AllyshipContent Radar

Deep dives 738

Paper và bài viết kỹ thuật

Crawl live-1795068900000 · 13:22 19 thg 11, 2026 · Xuất bản 13:30 19 thg 11, 2026

Tác tử

v0.22.1

<h2>What's Changed</h2> <ul> <li>feat(core): support image results in web search tools by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seratch/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seratch">@seratch</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5376206131" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4898" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4898/hovercard" href="https://github.com/openai/openai-agents-python/pull/4898">#4898</a></li> <li>feat(core): customize output guardrail blocked messages by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seratch/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seratch">@seratch</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5224272471" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4594" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4594/hovercard" href="https://github.com/openai/openai-agents-python/pull/4594">#4594</a></li> <li>feat(mcp): add server-wide guardrails to MCP tools by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seratch/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seratch">@seratch</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5240154474" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4632" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4632/hovercard" href="https://github.com/openai/openai-agents-python/pull/4632">#4632</a></li> <li>feat(sandbox): add configurable Unix-local environment isolation by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seratch/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seratch">@seratch</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5240560269" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4640" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4640/hovercard" href="https://github.com/openai/openai-agents-python/pull/4640">#4640</a></li> <li>feat(sandbox): allow labels on Docker sandbox containers by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seanxuu/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seanxuu">@seanxuu</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5214776113" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4564" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4564/hovercard" href="https://github.com/openai/openai-agents-python/pull/4564">#4564</a></li> <li>feat(voice): expose streamed transcription options by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/seratch/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/seratch">@seratch</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5241618727" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4645" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4645/hovercard" href="https://github.com/openai/openai-agents-python/pull/4645">#4645</a></li> <li>fix(core): detach aggregated request usage entries by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/green3sf/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/green3sf">@green3sf</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5191892708" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4519" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4519/hovercard" href="https://github.com/openai/openai-agents-python/pull/4519">#4519</a></li> <li>fix(core): fail closed on empty tool arguments by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/hsusul/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/hsusul">@hsusul</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5206323744" data-permission-text="Title is private" data-url="https://github.com/openai/openai-agents-python/issues/4545" data-hovercard-type="pull_request" data-hovercard-url="/openai/openai-agents-python/pull/4545/hovercard" href="https://github.com/openai/openai-agents-python/pull/4545">#4545</a></li> <li>fix(core): strip created_by when replaying RunItems as input by <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/hsusul/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/hsusul">@hsusul</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5217553687" data-permission-text="Title is private" data-url="https://github.com/openai/op

Nguyên văn nguồn
Chi tiết nguồn
  • v0.22.1 openai/openai-agents-python releases · 8 thg 9, 2026
Đánh giá

Top Alignment Researcher on OpenAI/HuggingFace Revelations, Fixing AI Safety & Takeover Odds

Câu hỏi 0:00 Intro 1:02 Buck's initial reaction upon first reading the report 2:37 How fast the AIs actually solved the "hack" 3:59 Why the AIs cheated in the first place 10:28 How this might have played out differently with human scorers 19:00 The most unexpected behaviors in the report 25:06 Buck's actual odds on a full AI takeover 27:33 Buck's proposed path forward for better alignment 36:19 Which criticisms of the report Buck agrees with, and which he doesn't 48:11 Can AI models even be trusted to evaluate each other?

Jacob sits down with Buck Shlegeris, CEO of Redwood Research, one of the organizations that led the independent investigation into OpenAI/Hugging Face's incident. They dig into the incident itself, Buck's reactions to it, and what he believes it reveals about the state of where we are today. 0:00 Intro 1:02 Buck's initial reaction upon first reading the report 2:37 How fast the AIs actually solved the "hack" 3:59 Why the AIs cheated in the first place 10:28 How this might have played out differently with human scorers 19:00 The most unexpected behaviors in the report 25:06 Buck's actual odds on a full AI takeover 27:33 Buck's proposed path forward for better alignment 36:19 Which criticisms of the report Buck agrees with, and which he doesn't 48:11 Can AI models even be trusted to evaluate each other? Jacob is an AI investor at Redpoint Ventures. He's led Redpoint's investments in companies like Abridge, Physical Intelligence & Legora. Follow Jacob on Twitter: https://x.com/jacobeffron On Unsupervised Learning we probe the sharpest minds in AI in search for the truth about what's real today, what will be real in the future and what it all means for businesses and the world. If you're a builder, researcher or investor navigating the AI world, this podcast will help you deconstruct and understand the most important breakthroughs and see a clearer picture of reality. Subscribe to this show to stay up to date on our latest episodes. Full Metr/Redwood report (by Ryan Greenblatt, Ajeya Cotra, and Hjalmar Wijk): https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/

Unsupervised Learning · Unsupervised Learning: With Jacob Effron Mở bản gốc ↗
Nguyên văn nguồn
Chi tiết nguồn
Kỹ thuật

How a Professional Writer Writes With AI

Two years ago, Katie Parrott was laid off from a crypto firm and couldn’t afford a career coach—so she turned to a $20-a-month ChatGPT subscription instead. The Every staff writer’s habit of feeding AI good context eventually became compound writing, a codified system for brainstorming, drafting, and editing with AI. Today, that system is a plugin any writer can use. On this week’s AI & I, Natalia Quintero talks with Katie about turning good ingredients into good writing, borrowing the taste of writers she admires, and why AI helped her fall back in love with the page. If you found this episode interesting, please like, subscribe, comment, and share! To hear more from Natalia Quintero: Subscribe to Every: https://every.to/subscribe Follow her on X: @NataliaZarina Go to https://attio.com/every and get 15% off your first year. Timestamps: 0:00 Start 0:49 Introduction 1:21 How Katie went from being laid off to using ChatGPT as a career coach 8:06 Turning AI into a “content agency of one” 10:59 Building the context files and style guides that make AI outputs useful 14:15 What makes a Claude project actually work 21:18 How AI helped with a mental health crisis and “computer errands” 27:03 Katie’s career coach, now a fully autonomous Codex project 32:27 What compounding means, and building the Compound Writing plugin 38:00 Borrowing Vonnegut, Hitchcock, and Sorkin as AI editing skills 44:49 Katie’s thesis: education and access will matter more than ever Links to resources mentioned in the episode: Katie Parrott on X: @kplikethebird Compound writing plugin: https://github.com/EveryInc/compound-writing Compound engineering plugin (Kieran Klassen): https://github.com/EveryInc/compound-engineering-plugin Parrott’s companion piece on compound writing: https://every.to/guides/compound-writing Parrott, “I Hired ChatGPT as My Career Coach”: https://every.to/working-overtime/i-hired-chatgpt-as-my-career-coach Parrott, “AI Turned Me Into a Content Agency of One”: https://every.to/working-overtime/ai-turned-me-into-a-content-agency-of-one

Nguyên văn nguồn
Chi tiết nguồn
Tác tử

Self-hosted machines

Cursor supports **self-hosted machines**, which let you keep tool execution entirely in your own network. Your codebase, build outputs, and secrets all stay on internal machines running in your infrastructure, while the agent handles tool calls locally. <figure><img src="https://ptht05hbb1ssoooe.public.blob.vercel-storage.com/assets/uploads/remote-machines-QjJ0ZzGDelSdT7v2CXDdh0BQHqjmH0.png" loading="lazy" alt="Run on menu showing Remote Machines with self-hosted options lambda-test, cloud-demo, and jacks-desktop" /><figcaption>Run on menu showing Remote Machines with self-hosted options lambda-test, cloud-demo, and jacks-desktop</figcaption></figure> Dynamic pool scheduling **My Machines** connects a single laptop or VM to your account for personal workflows. **Team pools** are named queues of workers for a team or enterprise. Capacity can grow as requests arrive and shrink when workers disconnect, so your self-hosted machines can scale with demand. Pools are not tied to one repository: name the pool, and any available worker can claim the request. Pools can also hibernate idle machines, then restore within a reconnect window when a follow-up arrives, so you don't keep expensive capacity warm just for the next prompt. <figure><img src="https://ptht05hbb1ssoooe.public.blob.vercel-storage.com/assets/uploads/pools-mNNea8rJyNxaBuz7rSVT2SJOKkRsmv.png" loading="lazy" alt="Cloud Agents dashboard listing Self-hosted Machines pools with active and idle worker counts" /><figcaption>Cloud Agents dashboard listing Self-hosted Machines pools with active and idle worker counts</figcaption></figure> Run on your sandboxes Cloud agents can now execute on infrastructure you already use , including from AWS Lambda, Coder, Cloudflare, Daytona, Modal, Namespace, Vercel, and E2B. Computer use on Linux and Mac Self-hosted workers now support computer use on Linux and Mac. With the right desktop packages, an agent can click, type, take screenshots, and drive the browser. You can watch its desktop or take control from Cursor.

Nguyên văn nguồn
Chi tiết nguồn
Tác tử

Making Cities Awesome: Peregrine’s Nick Noone & Ben Rudolph

Most public safety technology companies grow by collecting more data. Peregrine inverted the model: no sensors, no new data, a business built on connecting the data and information cities already own. Co-founders Nick Noone and Ben Rudolph received more than two dozen no's before San Pablo PD let them in the door in February 2018. Today, Peregrine powers law enforcement, emergency medical services, fire and rescue, and other services in more than 400 cities and communities globally. Nick and Ben explain their north star for data sovereignty, and discuss how Peregrine's philosophy and privacy-first approach to data access and ownership preserves individual privacy and cities' sovereignty. They walk through how AI and long-horizon agents are being deployed: a cold case agent that reproduced an exoneration detectives had reached by hand, a Wisconsin county that placed a suspect using cell records buried in 300GB of evidence, identifying threats to a synagogue, root-causing an escalation in weather-related incidents, and more. Hosted by Sonya Huang, Sequoia Capital 00:00 Introduction 02:07 What Forward Deployed Engineering Means 03:58 What Silicon Valley Gets Wrong 05:23 UNHCR, Dimagi And Downstream Data Problems 08:25 Why Cities, Why Safety 10:45 Two Dozen Nos And San Pablo PD 14:19 Building Through Defund The Police 18:16 The Inversion Of The Collection Model 21:20 Data Ownership And Governance 22:57 From Nice Search To Deep Analysis 29:59 Agents Writing The Integrations 31:45 The Cold Case Agent 35:02 The Anti-Network-Effect Proposition 38:40 Facial Recognition And Hard Decisions 40:48 Technology For The Underdogs 42:54 Trusting The Individual Contributor 48:50 Ten Thousand Cities

Nguyên văn nguồn
Chi tiết nguồn
Kỹ thuật

Developing Enterprise Frontier Safeguards with our customers

We’re announcing Enterprise Frontier Safeguards (EFS), a solution that combines the privacy of zero data retention with state-of-the-art safeguards for detecting misuse. EFS works by storing data in cloud infrastructure controlled by the customer, not Anthropic.

Nguyên văn nguồn
Chi tiết nguồn
Tác tử

Start from scratch, without a repo

Cloud Agents no longer require a connected GitHub or other third-party SCM provider to get started. Prompt from the get-go, then save your work to a Cursor Origin repo. Get started Select **Start from scratch** in the repo picker, then prompt the agent. In the background, Cursor creates an Origin repo for you. <figure><img src="https://ptht05hbb1ssoooe.public.blob.vercel-storage.com/assets/changelog/start-from-scratch-QflXlxizUwAvBbal0Kst46I0mEZT93.png" loading="lazy" alt="Start from scratch in the repo picker" /><figcaption>Start from scratch in the repo picker</figcaption></figure> Turn it into a real repo, whenever you want Once the agent's build is to your liking, click the **Create repo** button to save your work in an Origin repo. Choose a custom name or pick one of the suggested ones, then set the visibility to private or internal. You'll get a fully scaffolded Origin repo, ready to share or keep building on. Navigate to the Codebase tab to find your repo and access your project. A live preview, right in the browser Cursor now port-forwards your cloud agent's live environment straight to your browser, so you can preview it and use tools like design mode. Publish your work Connect a Vercel account and hit publish to get a live URL for what you built. A Vercel account is required to use the publish feature. Get started today .

Nguyên văn nguồn
Chi tiết nguồn
Kỹ thuật

v0.29.0

<h2>What's Changed</h2> <h3>Notable Changes</h3> <ul> <li>Removed the deprecated <code>start_method</code> setting that already had no effect.</li> <li>The <code>.length</code> property on paginated objects (e.g. <code>Api.runs()</code>, <code>Api.sweeps()</code>, <code>run.files()</code>) has been removed. Use <code>len(...)</code> instead.<br> <code>wandb.Video()</code> now requires the <code>format</code> argument when initializing from a numpy array or raw bytes.</li> <li>Removed the deprecated <code>summary="best"</code> option and <code>goal</code> argument of <code>run.define_metric()</code>. Use <code>summary="min"</code> or <code>summary="max"</code> instead.</li> <li>Removed the deprecated <code>quiet</code> argument of <code>run.finish()</code> and <code>wandb.finish()</code>. Use <code>wandb.Settings(quiet=...)</code> instead.</li> <li>Removed the deprecated <code>run.project_name()</code>, <code>run.get_url()</code>, <code>run.get_project_url()</code>, and <code>run.get_sweep_url()</code> methods. Use the <code>run.project</code>, <code>run.url</code>, <code>run.project_url</code>, and <code>run.sweep_url</code> properties instead.</li> <li><code>wandb.Image</code> no longer normalizes pixel values. Callers must now ensure their data already falls within the range [0, 255].</li> <li>On macOS, <code>wandb</code> now requires macOS 13 (Ventura) or newer.</li> </ul> <h3>Added</h3> <ul> <li>Added <code>run.console_logs()</code> to the public API for reading the console output that W&amp;B captured for a run — the whole log, or only the last N lines with <code>run.console_logs(last=N)</code>, for finished and still-running runs alike. Reading from the beginning requires W&amp;B Server 0.77 or newer (<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/dmitryduev/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/dmitryduev">@dmitryduev</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5123516179" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/12442" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/12442/hovercard" href="https://github.com/wandb/wandb/pull/12442">#12442</a>)</li> <li>The automations API now supports team and organization scopes. (<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/tonyyli-wandb/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/tonyyli-wandb">@tonyyli-wandb</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="4859789273" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/12197" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/12197/hovercard" href="https://github.com/wandb/wandb/pull/12197">#12197</a>, <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="4856758381" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/12194" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/12194/hovercard" href="https://github.com/wandb/wandb/pull/12194">#12194</a>)</li> <li>The automations API now supports creating and editing automations whose scope is a <code>Registry</code> object (<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/tonyyli-wandb/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/tonyyli-wandb">@tonyyli-wandb</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="3614364304" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/10867" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/10867/hovercard" href="https://github.com/wandb/wandb/pull/10867">#10867</a>)</li> <li>Press <code>g</code> in LEET to draw guides behind line charts: a dotted background or horizontal lines aligned with the axis ticks. The choice is saved and can also be set with <code>chart_guides</code> in <code>wandb leet config</code>. (<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/dmitryduev/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/dmitryduev">@dmitryduev</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5146917923" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/12463" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/12463/hovercard" href="https://github.com/wandb/wandb/pull/12463">#12463</a>)</li> <li>Drag the separator lines between LEET's run overview sections to resize them with the mouse. The proportions are saved per view; press <code>0</code> to reset them along with the other pane sizes (<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/users/dmitryduev/hovercard" data-octo-click="hovercard-link-click" data-octo-dimensions="link_type:self" href="https://github.com/dmitryduev">@dmitryduev</a> in <a class="issue-link js-issue-link" data-error-text="Failed to load title" data-id="5217693664" data-permission-text="Title is private" data-url="https://github.com/wandb/wandb/issues/12525" data-hovercard-type="pull_request" data-hovercard-url="/wandb/wandb/pull/12525/hovercard" href="https://github.com/wandb/wandb/pull/12525">#12525</a>)</li> <li>Registry API version queries (<code>Api.registries().versions()</code>, <code>Api.registries().collections().versions()</code>, <code>Registry.versions()</code>, <code>Registry.collections().versions()</code>) now accept an optional <code>order</code> string as a keyword argument for organizations with advan

Nguyên văn nguồn
Chi tiết nguồn
  • v0.29.0 Weights & Biases — SDK releases · 27 thg 8, 2026
Kỹ thuật

The Entertainment Industry’s Biggest Names Back Stability AI in Latest Funding Round

Total funding reaches $232M under new leadership with investor group comprised of Electronic Arts, Sony Music Group, Universal Music Group, Warner Music Group, and more LOS ANGELES - AUGUST 25, 2026 - Stability AI , the leader in purpose-built AI products for professional creatives across music, gaming, and entertainment, today announced a Series B fundraise of $76M in new capital. The news brings total funding to $232M, inclusive of two equity rounds and convertible notes, under CEO Prem Akkaraju since his appointment in June 2024. The capital will fuel the development of its product suite for creative production, deepen its applied research discipline, and expand its professional services arm. The round welcomes a new group of investors across entertainment and technology, including interactive entertainment innovator Electronic Arts, global leaders in music entertainment Sony Music Group, Universal Music Group and Warner Music Group, and investment firms AMD Ventures and Pacific Alliance Ventures. They join an already stacked roster of the who's who across the entertainment and technology industries backing Stability AI: Coatue, Greycroft, Kadmos Capital, Lightspeed Venture Partners, Mantis Capital, Sound Ventures, and WPP, alongside individual investors Sean Parker, Eric Schmidt, James Cameron, Kevin Mayer, Mark Burnett, Patrick Whitesell, Prem Akkaraju, Reuben Brothers, Robert Nelsen, and Vivi Nevo. Coatue, Greycroft, Kadmos Capital, Sean Parker, and Eric Schmidt are also participating in the Series B, investing for a second consecutive round under new leadership. “This unmatched group of investors is an affirmation of our vision where generative AI empowers every producer, musician, and storyteller,” said Prem Akkaraju, CEO of Stability AI. “Stability is unique in the AI field because we are creative people making tools for creatives. Developing industry-defining technology means assembling a team of industry-defining titans, and that’s what we have done here.” Akkaraju added, “Our investors set us up for long-term success because they provide not only capital, but expertise, credibility, and direct connection to artists.” This round includes investments from existing strategic partners Electronic Arts , Universal Music Group , and Warner Music Group . WPP , a global leader in marketing services, has been a strategic partner and investor throughout Stability AI's growth under new leadership. The announcement comes on the heels of the launch of Stable Audio 3.0 , a family of open-weight music models trained on fully licensed data. Artists can now bring Stable Audio 3.0 into their workflow through a new DAW (Digital Audio Workstation) plugin or directly on StableAudio.com . Additionally, Stability AI has announced that Coatue Co-Founder Thomas Laffont has joined its Board of Directors. “While others are building generalized AI, Stability AI is building creative tools and doing it alongside the artists, studios, and rights holders whose work defines the field,” said Thomas Laffont, Co-Founder of Coatue. “I'm glad to be joining Prem and the team.” Laffont rounds out a board that includes Academy Award-winning filmmaker James Cameron, entrepreneur, philanthropist, and former President of Facebook Sean Parker, Greycroft Co-Founder and Managing Partner Dana Settle, and Stability AI CEO and former CEO of Weta Digital Prem Akkaraju. With this new funding, Stability AI will continue to build creative tools designed around the artist's process, across music, gaming, and entertainment.

Nguyên văn nguồn
Chi tiết nguồn
Tác tử

Parallel’s Parag Agrawal: Building a New Web for AI Agents

Parag Agrawal is making a bet that goes against two decades of web search: agents will query the web a thousand times more than humans ever have, and the infrastructure built around human clicks is wrong for them. The former Twitter CEO, now founder and CEO of Parallel Web Systems, explains why Parallel treats human click data as a bug and trains on agent feedback instead. He unpacks the counterintuitive choice to ship a search agent before a search engine, building an index incrementally, and how the new Turbo product cut agentic search to 200 milliseconds. But the problem Parag keeps returning to is economic: the ad-supported internet collapses when agents show up instead of people. His fix draws on Shapley values to pay content owners for the value their pages provide agents, with real dollars reaching publishers, he predicts, within 12 to 24 months. Hosted by Sonya Huang and Andrew Reed, Sequoia Capital 00:00 Introduction 03:25 What Is Web Search 05:17 Why Start a New Index 07:52 Search Agents First 10:17 Not a Neolab 13:14 Agents vs Google Search 19:38 Inside the Search Stack 28:59 Search Multipliers With Agents 30:21 Meeting Prep Agent Workflows 31:46 Quality Cost Latency And Turbo 32:42 Are Agents Overtaking Humans 34:28 Ads Model Meets Agent Web 37:20 New Incentives For Content 40:48 Shapley Values Attribution 47:46 Parallel Web And Future Vision

Nguyên văn nguồn
Chi tiết nguồn
721–738 / 738
Trang 37 / 37