staged subscription access.


### [Granular Usage Attribution for dbt Pipelines with Query Tags](https://yomu.fyi/post/granular-usage-attribution-for-dbt-pipelines-with-query-tags.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Heeren Sharma, Lennart Reschke, JooHo Yeo
- Published: Jul 1, 2026
- Type: Tutorial

Databricks Query Tags give dbt pipelines granular usage attribution when identical query-history rows obscure model ownership, cost, and performance. The dbt-databricks adapter version 1.11+ automatically injects model, materialization, dbt-core, and adapter-version metadata, while profile-level and model-level configurations add custom dimensions such as team, cost center, project, and environment. These tags are stored as a MAP<STRING, STRING> in system.query.history, where SQL can extract them for cost attribution, performance debugging, monitoring, and dashboards. In the reference project, four mart tables accounted for 92% of compute time, whereas staging views and metric views were near-instantaneous. The accompanying open-source project demonstrates Genie exploration, a self-monitoring AI/BI dashboard, metric-view tagging, and deployment with Databricks bundles, while recommending consistent profile hierarchies, environment tags, and limited custom metadata.


### [Beyond dashboards: Introducing Decision Execution Platforms](https://yomu.fyi/post/beyond-dashboards-introducing-decision-execution-platforms.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Marc Solomon, Marcello Pedersen
- Published: Jul 1, 2026
- Type: Explainer

Databricks Forward Deployed Engineering introduces Decision Execution Platforms (DEPs), an enterprise analytics category intended to connect KPI signals, executive decisions, operational execution, and measured outcomes. The proposal addresses workflows in which dashboards reveal problems but meetings, decks, spreadsheets, and messaging threads leave implementation fragmented and impact measurement disconnected. DEPs run the four-stage loop on governed Databricks infrastructure: agents recommend actions, alternatives, predicted impact, and reasoning; approved choices execute through systems of record; and results persist in a Decision Log for continuous learning. Their architecture combines a foundation of Lakebase, Genie, Unity Catalog, Lakehouse, Agent Bricks, and MLflow with an SDK of reusable primitives and a Databricks Apps executive surface. A retailer case used a DEP to unify fulfillment data and enable simulated, controlled rerouting, with scaling aimed at measurable bottom-line and customer-satisfaction outcomes.


### [Mastering Agentic Techniques: AI Agent Reinforcement Learning](https://yomu.fyi/post/mastering-agentic-techniques-ai-agent-reinforcement-learning.md)
- Company: [NVIDIA](https://yomu.fyi/company/nvidia-developer-blog.md)
- Author: Elizabeth Goodman
- Published: Jul 1, 2026
- Type: Tutorial

Language model agents often fail across long-horizon workflows, repeat tool-call errors, and output invalid schemas when relying solely on prompting or retrieval. Reinforcement learning with verifiable rewards provides a practical training signal for domain-specific tasks by scoring model rollouts against deterministic verifiers such as schema checkers, unit tests, and execution sandboxes. Teams can implement group relative policy optimization to sample multiple trajectories per prompt and update policy weights based on relative performance within the group. Establishing environments with explicit harness, state, and verifier definitions allows agents to learn multi-step tool use, whereas static datasets fail to capture sequential dependencies. Inspecting checkpoint rollouts and testing against held-out tasks prevents reward hacking while continuously turning operational failures into verifiable benchmark environments.


### [How to measure the customer experience as AI scales](https://yomu.fyi/post/how-to-measure-the-customer-experience-as-ai-scales.md)
- Company: [Intercom](https://yomu.fyi/company/intercom.md)
- Author: William Lane
- Published: Jul 1, 2026
- Type: Explainer

CSAT captures fewer than 10% of conversations, and its responses skew toward delighted or furious customers. Most customers say nothing, leaving teams to assess customer experience from an unrepresentative sample that compresses product, policy, and service issues into one score. As AI handles more conversations end-to-end, the post argues for scoring every AI and human interaction on service quality, resolution, customer effort, and related attributes rather than relying on survey responses alone. Fin’s CX Score assigns each interaction a 1–5 score and surfaces reasons behind it, providing roughly five times more coverage than CSAT for most teams. Fin sets targets from CX Score attributes, operational correlations, historical performance, modelling, and team ambition, then uses recurring patterns to route fixes to support, product, or workflow owners.


### [Meta’s AI Storage Blueprint at Scale](https://yomu.fyi/post/meta-s-ai-storage-blueprint-at-scale.md)
- Company: [Meta](https://yomu.fyi/company/meta.md)
- Author: By Sidharth Bajaj, Venkatraghavan Srinivasan
- Published: Jul 1, 2026
- Type: Explainer

Meta describes how its BLOB-storage architecture evolved to address AI workloads' storage bottlenecks, which can stall GPUs and slow research iteration as datasets and GPU deployments expand. Legacy requests traversed multiple stateful metadata layers, sometimes across regions, producing latencies of hundreds of milliseconds even when flash made millisecond access possible. The redesign consolidates metadata into a flat schema backed by ZippyDB for O(1) path resolution, removes the dataplane proxy in favor of a fat client SDK, and deploys regional BLOB-storage stacks beside GPUs. For data loading, Meta combines host memory and flash with regional flash and global HDD-backed storage as tiered caches, using ordinary and deep prefetch, on-demand hydration, and TTL or LRU eviction policies. Production rollout saw rapid adoption, while both data-loading paradigms remain supported.


### [How Amazon tracks carbon intensity across its operations](https://yomu.fyi/post/how-amazon-tracks-carbon-intensity-across-its-operations.md)
- Company: amazon.science
- Author: Kerry Constabile
- Published: Jul 1, 2026
- Type: Explainer

Amazon measures carbon intensity across its diverse business operations to track progress toward its goal of reaching net-zero carbon by 2040. For retail operations, the company monitors carbon emissions per unit shipped, achieving a 39 percent reduction by the end of 2025 compared to 2019 baselines. This per-unit decrease stems from targeted investments in electric delivery vehicles, carbon-free energy, smarter package routing, lightweight packaging, low-carbon fuels, and alternative transport modes like rail. To improve industry comparability and transparency, the company updated its economic carbon intensity indicator in 2025 from emissions per dollar of gross merchandise sales to emissions per dollar of revenue. Carbon intensity is also evaluated regionally and nationally to isolate geographic variations and implement localized decarbonization interventions across supply chain networks.


### [How we scale PgBouncer in ClickHouse Managed Postgres](https://yomu.fyi/post/how-we-scale-pgbouncer-in-clickhouse-managed-postgres.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jul 1, 2026
- Type: Benchmark

PgBouncer is single-threaded, so a single process can bottleneck connection pooling on multi-vCPU machines even when Postgres has capacity. ClickHouse Managed Postgres runs a fleet of PgBouncer processes sized to available cores, sharing one port through so\_reuseport; peering forwards query-cancellation requests to the process that owns the session, while transaction pooling and per-process connection-budget splits limit Postgres connections. In tests on identical AWS EC2 c7i.4xlarge pooler instances, a 16-process fleet reached roughly 336,000 transactions per second at 256 clients, compared with 76,893 for one process, while using about 49% versus 8% in-guest CPU. The fleet's advantage emerged under concurrency, whereas the single process was slightly faster at eight clients, and the described setup ships by default with every ClickHouse Managed Postgres server.


### [Forecasting at the speed of modern retail](https://yomu.fyi/post/forecasting-at-the-speed-of-modern-retail.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Ryuta Yoshimatsu, Puneet Jain, Lourdes Angélica Martinez Medina, Lucas Bruand, Dael Williamson
- Published: Jul 1, 2026
- Type: Announcement

Retail and CPG forecasting now spans hundreds of thousands, sometimes far more, time series across fragmented channels, promotions, and short-lived products, making legacy methods and manual exception management difficult. Multi-model forecasting addresses this complexity by evaluating a range of techniques against actual data and selecting the best-performing model for each series, but enterprise-scale experiments require scarce forecasting and distributed-systems expertise. Released in 2024, Databricks’ open-source Many Model Forecasting (MMF) integrates more than 35 statistical, deep-learning, and foundation time-series models and runs on distributed Databricks compute. MMF Agent, built on Genie Code, guides users through data quality, series classification, compute configuration, forecasting, post-processing, and model selection, while Unity Catalog helps it use organizational data context. The workflow is intended to reduce setup from days to hours, improve targeting and accuracy, and make rigorous forecasting more accessible while remaining customizable for technical teams.


### [How Artemis Security runs 69x faster detection queries with ClickHouse Cloud](https://yomu.fyi/post/how-artemis-security-runs-69x-faster-detection-queries-with-clickhouse.md)
- Company: [Clickhouse](https://yomu.fyi/company/clickhouse.md)
- Published: Jul 1, 2026
- Type: Benchmark

Artemis Security uses ClickHouse Cloud on AWS for real-time threat detection across many terabytes of compressed logs, tens of billions of daily rows, and queries over trillions of rows. To reduce repeated scans, its engineers coalesce hundreds of detection-rule predicates into one query, tagging matching rows by rule ID; batches of more than 100 rules now run in under a second, with a measured 69x speedup. A Claude Code skill called /poke connects ClickHouse system tables, CloudWatch, Datadog, application data, and source code to investigate infrastructure issues through natural-language prompts. For investigative queries over semi-structured JSON, Artemis extracts frequently queried fields into a parallel materialized table and uses a hybrid two-step lookup, producing 30–60x faster queries and up to 400x lower CPU consumption.


### [Celebrating the Winners of the 2026 Built-On Databricks Startup Challenge - Cloned](https://yomu.fyi/post/celebrating-the-winners-of-the-2026-built-on-databricks-startup-challe-cloned.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Andrew Ferguson, Joslyn O'Connell, Sully Clark
- Published: Jul 1, 2026
- Type: Announcement

The 2026 Built-On Databricks Startup Challenge recognized early-stage startups building B2B applications on the Databricks platform, with winners announced at the Data + AI Summit in San Francisco on June 16. VisionHeight won the grand prize for an agentic threat-intelligence platform that maps adversary infrastructure across the Internet while it is being built; Linkup placed second with a production-grade Web Search API delivering results in about two seconds, and Intelo placed third with an agentic workforce for retail merchandising and planning. Clarecast, Gemini Sports, and LakeFusion received honorable mentions for predictive intelligence, football squad planning, and an AI-native data foundation combining MDM, PIM, and LakeGraph. Judges assessed market potential, founding-team caliber, and innovative use of Databricks, while the startup-program offer provides qualifying startups up to $200,000 in credits across Databricks and Neon.


### [Reading the agent traces is how you make the call your eval can't](https://yomu.fyi/post/reading-the-agent-traces-is-how-you-make-the-call-your-eval-can-t.md)
- Company: [Sentry](https://yomu.fyi/company/sentry.md)
- Author: Sergiy Dybskiy
- Published: Jul 1, 2026
- Type: Explainer

Automated evaluations often miss unknown failure modes in AI agents because developers only write tests for anticipated errors. While building an itinerary agent for AI Engineer World's Fair, an open-weight model routed a speaker query to a tool returning only tracks, leading the model to hallucinate famous speakers and falsely claim the tool provided that data. Because a tool was executed, the trace appeared grounded until inspectable tool outputs revealed the discrepancy. To resolve this, the author retained the inexpensive model, corrected query routing, tightened system prompts against ungrounded entities, and added a custom entity-diffing eval in CI. Ultimately, directly reading agent telemetry traces alongside errors and logs is essential for diagnosing root causes and deciding trade-offs between model costs, prompt adjustments, and routing fixes.


### [Multi-token Residual Prediction](https://yomu.fyi/post/multi-token-residual-prediction.md)
- Company: [Modal](https://yomu.fyi/company/modal.md)
- Published: Jul 1, 2026
- Type: Explainer

Multi-Token Residual Prediction (MRP) adapts multi-token prediction for diffusion language models by training a small module to predict the residual between adjacent denoising steps rather than the full next-step distribution. This residual target is easier to learn, allowing a tiny 3-layer transformer to accurately predict corrections across multiple steps. In the static regime, MRP enables lossless speculative decoding with up to 1.56× throughput in SGLang, or direct decoding with a small quality cost. In the dynamic regime, MRP recovers up to +16 accuracy points lost by aggressive low-threshold decoding through a remasking mechanism. The module is trained on a frozen backbone and composes cleanly with existing inference methods.


### [Announcing our $800M Series C to accelerate the shift to open-source AI](https://yomu.fyi/post/announcing-our-800m-series-c-to-accelerate-the-shift-to-open-source-ai.md)
- Company: [Together AI](https://yomu.fyi/company/together-ai.md)
- Published: Jul 1, 2026
- Type: Announcement

Together AI announces an $800M Series C to accelerate its open-source AI platform as production demand for inference grows. The company says the round includes $800 million from investors including Aramco Ventures, NVIDIA, Vista Equity, General Catalyst, and others, plus commitments for more than 500 MW of compute capacity to be capitalized independently by new investors. Together argues that closed frontier-model economics become unsustainable as agents and workflows scale, while open-weights models such as DeepSeek, Nemotron, MiniMax, Kimi, and GLM can provide equal or better performance at lower cost. Its full-stack approach spans models, kernels, compilers, inference systems, training infrastructure, hardware utilization, and software, with cited work including FlashAttention-4, Together Megakernel, together.compile, and expanded post-training APIs for tool calling, reasoning, and vision-language models.


### [Build agentic full-stack apps with Genkit](https://yomu.fyi/post/build-agentic-full-stack-apps-with-genkit.md)
- Company: [Google](https://yomu.fyi/company/google.md)
- Author: Chris Gill
- Published: Jul 1, 2026
- Type: Announcement

Genkit Agents is an open-source, full-stack foundation for conversational AI and agentic applications, supporting TypeScript, Go, Dart, and Python. The Agents API packages message history, tool loops, streaming, persistence, and frontend communication behind a shared interface: agents can run in process or behind HTTP, and clients use chat() through remoteAgent(). Developers can choose client-managed state or session-store-backed snapshots, resume by session ID, branch from snapshot ID, stream text, custom state, and artifacts, and pause tools for human approval. The framework also supports detached long-running work, specialist delegation, persistence choices including in-memory, file, Firestore, and custom stores, plus an Agent Runner in the Developer UI. The Agents API is in preview for TypeScript and Go, with breaking changes possible in minor releases.


### [ML Development in VS Code with Google Cloud Power: Workbench Extension Now Available](https://yomu.fyi/post/ml-development-in-vs-code-with-google-cloud-power-workbench-extension.md)
- Company: [Google](https://yomu.fyi/company/google.md)
- Author: Andrii Lobanov, Alex Kallaur, Diego Granados
- Published: Jul 1, 2026
- Type: Announcement

Google Cloud has launched the Google Cloud Workbench Notebooks extension for VS Code, allowing developers to use managed cloud notebook environments from a local IDE. Workbench, described as a managed Jupyter platform optimized for data science, can provide access to Google Cloud's AI-optimized infrastructure and high-performance compute when local machines need more capacity. The workflow connects through VS Code's Extensions view, the Jupyter extension, and a notebook's Select Kernel menu, where users choose Google Cloud, Workbench, authenticate, select a project, and choose an active instance. The extension runs notebooks on Workbench instances, preserves local settings and extensions, is fully open-sourced, and is available through the VS Code Marketplace, with GitHub contributions invited.


[Newer posts](https://yomu.fyi/posts/page/101.md) · [Older posts](https://yomu.fyi/posts/page/103.md)