<!DOCTYPE html><html lang="en"><head><meta http-equiv="Content-Type" content="text/html charset=UTF-8"><meta charset="UTF-8"><meta name="viewport" content="width=device-width"><meta name="x-apple-disable-message-reformatting"><title>TLDR Data</title><meta name="color-scheme" content="light dark"><meta name="supported-color-schemes" content="light dark"><style type="text/css">
:root {
color-scheme: light dark; supported-color-schemes: light dark;
}
*,
*:after,
*:before {
-webkit-box-sizing: border-box; -moz-box-sizing: border-box; box-sizing: border-box;
}
* {
-ms-text-size-adjust: 100%; -webkit-text-size-adjust: 100%;
}
html,
body,
.document {
width: 100% !important; height: 100% !important; margin: 0; padding: 0;
}
body {
-webkit-font-smoothing: antialiased; -moz-osx-font-smoothing: grayscale; text-rendering: optimizeLegibility;
}
div[style*="margin: 16px 0"] {
margin: 0 !important;
}
table,
td {
mso-table-lspace: 0pt; mso-table-rspace: 0pt;
}
table {
border-spacing: 0; border-collapse: collapse; table-layout: fixed; margin: 0 auto;
}
img {
-ms-interpolation-mode: bicubic; max-width: 100%; border: 0;
}
*[x-apple-data-detectors] {
color: inherit !important; text-decoration: none !important;
}
.x-gmail-data-detectors,
.x-gmail-data-detectors *,
.aBn {
border-bottom: 0 !important; cursor: default !important;
}
.btn {
-webkit-transition: all 200ms ease; transition: all 200ms ease;
}
.btn:hover {
background-color: #f67575; border-color: #f67575;
}
* {
font-family: Arial, Helvetica, sans-serif; font-size: 18px;
}
@media screen and (max-width: 600px) {
.container {
width: 100%; margin: auto;
}
.stack {
display: block!important; width: 100%!important; max-width: 100%!important;
}
.btn {
display: block; width: 100%; text-align: center;
}
}
body,
p,
td,
tr,
.body,
table,
h1,
h2,
h3,
h4,
h5,
h6,
div,
span {
background-color: #FEFEFE !important; color: #010101 !important;
}
@media (prefers-color-scheme: dark) {
body,
p,
td,
tr,
.body,
table,
h1,
h2,
h3,
h4,
h5,
h6,
div,
span {
background-color: #27292D !important; color: #FEFEFE !important;
}
}
a {
color: inherit !important; text-decoration: underline !important;
}
</style><!--[if mso | ie]>
<style type="text/css">
a {
background-color: #FEFEFE !important; color: #010101 !important;
}
@media (prefers-color-scheme: dark) {
a {
background-color: #27292D !important; color: #FEFEFE !important;
}
}
</style>
<![endif]--></head><body class="">
<div style="display: none; max-height: 0px; overflow: hidden;">AI often fails not at the model layer but because the underlying data isn’t AI-ready. Barriers include semantic ambiguity </div>
<div style="display: none; max-height: 0px; overflow: hidden;">
<br>
</div>
<table align="center" class="document"><tbody><tr><td valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" class="container" width="600"><tbody><tr class="inner-body"><td>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr class="header"><td bgcolor="" class="container">
<table width="100%"><tbody><tr><td class="container">
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" style="margin-top: 0px;" width="100%"><tbody><tr><td style="padding: 0px;">
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div style="text-align: center;">
<span style="margin-right: 0px;"><a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Ftldr.tech%2Fdata%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/0hu8eq0L4JT4nliXujZyd1TuxEHsV3guVCkQo5ba7ew=422" rel="noopener noreferrer" target="_blank"><span>Sign Up</span></a>
|<span style="margin-right: 2px; margin-left: 2px;"><a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fadvertise.tldr.tech%3Futm_source=tldrdata%26utm_medium=newsletter%26utm_campaign=advertisetopnav/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/pcnwE1po22oRo5f-NWkMb_UCfJiH90p6KNGiGZUASv4=422" rel="noopener noreferrer" target="_blank"><span>Advertise</span></a></span>|<span style="margin-left: 2px;"><a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fa.tldrnewsletter.com%2Fweb-version%3Fep=1%26lc=1670a604-84b7-11f0-bcf5-55fc1d40139c%26p=9518a61c-91ed-11f0-8131-f301a8837e3d%26pt=campaign%26t=1757930790%26s=cc43fe779ef9302328cafa05a1c4fe765d912c8872562763887646021ca028cd/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/EbsPC2POSn2uslVbggbChOVKlgvXbBYfaNgdXEbdxt8=422"><span>View Online</span></a></span>
<br>
</span></div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="text-align: center;"><span data-darkreader-inline-color="" style="--darkreader-inline-color:#3db3ff; color: rgb(51, 175, 255) !important; font-size: 30px;">T</span><span style="font-size: 30px;"><span data-darkreader-inline-color="" style="color: rgb(232, 192, 96) !important; --darkreader-inline-color:#e8c163; font-size:30px;">L</span><span data-darkreader-inline-color="" style="color: rgb(101, 195, 173) !important; --darkreader-inline-color:#6ec7b2; font-size:30px;">D</span></span><span data-darkreader-inline-color="" style="--darkreader-inline-color:#dd6e6e; color: rgb(220, 107, 107) !important; font-size: 30px;">R</span>
<br>
</td></tr></tbody></table>
<br>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr id="together-with"><td align="center" height="20" style="vertical-align:middle !important;" valign="middle" width="100%"><strong style="vertical-align:middle !important; height: 100%;">Together With </strong>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Fdata-science-guide%3Fe=48754805%26hl=en/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/kqV8kqWdcul2SDh6o5rziUpsndEMKEDy1aJaHskWyNk=422"><img src="https://images.tldr.tech/googlecloud50.png" valign="middle" style="vertical-align: middle !important; height: 100%;" alt="Google"></a></td></tr></tbody></table>
<table style="table-layout: fixed; width:100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;">
<div style="text-align: center;">
<h1><strong>TLDR Data <span id="date">2025-09-15</span></strong></h1>
</div>
</td></tr></tbody></table>
<table style="table-layout: fixed; width:100%;" width="100%"><tbody><tr id="sponsy-copy"><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Fdata-science-guide%3Fe=48754805%26hl=en/2/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/BqOZcesbCUnCZug1g2WmPxNVWLkrezu7_DoDpknbckg=422">
<span>
<strong>Practical data science on Google Cloud - with 8 use cases you can build today (Sponsor)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Literally everyone is talking about AI. But in the real world, data scientists still spend too much time wrestling with infrastructure instead of building models.<p></p><p>This <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Fdata-science-guide%3Fe=48754805%26hl=en/3/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/65QdnYakiCYL1hFwrJLKJzHr86J-bJP1wsnvw_sa_3o=422" rel="noopener noreferrer nofollow" target="_blank"><span>practical guide by Google</span></a> shows how you can use BigQuery, Vertex AI, and other Google Cloud tools to automate routine tasks, use previously untapped unstructured data, and achieve new levels of efficiency.</p>
<p>What's inside:</p>
<ul>
<li>8 hands-on use cases - retail demand forecasting, customer segmentation, assessing environmental risks, and more.</li>
<li>Building <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Fdata-science-guide%3Fe=48754805%26hl=en/4/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/7tTn6nnqpDlImfdT5ZiLBAymVEBUM7fx8cQWzsZmYmE=422" rel="noopener noreferrer nofollow" target="_blank"><span>AI-first workflows</span></a>: vector search, multimodal analysis, and automated model deployment.</li>
<li>Architecture and code examples..</li>
</ul>
<p><a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Fdata-science-guide%3Fe=48754805%26hl=en/5/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/0A8CryW_uZlbfrbl2Q3Xl18zuE775IrEsOdT336BjVA=422" rel="noopener noreferrer nofollow" target="_blank"><span>Get the guide →</span></a>
</p>
</span></span></div>
</td></tr></tbody></table>
</td></tr></tbody></table>
</td></tr></tbody></table>
</td></tr>
<tr bgcolor=""><td class="container">
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td style="padding: 0px;">
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><span style="font-size: 36px;">📱</span></div></div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;">
<h1><strong>Deep Dives</strong></h1>
</div>
</div>
</td></tr></tbody></table>
<table style="table-layout: fixed; width: 100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;" valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fmoderndata101.substack.com%2Fp%2Fai-ready-data-a-technical-assessment%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/swUP9kk8TOJBfq2LwckHvtHVW9jD2KN8oULSYrt9f6w=422">
<span>
<strong>AI-Ready Data: A Technical Assessment. The Fuel and the Friction (14 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
AI often fails not at the model layer but because the underlying data isn't AI-ready. Barriers include semantic ambiguity, poor quality, temporal misalignment, and inconsistent formats. Fixes require AI-native pipelines with built-in validation, aligned semantics, standardized schemas, and time-aware data flows. The key shift is treating data as a governed product, enabling reliability, reusability, and trust essential for AI at scale.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fmedium.com%2Fpinterest-engineering%2Fnext-gen-data-processing-at-massive-scale-at-pinterest-with-moka-part-2-of-2-d0210ded34e0%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/wzDHAMeKXX3r8uw3nc8XB-yxoG-wIcki7VwK_jP0-MY=422">
<span>
<strong>Next Gen Data Processing at Massive Scale at Pinterest with Moka (14 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Moka, Pinterest's next-generation data processing platform built to replace the aging Hadoop-based Monarch system, has driven infrastructure modernization at Pinterest, including AWS multi-account setups, networking topology changes, pod-level identity management, and logging system enhancements. Moka also enables EKS adoption for other use cases like TiDB, Flink, Ray, and PyTorch.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fmaknee.github.io%2Fblog%2F2025%2FNetwork-And-Storage-Training-Skypilot%2F%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/zIYJjq-8uslefnRKFGqBbHpp8wghQWzvBJNs714vTdQ=422">
<span>
<strong>Network and Storage Benchmarks for LLM Training on the Cloud (5 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Network and storage infrastructure configuration has a decisive impact on distributed LLM training performance, yielding 6–10x speedups compared to default cloud settings. Benchmarks with Gemma 3 12B and GPT-OSS-120B showed that InfiniBand (400 GB/s) delivers 10x faster training than Ethernet, while optimized storage cuts checkpointing time in half. Benchmark results and configurations using Skypilot are linked in the article.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.confessionsofadataguy.com%2Fhow-to-tune-spark-shuffle-partitions%2F%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/Lc8Q16zZoHiRHplGt84J-jk1dhvR04BDDkBZryahpT4=422">
<span>
<strong>How to Tune Spark Shuffle Partitions (6 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Tuning Spark shuffle partitions is crucial to avoid performance issues: too few cause memory overload and slow tasks, while too many add scheduling overhead. Use Adaptive Query Execution (AQE) for automatic adjustments, aim for ~128 MB per partition, and monitor via Spark UI for skew or spills, while aligning with cluster cores (2–4 partitions per core) for optimal balance.
</span>
</span>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><span style="font-size: 36px;">🚀</span></div>
</div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;">
<h1><strong>Opinions & Advice</strong></h1>
</div>
</div>
</td></tr></tbody></table>
<table style="table-layout: fixed; width: 100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;" valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.infoq.com%2Fpresentations%2Fllm-data-code-model-prompt%2F%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/zpu3eiT59aUWcrufMqBDZkInLTBLN13iA4Oc5zrWCgQ=422">
<span>
<strong>The Data Backbone of LLM Systems (51 minute presentation)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Most AI engineering today is less about ML and more about data, retrieval, and system design. RAG dominates as the core pattern, with effective systems built around a feature–training–inference–observability architecture. Success hinges on decoupling ingestion from inference, versioning datasets, supporting reproducibility and experimentation, and treating observability as first-class. Agentic apps add memory and guardrails but remain software-engineering problems at heart.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.eckerson.com%2Farticles%2Fsix-kpis-every-leader-should-track-across-the-data-engineering-lifecycle%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/ZkEVl57JrfRbmwk64EyDm4FtdRSa6YRXYQv_Qw1bGxk=422">
<span>
<strong>Six KPIs Every Leader Should Track Across the Data Engineering Lifecycle (4 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Measuring data engineering productivity requires tracking six critical lifecycle KPIs: percentage of requests tied to business value, planning accuracy, time-to-delivery, pipeline success rate, time-to-resolve incidents, and data product reuse rate. These metrics expose bottlenecks across intake, planning, development, testing, deployment, and feedback stages, highlighting hidden friction points missed by output metrics alone.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.datasciencecentral.com%2Fhow-business-leaders-are-using-ai-to-make-data-driven-decisions%2F%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/y7C2N5CvmT9KVQry-jAGqXRD_CCEcK6r3dyoLp9CgkE=422">
<span>
<strong>How Business Leaders are Using AI to Make Data-driven Decisions (4 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
AI enhances predictive analytics to forecast trends and risks, supercharges business intelligence by automating data analysis, improves customer understanding through sentiment analysis and behavioral insights, streamlines operations like supply chain management, and fosters a data-driven culture across organizations.
</span>
</span>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><span style="font-size: 36px;">💻</span></div>
</div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;">
<h1><strong>Launches & Tools</strong></h1>
</div>
</div>
</td></tr></tbody></table>
<table style="table-layout: fixed; width: 100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;" valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.cdata.com%2Fevents%2Ffoundations-2025%2F%3Futm_source=tldr%26utm_medium=%26utm_campaign=25Q3_Foundations_Event/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/ikZQl3n1Js91DhCpXuh_WlpjUoRyjG0N09eT0kZkPZI=422">
<span>
<strong>You've got AI ambitions. Do you have the foundation to support them? (Sponsor)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
If you don't have a strong data foundation, you're building your analytics and AI on sand. At the upcoming <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.cdata.com%2Fevents%2Ffoundations-2025%2F%3Futm_source=tldr%26utm_medium=%26utm_campaign=25Q3_Foundations_Event/2/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/ke23quAtV34CUiRGTQtEwBl0UVLDNUM6d4K8jyPLS-Q=422" rel="noopener noreferrer nofollow" target="_blank"><span>CData Foundations 2025 event (free, virtual)</span></a>, you can hear directly from leaders at Google, AWS, Databricks, and ServiceNow on their approach to data strategy and AI-readiness. <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.cdata.com%2Fevents%2Ffoundations-2025%2F%3Futm_source=tldr%26utm_medium=%26utm_campaign=25Q3_Foundations_Event/3/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/ax_kF-5wie03X10MxHHTcIQ5vvR05qZYrkVLnD7MxF0=422" rel="noopener noreferrer nofollow" target="_blank"><span>See the full agenda</span></a>
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fspiraldb.com%2Fpost%2Fannouncing-spiral%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/Tqlw073-Bt5UqcUulboMFaPdawtdV1HPK7J-Z5E5t24=422">
<span>
<strong>Announcing Spiral (5 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Spiral introduces a new data system for the "Third Age" of AI-driven workloads, using the Vortex file format to deliver 10-20x faster scans and 100-200x faster random reads compared to Parquet, enabling direct S3-to-GPU data transfer to fully utilize NVIDIA H100 GPUs. Backed by $22 million from Amplify Partners and General Catalyst, Spiral offers unified governance, secure "fearless permissioning," and a single API to handle diverse data types.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.crunchydata.com%2Fblog%2Fget-excited-about-postgres-18%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/KTlD2K6T3Ujn_NR0tqJiD7LYPhP0N1K6LpN9gFJpAt4=422">
<span>
<strong>Get Excited About Postgres 18 (6 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Postgres 18, to be released later this month, will introduce several compelling features: asynchronous I/O for faster reads (especially for sequential scans and bitmap heap scans), UUID v7 for better locality/indexing of time-based UUIDs, B-tree skip scans allowing multi-column indexes to be used when leading columns aren't supplied (in some cases), virtual generated columns by default (reducing storage & write overhead), and native OAuth 2.0 support in pg_hba.conf. With 3,000 commits from 200+ contributors, even users who don't use all these features will benefit from general planner and performance improvements.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fgithub.com%2Fdocling-project%2Fdocling%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/b5Bxmih6JPq8vLkhz3Gv2MTn0zJbxti7YSCCwJnPxDU=422">
<span>
<strong>Dockling (GitHub Repo)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Docling processes diverse document formats (PDF, DOCX, PPTX, images, audio, etc.), extracting layout, tables, formulas, and more to produce a unified document representation. It integrates with tools like LangChain, LlamaIndex, and MCP servers for agentic workflows, supports air-gapped/local execution, and offers many export formats (Markdown, lossless JSON, etc.).
</span>
</span>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><span style="font-size: 36px;">🎁</span></div></div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><strong><h1>Miscellaneous</h1></strong></div>
</div>
</td></tr></tbody></table>
<table bgcolor="" style="table-layout: fixed; width: 100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;" valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.dataversity.net%2Fcomparing-eu-and-u-s-state-laws-on-ai-a-checklist-for-proactive-compliance%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/p763CT807rwj1ERuImc6N9TzZzufjdqQ3dGWWS2mY-I=422">
<span>
<strong>Comparing EU and U.S. State Laws on AI: A Checklist for Proactive Compliance (1 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
The EU AI Act creates the world's first comprehensive, enforceable AI regulation with a risk-tiered framework, sweeping obligations for high-risk and foundation models, and strict fines, while the U.S. pursues a patchwork of lighter, state-level laws. For multinationals, chasing minimum compliance across jurisdictions adds cost and risk. A more strategic path is adopting “EU-plus” governance globally (i.e., meeting the EU bar everywhere) to simplify operations, build trust, and turn compliance into a market differentiator.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.databricks.com%2Fblog%2Fkey-production-ai-agents-evaluations%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/redh9dJlK_hHK8y5gzOuxmt1J_pt_4ModcLRoxZf0r4=422">
<span>
<strong>The Key to Production AI Agents: Evaluations (5 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
AI Agents with generic metrics and informal "vibe checks" fail to address business-specific needs, leading to stalled GenAI projects. Effective evaluation involves task-level benchmarking, grounded assessments using enterprise context, and change tracking to ensure consistent, trustworthy performance, transforming agents into continuously improving systems.
</span>
</span>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;"><span style="font-size: 36px;">⚡</span></div></div>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding-top: 0px; padding-bottom: 0px;">
<div class="text-block">
<div style="text-align: center;">
<h1><strong>Quick Links</strong></h1>
</div>
</div>
</td></tr></tbody></table>
<table bgcolor="" style="table-layout: fixed; width: 100%;" width="100%"><tbody><tr><td style="padding:0;border-collapse:collapse;border-spacing:0;margin:0;" valign="top">
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fplassard.substack.com%2Fp%2Fthe-coding-benchmark-leaderboard%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/CxdTn_URYE5vDv7OWIerdzRTm2PqAS0f8SMspDYd_bI=422">
<span>
<strong>Beyond Accuracy: Building Cost-Conscious AI Benchmarks (3 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
AI coding benchmarks traditionally favor larger models, but a new approach, "Pass @ Budget," reveals that smaller, cost-efficient models outperform them when given a fixed budget for problem-solving.
</span>
</span>
</div>
</td></tr></tbody></table>
<table align="center" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block">
<span>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.getdbt.com%2Fblog%2Fwhat-to-expect-from-sessions-at-coalesce-2025%3Futm_source=tldrdata/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/Q1QyiUq2Zi5ltnYpjj6OPJCv7pAQzHqrovpe1hcWMjo=422">
<span>
<strong>What to Expect from Sessions at Coalesce 2025 (3 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Coalesce 2025 offers practitioner-focused breakout tracks addressing key data challenges such as platform modernization, AI readiness, scaling development with dbt, self-service analytics enablement, process optimization, and governed data quality backed by real-world case studies.
</span>
</span>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td align="left" style="word-break: break-word; vertical-align: top; padding: 5px 10px;">
<p style="padding: 0; margin: 0; font-size: 22px; color: #000000; line-height: 1.6; font-weight: bold;">
Want to advertise in TLDR? 📰
</p>
<div class="text-block" style="margin-top: 10px;">
If your company is interested in reaching an audience of data engineering professionals and decision makers, you may want to <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fadvertise.tldr.tech%2F%3Futm_source=tldrdata%26utm_medium=newsletter%26utm_campaign=advertisecta/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/3ThnTmMPwxovpLoEs0_M_0yrX6s2RqdzR0lffp_knNA=422"><strong><span>advertise with us</span></strong></a>.
</div>
<br>
<!-- New "Want to work at TLDR?" section -->
<p style="padding: 0; margin: 0; font-size: 22px; color: #000000; line-height: 1.6; font-weight: bold;">
Want to work at TLDR? 💼
</p>
<div class="text-block" style="margin-top: 10px;">
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fjobs.ashbyhq.com%2Ftldr.tech/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/DhRHfZ9QHvFHanPOBl4ObmSLv0xlh9bs0LMXqHFHOA4=422" rel="noopener noreferrer" style="color: #0000EE; text-decoration: underline;" target="_blank"><strong>Apply here</strong></a> or send a friend's resume to <a href="mailto:jobs@tldr.tech" style="color: #0000EE; text-decoration: underline;">jobs@tldr.tech</a> and get $1k if we hire them!
</div>
<br>
<div class="text-block">
If you have any comments or feedback, just respond to this email!
<br>
<br> Thanks for reading,
<br>
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.linkedin.com%2Fin%2Fjoelvanveluwen%2F/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/GzKQhewAMhmVFPs0bC9qk815-d93WkkaegUqM3H-98c=422"><span>Joel Van Veluwen</span></a>, <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.linkedin.com%2Fin%2Fjennytzurueyching%2F/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/rN4eXtnlSR_Z275O_Oees-3pdz2TiyfdTLc7OhmH5eo=422"><span>Tzu-Ruey Ching</span></a> & <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.linkedin.com%2Fin%2Fremi-turpaud%2F/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/dp2qKxeKjkV6I5o69E_wnUm5gAJGegdPZt_WyGQTC20=422"><span>Remi Turpaud</span></a>
<br>
<br>
</div>
<br>
</td></tr></tbody></table>
<table align="center" bgcolor="" border="0" cellpadding="0" cellspacing="0" width="100%"><tbody><tr><td class="container" style="padding: 15px 15px;">
<div class="text-block" id="testing-id">
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Ftldr.tech%2Fdata%2Fmanage%3Femail=silk.theater.56%2540fwdnl.com/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/YlIb77t3z5wU_whIkD2WBIhAmBa3B7ZfHp081j_wc2o=422">Manage your subscriptions</a> to our other newsletters on tech, startups, and programming. Or if TLDR Data isn't for you, please <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fa.tldrnewsletter.com%2Funsubscribe%3Fep=1%26l=037ede50-92cc-11ee-b0f2-b761aa2217ad%26lc=1670a604-84b7-11f0-bcf5-55fc1d40139c%26p=9518a61c-91ed-11f0-8131-f301a8837e3d%26pt=campaign%26pv=4%26spa=1757930489%26t=1757930790%26s=8df6a64d94298ee475ed732715a55dfbd08f594b6fa67401bd8c888b7826910f/1/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/OHKLuQxXIihVI8qx7_nST5hv9aKKWO-yaJZpgj8YqXc=422">unsubscribe</a>.
<br>
</div>
</td></tr></tbody></table>
</td></tr></tbody></table>
</td></tr></tbody></table>
</td></tr></tbody></table>
</td></tr></tbody></table>
<img alt="" src="http://tracking.tldrnewsletter.com/CI0/010001994cd71d60-4f4bb85c-1e8f-48d8-9466-46d3c4607e8f-000000/duyhnyL7289T6BaqT6GqwiUCUfUZ0Mmp8ams4LEcaEA=422" style="display: none; width: 1px; height: 1px;">
</body></html>