AllyshipContent Radar

Hands-on 278

Công cụ, thử nghiệm và hướng dẫn thực hành

Crawl live-1795090500000 · 19:15 19 thg 11, 2026 · Xuất bản 19:35 19 thg 11, 2026

19

How Social Department built culture-aware video discovery

58.6

Hướng dẫn Mariah Miller

<p><em>This post is co-written by Mariah Miller and Sirajus Salekin from AWS and Nicholas Chase from Social Department.</em></p> <p>Entertainment marketing teams manage tens of thousands of video assets across studios, networks, and streaming applications. Most campaigns begin with the same video discovery problem: finding the one moment in the footage that will resonate with an audience. When a brief lands, the clock starts, and what they need is almost never something a database can describe.</p> <p>Semantic search, on its own, has a ceiling: It answers the question you can articulate. Ask it for a character holding a coffee mug and it will find every coffee mug in a thousand episodes. The queries that actually drive entertainment campaigns look nothing like that. They are questions of judgment: which moment in a show will travel, which line is quotable, and which single frame will work as a meme.</p> <p>Those questions can’t be typed into a search box. Answering them requires two things the search index doesn’t have: an editorial definition of what “quotable” or “memeable” means for a given library, and a continuously refreshed view of what is moving outside the library. Historically, answering them meant a person with deep knowledge of the show watching footage for a week, and that person doesn’t scale across a catalog.</p> <p>In this post, we describe how Social Department, the AI operating system for entertainment marketing, set out to encode those editorial questions as executable retrieval pipelines. The result is QuickSearch, a set of single-action queries that return the clips a marketer is looking for without requiring them to describe it: Shareable Scenes, Quotable Quotes, Memeable Moments, and Trend Align.</p> <p>The models in this solution run on Amazon Bedrock, so customer data stays within the AWS data boundary. It isn’t shared with model providers and isn’t used to train the underlying models.</p> <p>Social Department’s approach to responsible AI is that AI proposes and people decide. In production, code checks generated output against the source material (for example, a quote is discarded if it isn’t in the episode transcript). Amazon Bedrock Guardrails helps screen customer search text for prompt attacks, and a person on the customer’s team approves every post before it publishes.</p> <p>Social Department owns the pipeline architecture, the retrieval system, and the editorial criteria behind each action. TrackIt, an AWS Premier Tier Services Partner, implemented and delivered against that design. TwelveLabs’ Marengo 3.0 and Pegasus 1.2 models on <a href="https://aws.amazon.com/bedrock/" target="_blank" rel="noopener">Amazon Bedrock</a> provide the scene-level video reasoning the pipeline builds on. AWS supplies the infrastructure, the model access including Anthropic’s Claude models, and <a href="https://aws.amazon.com/opensearch-service/" target="_blank" rel="noopener">Amazon OpenSearch Service</a>, using its serverless option, Amazon OpenSearch Serverless, for retrieval.</p> <p><strong>A note on scope:</strong>&nbsp;This post describes the AWS portions of the system: the ingest and indexing pipeline, the retrieval path, and the services that run them. The components Social Department owns, including its media and data backbone, content analysis and discovery engine, editorial criteria, and evaluation program, are described only to the depth needed to follow the architecture. The named services that follow are the ones this post is about, not an inventory of Social Department’s full system.</p> <h2 id="about-social-department">About Social Department</h2> <p>Social Department is the AI operating system for entertainment marketing, serving major studios, networks, streamers, and FAST applications. It automates campaign strategy, asset creation, publishing, and performance optimization across a client’s full content library. Most of those functions depend on locating the right moment inside the footage first, and semantic clip discovery has been core to the product from the beginning.</p> <h2 id="why-semantic-search-alone-was-insufficient">Why semantic search alone was insufficient</h2> <p>Entertainment marketing operates under two constraints that a general-purpose search layer can’t satisfy.</p> <h3 id="the-valuable-queries-are-the-ones-nobody-can-phrase">The valuable queries are the ones nobody can phrase</h3> <p>A marketer can describe an object, a location, or a line of dialogue, but not what will travel. That judgment, this moment lands and that one doesn’t, is the actual skill in the job. It was never automatable because it depends on knowing how an audience behaves rather than what a video contains. A search layer that only understands content requires a human to supply the intent, and supplying that intent well requires having already watched the show.</p> <h3 id="trend-signal-lives-outside-the-library-and-moves-on-a-different-clock">Trend signal lives outside the library and moves on a different clock</h3> <p>No amount of video understanding tells you what audiences are talking about this morning. That information is created outside the library and changes hourly, while a library changes only when new content is added. A system that answers these queries has to acquire trend signal continuously and reconcile it against the library at request time.</p> <p>That distinction drove the core architectural decision: content understanding and trend acquisition run as independent pipelines on their own schedules and are joined only when a user selects a QuickSearch action.</p> <h2 id="solution-overview">Solution overview</h2> <p>With QuickSearch, teams can interact with a video library in a different way, one where the system proposes rather than waits. It provides four actions, each with its own retrieval strategy, editorial criteria, and ranking model. The walkthrough that follows describes how each is implemented. The list here introduces what they do.</p> <ol type=

AWS Machine Learning Blog · 19 thg 11, 2026

Nguyên văn nguồn
Chi tiết nguồn
20

Do LLM agents still need to be told how to self-improve?

58.6

Hướng dẫn omarsar0

Do LLM agents still need to be told how to self-improve? A new paper from MIT CSAIL and the University of Tübingen tests this directly. They give a frozen model a compute budget and simple tools. It can write practice games, watch another copy of itself play them, and edit that copy's prompt and skill files. It does not see real tasks or validation scores, and it is not told how to learn. Capable models built a working learning procedure on their own and improved on held-out games with hidden mechanics. Models from different families arrived at similar strategies. They wrote games at a calibrated difficulty, ran controlled A/B comparisons, built their own validation suites, and spent more effort on harder tasks. The text files they produced also improved play on external benchmarks with different horizons, mechanics, and 3D navigation. This matters for safety as well. Leaving out an explicit self-improvement loop does not stop a capable model from improving itself when it has the tools. https://arxiv.org/abs/2611.23219 Chat with Paper: https://academy.dair.ai/papers/do-llm-agents-still-need-to-be-told-how-to-self-improve-2611.23219

X · 19 thg 11, 2026

Nguyên văn nguồn
Chi tiết nguồn
21

From reporting to reasoning: Getting enterprise data ready for AI agents

58.6

Hướng dẫn Jane Ridge

<p><a href="https://www.linkedin.com/groups/25720044/" target="_blank" rel="noopener"><em>AWS Quick User Group DC</em></a> <em>inaugural meetup recap</em></p> <p>Many organizations are asking how to move beyond static reporting dashboards and begin using generative and agentic AI. But before an AI agent can answer questions or help teams act, it requires a trusted data foundation.</p> <div style="width: 810px" class="wp-caption alignnone"> <a href="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-1-2.jpeg" target="_blank" rel="noopener"><img src="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-1-2.jpeg" alt="Attendees at the AWS Quick User Group DC inaugural meetup" width="800"></a> <p class="wp-caption-text">Figure 1: AWS Quick User Group DC inaugural meetup, October 2026</p> </div> <p>On Wednesday, October 7, 2026, data leaders, developers, and practitioners gathered at the <a href="https://aws.amazon.com/training/skills-centers/arlington-skills-center/" target="_blank" rel="noopener">AWS Skills Center</a> in Arlington, Virginia, for the AWS Quick User Group DC meetup. The session brought together founders, software engineers, and enterprise leaders, among them a team from Financial Independence Group (FIG). Attendees explored practical strategies for modernizing legacy architectures, preparing structured data for agentic workflows, and automating routine tasks with no-code tools.</p> <blockquote> <p><em>User groups like this provide a forum for peer-to-peer learning and collaborative problem-solving. The event offered attendees an inside look at how organizations are putting <a href="https://aws.amazon.com/quicksuite/" target="_blank" rel="noopener">AWS Quick</a> to work to drive productivity gains.</em></p> </blockquote> <h2 id="whats-new-with-aws-quick-agents-skills-and-connectors">What’s new with AWS Quick: Agents, skills, and connectors</h2> <p>For many in the room, this meetup was a first look at what agentic AI looks like in practice with AWS Quick.</p> <div style="width: 810px" class="wp-caption alignnone"> <a href="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-2-2.jpeg" target="_blank" rel="noopener"><img src="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-2-2.jpeg" alt="Arun Ganesan presenting AWS Quick updates to the group" width="800"></a> <p class="wp-caption-text">Figure 2: Arun Ganesan presenting what’s new with AWS Quick</p> </div> <p>The meetup kicked off by exploring the latest product capabilities across AWS Quick, focusing on its evolution into a unified AI assistant for work. Rather than functioning solely as a traditional business intelligence visualization tool, AWS Quick unifies analytics, natural-language interaction, and multi-step task orchestration.</p> <ul> <li><strong>AWS Quick agents and skills</strong>: <a href="https://docs.aws.amazon.com/quicksuite/latest/userguide/working-with-agents.html" target="_blank" rel="noopener">Turn natural-language questions</a> into multi-step reusable actions that aid in decision making across enterprise data.</li> <li><strong>Third-party connectors</strong>: AWS Quick includes built-in integrations that connect tools like Salesforce, Slack, Google Drive, and Microsoft 365, breaking down data silos.</li> <li><strong>Unified workspace</strong>: A cohesive environment that bridges research, automated visual generation, and business artifact creation (presentations, documents, and spreadsheets) directly where users work.</li> </ul> <h2 id="real-world-innovation-moving-to-ai-analytics-and-agentic-systems">Real-world innovation: Moving to AI analytics and agentic systems</h2> <p>The core of the day highlighted a real-world customer journey showing how one enterprise is rethinking its data architecture so AI agents can work from governed, trusted data.</p> <h2 id="financial-independence-group-innovating-at-the-pace-of-technology">Financial Independence Group: Innovating at the pace of technology</h2> <p>Vanessa Gonzalez, Vice President of AI at Financial Independence Group (FIG), shared how FIG keeps pushing past what’s possible today. They’re adopting new tools and platforms as technology changes, and building toward AI-powered analytics and agentic systems on top of their data with AWS Quick.</p> <div style="width: 810px" class="wp-caption alignnone"> <a href="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-3-2.png" target="_blank" rel="noopener"><img src="https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/11/16/ML-22393-3-2.png" alt="Vanessa Gonzalez presenting FIG’s AI analytics journey" width="800"></a> <p class="wp-caption-text">Figure 3: Vanessa Gonzalez, Vice President of AI at FIG, presenting</p> </div> <p>Technology is changing fast, and FIG wants to be at the front of that change. Building on its existing analytics, FIG moved to AWS Quick, a platform that brings AI into its dashboards, reporting, and analytics. AWS Quick is FIG’s “single front door” for enterprise analytics. It brings together structured data such as customer relationship management (CRM) systems, databases, and finance data. It also pulls in unstructured content such as internal knowledge bases, policy documents, and collateral, through a governed data layer. That gives FIG a base for AI-powered insights and agentic workflows built on its own data.</p> <ul> <li><strong>The AI analytics loop</strong>: Moving from reporting on what already happened to a continuous cycle: Ask → Analyze → Automate → Act.</li> <li><strong>Centralized discovery</strong>: An analytics portal organized by business area (sales, operations, marketing, commissions), with built-in AI chat assistants so teams can talk directly to their documents and datasets.</li> <li><strong>Democratized insigh

AWS Machine Learning Blog · 19 thg 11, 2026

Nguyên văn nguồn
Chi tiết nguồn
22

13 years ago, Ivan Zhao (@ivanhzhao) started Notion so anyone could make software without knowing how to code.

58.5

Hướng dẫn danshipper

13 years ago, Ivan Zhao (@ivanhzhao) started Notion so anyone could make software without knowing how to code. Now AI agents do that out of the box. Rebuilding Notion for that world is one of the hardest things he’s ever done. Ivan and I talked at Thesis: 2027, @Every’s conference on what great human work looks like after automation. I came away thinking Notion is a dark-horse candidate as one of the most important companies at the app layer in AI. Here’s what we talked about: • How Notion went from pre-AI to AI-native. In late 2022, Ivan and his cofounder locked themselves in a hotel room and shipped their first AI product weeks before ChatGPT—even while people inside the company thought AI was a fad like crypto. After two years grinding on agents, they realized everyone is building agents now, so Notion’s unique edge had to come back to its core. • Why context outlasts UI and workflows. For the past 15 years, software meant a company picking one workflow and building buttons around it—something an agent can now vibe code on demand. Ivan says Notion isn’t married to the UI anymore. What lasts is a team’s data, semantics, and context, and Notion wants to be the keeper of that context for both humans and the next billion knowledge workers, most of whom he expects will be agents. • Legos versus clay for vibe coding. Coding agents hand you clay—granular code that falls apart when you try to build a tall tower without expertise. Notion is leaning into the Legos: opening its UI pattern libraries to agents while taking care of permissions, syncing, and collaboration under the hood, alongside Ivan’s framework for what you should vibe code yourself versus leave to crafted software. • The TikTokification of AI—and keeping human understanding. Ivan admits he catches himself chatting with Claude instead of reading on his Kindle, getting grossed out after 10 minutes like he’s on TikTok. His deeper warning for builders is not to surrender understanding when agents write code inside black boxes—which is why Notion’s own teams went back to Kanban boards to visualize and orchestrate their agents’ work. This is a must-listen for anyone transitioning their company to the AI age. Listen to the full conversation on The Every Podcast! 0:00 Introduction 1:18 Why vibe coding revives the 1960s vision of software 4:40 When people should build their own tools—and when they shouldn't 7:47 Why data and context outlast UI and procedural logic 9:46 The history game Ivan is vibe coding with his wife 12:23 Bringing a 13-year-old company from pre-AI to post-AI 16:00 The TikTokification of AI and its second-order effects 19:46 Don't surrender your understanding to agents 22:01 Why builders need Lego-like blocks to vibe code 28:04 Why the next billion knowledge workers will be agents Media

X · 19 thg 11, 2026

Nguyên văn nguồn
Chi tiết nguồn
23

It's been nearly 6 months since I posted a 'watch me work' deep dive on my skills.

58.4

Hướng dẫn mattpocockuk

It's been nearly 6 months since I posted a 'watch me work' deep dive on my skills. Since then, they've become the most-starred skills repo in the world. So, here's 70 minutes of me building a 21,000-line PR, especially focusing on: - DDD with /grill-with-docs - Improving your codebase with /retro - How to do human review effectively Enjoy! Media

X · 18 thg 11, 2026

Nguyên văn nguồn
Chi tiết nguồn
19–24 / 179
Trang 4 / 30