<!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 AI</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;">An employee-only chatbot, codenamed βVeritasβ, is speculated to be Appleβs answer to ChatGPT, but itβs unknown if or when itβll be released publicly β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β β </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%2Fai%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/JlUy91IAbz1duq9zB5-_gqsdj900Ac--3rIGsJIolX0=424" 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%2F%3Futm_source=tldrai%26utm_medium=newsletter%26utm_campaign=advertisetopnav/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/Ru-yXq_ZOaTz4xxSfKEC9oTGs9TeTZPyGrtOft21KaQ=424" 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=1571a9f6-84b7-11f0-b464-47c5c04ad337%26p=e2677a24-9d2d-11f0-85b9-df7ce12a65ad%26pt=campaign%26t=1759151752%26s=f5a140c313fd910ed827165db1a5b102fe892f0a808d2343b1419f4355d7b458/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/KMZOpgRJfcgJtnqSFMgd_c_nW1PiuG_610YDMZ-9Bmo=424"><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%2Freal-world-generative-ai-adoption%3Futm_source=cloud_sfdc%26utm_medium=email%26utm_campaign=FY25-Q2-GLOBAL-ENT34861-website-dl-aiadoption-63328%26utm_content=tldr%26utm_term=september_29%26hl=en/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/QrcLw5xtxNUdcR7arirFqVhuh_Xc2Q_qQ1-O0aKNH8E=424"><img src="https://images.tldr.tech/googlecloud50.png" valign="middle" style="vertical-align: middle !important; height: 100%;" alt="Google Cloud"></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 AI <span id="date">2025-09-29</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%2Freal-world-generative-ai-adoption%3Futm_source=cloud_sfdc%26utm_medium=email%26utm_campaign=FY25-Q2-GLOBAL-ENT34861-website-dl-aiadoption-63328%26utm_content=tldr%26utm_term=september_29%26hl=en/2/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/ng0iklBvRRZO6ZC59GCnDbjZQCDpmhimwnop27V0wa4=424">
<span>
<strong>9 hard-learned lessons from implementing Gen AI in enterprise environments (Sponsor)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Many companies struggle to reconcile exciting AI demos and actual production value. Where are things going wrong - and what are leaders doing right?<p></p><p>Drawing from real customer engagements from Google Cloud's consulting team - including at Commerzbank, Erste Group, and SIGNAL IDUNA - this <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Freal-world-generative-ai-adoption%3Futm_source=cloud_sfdc%26utm_medium=email%26utm_campaign=FY25-Q2-GLOBAL-ENT34861-website-dl-aiadoption-63328%26utm_content=tldr%26utm_term=september_29%26hl=en/3/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/sEkGFGqICJEdSK2j3UruQKqEr3kYL4p9refGEsI_Ris=424" rel="noopener noreferrer nofollow" target="_blank"><span>ebook</span></a> exposes 9 critical lessons for scaling generative AI.</p>
<p>Download the guide to learn why <a class="underline" href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Freal-world-generative-ai-adoption%3Futm_source=cloud_sfdc%26utm_medium=email%26utm_campaign=FY25-Q2-GLOBAL-ENT34861-website-dl-aiadoption-63328%26utm_content=tldr%26utm_term=september_29%26hl=en/4/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/DT6mdm2Gx3nElHZUwgMZBSOF0hg6J_KorHctLMDGHso=424" rel="noopener noreferrer nofollow" target="_blank"><span>data quality matters more than model sophistication</span></a>, specific strategies that boost adoption by 125-451%, and how to build developer trust in AI systems.</p>
<p>Learn from real-world scenarios and expert guidance to avoid common pitfalls and accelerate your path to gen AI success. <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fcloud.google.com%2Fresources%2Fcontent%2Freal-world-generative-ai-adoption%3Futm_source=cloud_sfdc%26utm_medium=email%26utm_campaign=FY25-Q2-GLOBAL-ENT34861-website-dl-aiadoption-63328%26utm_content=tldr%26utm_term=september_29%26hl=en/5/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/x0T4PO00oUZ3coKjqykdchd6Kn6zXU-AzmFXTFbFWoY=424" rel="noopener noreferrer nofollow" target="_blank"><span>Get the guide βοΈ</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 style="font-size:36px;">π</span></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>Headlines & Launches</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.theverge.com%2Fnews%2F787046%2Fapples-veritas-siri-ai-chatbot%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/zHGd_2dWpcrU_VCGCaYwvhfI2q1rBNd7JsOGiB4TUw4=424">
<span>
<strong>Apple's βVeritas' chatbot is reportedly an employee-only test of Siri's AI upgrades (1 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
An internal employee-only chatbot, codenamed βVeritasβ, is speculated to be Apple's answer to ChatGPT, but it's unknown if or when it'll be released publicly. The company has struggled to get its AI efforts off the ground after a lukewarm reception to its Apple Intelligence features and multiple delays to Siri upgrades.
</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%2Fventurebeat.com%2Fai%2Fdeepseek-v3-1-terminus-launches-with-improved-agentic-tool-use-and-reduced%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/i37IDvRbXVtDhrzXGA_3g4v0BKbE-Cunr6Mvz1mTOI4=424">
<span>
<strong>DeepSeek-V3.1-Terminus launches with improved agentic tool use and reduced language mixing errors (2 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
DeepSeek released V3.1-Terminus, which strengthens its Code Agent and Search Agent and addresses users' feedback about Chinese/English mix-ups. The model shows benchmark gains in agentic tool tasks (BrowseComp, SWE Verified, etc.) and offers both βchatβ and βreasonerβ modes for different use cases. Terminus remains open-weight under an MIT license, making it viable for customization and deployment both via API and local environments.
</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%2Flinks.tldrnewsletter.com%2FfOI08H%26utm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/AolGpNVTemtjQ6DXIzgPihHmtNaVSPp96K_Qef4XStI=424">
<span>
<strong>Head of ChatGPT addresses controversial model routing (1 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
OpenAI confirmed it's testing a safety system that automatically switches users from GPT-4o to reasoning models or GPT-5 when conversations touch on "sensitive and emotional topics." Users have complained that the routing triggers too frequently in harmless situations, pulling them away from GPT-4o, which many prefer for personal conversations.
</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;">π§ </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 & Analysis</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.lesswrong.com%2Fposts%2FKmH83cDTh54EjTjbi%2Fthe-ai-village-in-numbers%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/CYeu7TbtiOeLF10wT-zQPiFYhdQPbXjbPEUlj4TdOHU=424">
<span>
<strong>The AI Village in Numbers (7 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
OpenAI's models lead on various stylistic measures like lexical diversity, positive sentiment, and formality of language, but Anthropic's models lead in goal-directed behavior. While OpenAI's models vary in which stylistic measures they excel at, Anthropic's models all perform well on agentic behavior. OpenAI offers the brightest, most cheerful, and most eloquent model. Anthropic's models tend to excel at getting stuff done.
</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%2Fcdn.openai.com%2Fpdf%2F3c7f7e1b-36c4-446b-916c-11183e4266b7%2Fchatgpt-usage-and-adoption-patterns-at-work.pdf%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/QkU4OwYXWGZQYV9C7C35vo5UPCcpgvU3s14d5KkG-9E=424">
<span>
<strong>OpenAI releases workplace usage report showing rapid ChatGPT adoption (20 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
28% of US workers now use ChatGPT at work, with adoption highest among highly educated white collar workers in tech, finance, and manufacturing. It's vague exactly how the tool is being used - OpenAI is primarily for writing, research, programming, and data analysis - but technical teams have emerged as the clear power users.
</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;">π§βπ»</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>Engineering & Research</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%2Fsnyk.io%2Fevents%2Fdevseccon%2F%3Futm_campaign=dm_pp_tldr_250929_evt_251022_devseccon_flagship_2025%26utm_medium=em-pa%26utm_source=tldr%26utm_content=evt_251022_devseccon_flagship_2025/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/SU8KLf5ehTzo8aMom0JwaLUVkun7tKPDq1ovNfy1Lb8=424">
<span>
<strong>Can you build secure AI at scale? Take the challenge at DevSecCon 2025 (Sponsor)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Something big is brewing October 22: the FIRST ever <a class="underline" href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fsnyk.io%2Fevents%2Fdevseccon%2F%3Futm_campaign=dm_pp_tldr_250929_evt_251022_devseccon_flagship_2025%26utm_medium=em-pa%26utm_source=tldr%26utm_content=evt_251022_devseccon_flagship_2025/2/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/XtTZOtH-AJLq0rOTteEFpAPqtjpOWdf4e2_QVb7eQeY=424" rel="noopener noreferrer nofollow" target="_blank"><span>AI Security Developers Challenge</span></a> at DevSecCon 2025 (virtual). Team up with brilliant minds globally to solve real AI security problems. Also on the agenda: hands-on demos, cutting-edge talks from Tessl and Ragie.ai experts, and practical strategies for securing AI applications at scale. <a class="underline" href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fsnyk.io%2Fevents%2Fdevseccon%2F%3Futm_campaign=dm_pp_tldr_250929_evt_251022_devseccon_flagship_2025%26utm_medium=em-pa%26utm_source=tldr%26utm_content=evt_251022_devseccon_flagship_2025/3/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/MWWvvn-uRFaRXD_6CN-DfkVGibKehdNounXgNtrs2vU=424" rel="noopener noreferrer nofollow" target="_blank"><span>Save your spot</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%2Fthinkingmachines.ai%2Fblog%2Fmodular-manifolds%2F%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/tbbUZpvSuXPvtVRCReiM9Mqr2C9uNs11BW2zqBTMGgs=424">
<span>
<strong>Modular Manifolds - Thinking Machines Lab (23 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Normally, neural networks are trained by letting their weights move freely in any mathematical direction. A new method that restricts the weights to specific curved surfaces makes training more stable and predictable. This "manifold Muon" optimizer forces weight matrices to maintain consistent properties (like not stretching or shrinking data too dramatically) by constraining them to geometric shapes called manifolds. While promising, the experiments required computational overhead, which limits real-world adoption.
</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%2Fmodal.com%2Fblog%2Freverse-engineer-flash-attention-4%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/TnvFq1dFOveN-atPMoqC2s12GODMA3RJPRHzhWt1ooA=424">
<span>
<strong>We reverse-engineered Flash Attention 4 (15 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Flash Attention 4 is a brand new, highly optimized CUDA kernel that speeds up the attention calculations in transformers, which are the main mathematical operations that bottleneck AI models like ChatGPT. However, the improvements from FA4 are not from clever new math but from improvements in how an incredibly complex async pipeline splits computations across 32-thread βwarps'.
</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.gojiberries.io%2Fwhat-problemdspy-fixed-budgets-and-what-were-really-optimizing%2F%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/zUAbXQlX9ZVw4g-TGD15tc3ynyUHlh0rJ3BWzehEuFE=424">
<span>
<strong>What Problem is DSPy Solving With What Assumptions (9 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
DSPY's programming model is a good base, but it has issues with defaults and missing affordances. This post identifies these issues and suggests some changes. Adopting these changes would sharpen DSPy's design without fighting it. It would keep the ergonomics that makes teams productive while aligning the optimization layer with the realities of fixed budgets and production constraints.
</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;">π</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%2Finterestingengineering.com%2Fculture%2Falibaba-nvidia-ai-partnership-expansion%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/Cjjy-vwowCXN5Py_z2FbfBXR_5IleVZiK3hP3zvqzj0=424">
<span>
<strong>Alibaba bets big on AI with Nvidia tie-up, new data center plans (4 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Alibaba has partnered with Nvidia to enhance its cloud platform with advanced AI capabilities, focusing on robotics and autonomous driving. It plans to expand global data centers, opening facilities in Brazil, France, and the Netherlands to meet increasing AI infrastructure demand. The company also launched Qwen3-Max and Qwen3-Omni AI models, emphasizing improved performance in code generation and virtual reality applications.
</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%2Fscale.com%2Fblog%2Fphysical-ai%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/pR2ik6U6rj6XCwg74GTP_sFssZQ2956cv6ufpZNOMgQ=424">
<span>
<strong>Expanding Our Data Engine for Physical AI (4 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Scale is expanding its Data Engine for Physical AI, providing robotics companies with high-quality, annotated datasets needed for training foundation models in physical environments. It addresses the robotics data gap by ensuring datasets are abundant, diverse, and enriched, collecting data through dedicated robots and human demonstrations. This aims to accelerate the development of reliable AI systems capable of handling real-world complexity in various applications.
</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;">β‘</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%2Fwww.callrail.com%2Fvoice-assist-in-action%3Futm_campaign=tldr_ai_september_2025%26utm_medium=thirdparty_advertising%26utm_source=tldr_ai%26utm_content=september_29_2025_newsletter/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/dH5I5OORhcbIrS8m1gJm0Qv7FnkToVcr0oxg02K64FU=424">
<span>
<strong>AI that answers calls 24/7: Call & talk to it (Sponsor)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
CallRail's Voice Assist can field calls more fluently and efficiently than most humans. It answers calls 24/7, qualifies and captures leads, filters out spam, and answers FAQs. <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.callrail.com%2Fvoice-assist-in-action%3Futm_campaign=tldr_ai_september_2025%26utm_medium=thirdparty_advertising%26utm_source=tldr_ai%26utm_content=september_29_2025_newsletter/2/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/TQf8UdNJE_xj2-kbdsPG99vN3lARbzcNQ8bZFuZGPGY=424" rel="noopener noreferrer nofollow" target="_blank"><span>π Call Voice Assist and hear it live</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%2Ffidjisimo.substack.com%2Fp%2Fa-new-paradigm-of-proactive-steerable%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/_oQX_xbj_utkfv7k1TqZ2WPzZRjEUZyCWWlnEgcHKIE=424">
<span>
<strong>A new paradigm of proactive, steerable AI (3 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
OpenAI's new CEO for Applications says the long-heralded dream of AI optimists is near - the personalized AI assistant that proactively helps you achieve your goals.
</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%2Fblog.cloudflare.com%2Fan-ai-index-for-all-our-customers%2F%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/o3WgE1TXqEE2gYLP3pipIScIbEAUlbsC9lS_O8BN0A0=424">
<span>
<strong>An AI Index for all our customers (6 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
AI Index for domains on Cloudflare is a web index that gives content creators the tools to make their data discoverable by AI, giving AI builders access to better data for fair compensation.
</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.dwarkesh.com%2Fp%2Frichard-sutton%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/IsiIKmmqmUEakBeHf5j7EbftfdAIewQKdhx44h7iqIs=424">
<span>
<strong>Richard Sutton β Father of RL thinks LLMs are a dead end (54 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Large language models aren't capable of learning on-the-job, so no matter how much we scale, we will still need a new architecture to enable continual learning.
</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.cnbc.com%2F2025%2F09%2F26%2Fanthropic-global-ai-hiring-spree.html%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/3UPhPKS9W4h3jXvEHNRmOoyVO7Tib7qyqFC1dbJsdyo=424">
<span>
<strong>Anthropic to triple international workforce in global AI push (8 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Anthropic's international growth is outpacing the company's most ambitious forecasts.
</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.testingcatalog.com%2Fanthropic-experiments-with-an-agent-for-gereating-ui-on-the-fly%2F%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/KbRgQS_o2x9uyckWyqPGadpTrMJrKX5EXEWCffNxhfs=424">
<span>
<strong>Anthropic experiments with an agent for real-time UI generation (6 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
Imagine with Claude, which lets users interact with a classic desktop UI where windows and apps are managed by AI, will be released as a temporary demo for certain plans.
</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%2Fblog.youtube%2Fnews-and-events%2Fintroducing-youtube-labs%2F%3Futm_source=tldrai/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/rTx0oqQZDCWECpzmFSH8H6H3b0eWo_BIYPe644foC24=424">
<span>
<strong>YouTube Labs (1 minute read)</strong>
</span>
</a>
<br>
<br>
<span style="font-family: "Helvetica Neue", Helvetica, Arial, Verdana, sans-serif;">
YouTube Labs is a new initiative that gives users early access to AI experiments.
</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;">
Love TLDR? Tell your friends and get rewards!
</p>
</td></tr>
<tr><td class="container" style="padding: 0px 10px 15px;">
<div class="text-block">
Share your referral link below with friends to get free TLDR swag!
</div>
</td></tr>
<tr><td align="left" style="padding: 10px;">
<div class="text-block">
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Frefer.tldr.tech%2F51d793d6%2F2/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/ss9QrXFZ6pxuGMWMXYkN8Extgwv5tJSj4PBJMScmJTc=424" style="color: #464ba4; text-decoration: underline;">https://refer.tldr.tech/51d793d6/2</a>
</div>
</td></tr>
<tr></tr>
<tr><td align="left" style="padding:5px 10px;">
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fhub.sparklp.co%2Fsub_9a4c16be99eb%2F2/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/_9inKxoNWkosIjrRBKUKkldNPQ0uZVT6LdlXUVj6avA=424" style="font-size: 16px; line-height: 1.6; padding: 10px 0; display: inline-block; text-decoration: underline;"><span style="mso-text-raise:13pt; text-decoration: underline;">Track your referrals here.</span></a>
</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? π°
</p>
<div class="text-block" style="margin-top: 10px;">
If your company is interested in reaching an audience of AI professionals and decision makers, you may want to <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fadvertise.tldr.tech%2F%3Futm_source=tldrai%26utm_medium=newsletter%26utm_campaign=advertisecta/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/ECRAVl3w68P59SzcJgDKMgV-1sjO5AwI-yGZfFi0DZg=424"><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? πΌ
</p>
<div class="text-block" style="margin-top: 10px;">
<a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fjobs.ashbyhq.com%2Ftldr.tech/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/Mfuxj5KIWGLiyCLyg2NHHYkuQqgbt5Y7BhSPVGb03NA=424" 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%2Ftwitter.com%2Fandrewztan/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/pBl9IB0rhixzGh4o7vmUbNv_srETY8elW0KKiDL4R9s=424"><span>Andrew Tan</span></a>, <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.linkedin.com%2Fin%2Faliiaminian%2F/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/maSP1v4d9R-fT9R-VVTm6cvr520q5GIPxjz7lrbEpGI=424"><span>Ali Aminian</span></a>, & <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fwww.linkedin.com%2Fin%2Fjacob-turner-7521a8198%2F/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/F-AHpxiUJSrGhlhYIgRC481b6hRiLINVL7fBD8Z_obM=424"><span>Jacob Turner</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%2Fai%2Fmanage%3Femail=silk.theater.56%2540fwdnl.com/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/79Z3-LHJ9Hd2_B5W1H-YVVkihCXSC58czutF0slPcwI=424">Manage your subscriptions</a> to our other newsletters on tech, startups, and programming. Or if TLDR AI isn't for you, please <a href="https://tracking.tldrnewsletter.com/CL0/https:%2F%2Fa.tldrnewsletter.com%2Funsubscribe%3Fep=1%26l=eedf6b14-3de3-11ed-9a32-0241b9615763%26lc=1571a9f6-84b7-11f0-b464-47c5c04ad337%26p=e2677a24-9d2d-11f0-85b9-df7ce12a65ad%26pt=campaign%26pv=4%26spa=1759150913%26t=1759151752%26s=e30fcb3b7aa0c96c86e6eec18a5139ddc869181508c8a7da590c172b6f2145cc/1/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/_XNQ0cqZ5Wh5z1A_XEcKjwogc8PkHLc5YyJAhdapeCo=424">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/CI0/01000199959d87bf-81206901-6687-4c0c-ab84-2032aa5328d9-000000/MhSeSspGM1MTz8d4xXwI0ocGfyIcVqdVJGK4QmKvULA=424" style="display: none; width: 1px; height: 1px;">
</body></html>