Mọi bản cập nhật chức năng và bản phát hành mẫu chính — ai đã vận chuyển cái gì và nó thay đổi ranh giới hiệu suất giá như thế nào.
Câu chuyện 301
Các mô hình mới thiết lập lại giới hạn về năng lực và hiệu suất giá cả. Các nhóm đánh giá lại những gì cần xây dựng bất cứ khi nào đợt ra mắt thay đổi số tiền có thể có trên mỗi đô la.
<p><strong><a href="https://tools.simonwillison.net/openai-webrtc">Phiên âm thanh OpenAI WebRTC, hiện có ngữ cảnh tài liệu</a></strong></p> Tôi đã xây dựng phiên bản đầu tiên của công cụ này <a href="https://simonwillison.net/2024/D…
“Tôi không chắc liệu công ty này có ủng hộ văn hóa hackathon nữa hay không,” một nhân viên đăng trên một diễn đàn dành cho toàn bộ nhân viên.
Một báo cáo mới cho thấy đơn vị sử dụng 6.500 người này đang trên bờ vực nổi dậy.
Theo các nguồn tin và cuộc thảo luận nội bộ được WIRED xem xét, các giám đốc điều hành cũng như nhân viên đều đang phải vật lộn với chiến lược AI hỗn loạn của Meta.
Chúng tôi xem xét Gemini-SQL2, khả năng chuyển văn bản sang SQL mà Google Research đã công bố vào ngày 12 tháng 6 năm 2026. Được hỗ trợ bởi Gemini 3.1 Pro, nó đạt độ chính xác thực thi 80,04% trên bảng xếp hạng mô hình đơn BIRD. Chúng tôi giải thích những gì sco…
AgentPerf từ Phân tích nhân tạo, tiêu chuẩn AI tác nhân đầu tiên trong ngành, mang đến cho các nhà phát triển, doanh nghiệp và nhà cung cấp cơ sở hạ tầng một cách rõ ràng để so sánh các hệ thống dành cho AI tác nhân. Trong đợt xuất bản đầu tiên…
Trong bài đăng này, chúng ta khám phá cách Rocket Close xây dựng giải pháp bằng cách sử dụng Strands Agents, mô hình ngôn ngữ lớn (LLM), Amazon Bedrock, Cơ sở kiến thức Amazon Bedrock và các công cụ Giao thức bối cảnh mô hình (MCP). Chúng tôi đề cập đến giải pháp…
Gã khổng lồ công nghệ cho biết một nhóm có tên "Outsider Enterprise" đã sử dụng AI để lừa đảo hàng trăm nghìn nạn nhân, gửi 2,5 triệu tin nhắn văn bản trong khoảng thời gian hai tuần.
Đây không phải là công ty khởi nghiệp duy nhất giải quyết vấn đề AI vật lý, nhưng đây là một trong những công ty được tài trợ tốt nhất.
Anthropic đã khảo sát gần 52.000 người Mỹ về hy vọng và nỗi sợ hãi của họ đối với AI. 64% lo sợ mất việc làm và 56% lo lắng về việc mất khả năng tự suy nghĩ. Người dùng AI hàng ngày ít hợp tác hơn…
Quyền tự chủ hoàn toàn là rất hiếm, nhưng Ukraine đang cài đặt các mô-đun AI trên máy bay không người lái và robot.
OpenAI hiện cho phép người dùng Codex lưu trữ các thiết lập lại giới hạn tỷ lệ của họ và kích hoạt chúng theo cách thủ công thay vì xem chúng hết hạn theo lịch trình cố định. Nếu bạn đạt đến giới hạn sử dụng giữa phiên, bạn có thể đổi tiền mặt bằng khoản đặt lại đã lưu ngay lập tức trong…
Gã khổng lồ công nghệ cho biết một nhóm có tên "Outsider Enterprise" đã sử dụng AI để lừa đảo hàng trăm nghìn nạn nhân, gửi 2,5 triệu tin nhắn văn bản trong khoảng thời gian hai tuần.
Bạn sẽ được tha thứ nếu nghĩ rằng ngày này sẽ không bao giờ đến. Siri đã dành một thập kỷ rưỡi ở đâu đó giữa "hữu ích ở một số thứ" và "hoàn toàn tai hại, tại sao tôi thậm chí còn thử, liệu nó có thực sự không đặt ra mục tiêu gì không?"
Claude Fable 5 đứng đầu Chỉ số trí tuệ phân tích nhân tạo với 64,9 điểm và lập kỷ lục ở năm trên mười điểm chuẩn. Nhưng mức tăng so với Opus 4.8 chỉ là 5,7% với giá gấp đôi mã thông báo. Bộ lọc an toàn với f…
Những kẻ lừa đảo được cho là đã nhắm mục tiêu hàng trăm nghìn người bằng các trang web lừa đảo được mã hóa bởi Gemini.
Bài đăng này cho biết cách xây dựng trợ lý theo dõi và chuẩn bị cuộc họp tùy chỉnh bằng cách sử dụng máy chủ Amazon Quick và Cisco Webex MCP. Từ một lời nhắc duy nhất, nhân viên hỗ trợ tìm thấy cuộc họp Webex sắp diễn ra, xem xét các bản tóm tắt cuộc họp trước đó…
Bài đăng này phác thảo quá trình phát triển quy trình xử lý tài liệu thông minh có khả năng mở rộng và tiết kiệm chi phí trên AWS, được hỗ trợ bởi Amazon Bedrock và các tính năng của nó. BDA là một dịch vụ được quản lý trong Amazon Bedrock tự động…
Người sáng lập Amazon Jeff Bezos cho biết công ty khởi nghiệp AI mới của ông sẽ nỗ lực phát triển một "kỹ sư tổng hợp nhân tạo", theo báo cáo từ The New York Times và CNBC. Công ty khởi nghiệp có tên Prometheus nhằm mục đích phát triển AI-p…
AWS Professional Services (AWS ProServe) đã nén các mốc thời gian tương tác từ hàng tháng xuống hàng ngày, không phải bằng cách thêm các công cụ trí tuệ nhân tạo (AI) vào quy trình hiện có mà bằng cách xây dựng lại một cách cơ bản cách chúng tôi phân phối từ…
Việc tái sử dụng dữ liệu Pokémon Go để đào tạo AI tiếp tục thu hút sự chú ý.
OpenAI giới thiệu ba khóa học của Học viện giúp mọi người xây dựng các kỹ năng AI thực tế, tạo quy trình làm việc có thể lặp lại và áp dụng tác nhân vào công việc hàng ngày.
Kimi Work của Moonshot AI là một tác nhân máy tính để bàn cục bộ dành cho macOS và Windows. Nó chạy một nhóm gồm 300 tác nhân phụ, điều khiển trình duyệt đã đăng nhập của bạn thông qua WebBridge và lên lịch các công việc ở chế độ nền. Bài viết Moonshot AI ra mắt Kimi Work,…
Trong hướng dẫn này, chúng tôi xây dựng quy trình phân đoạn hình ảnh y tế 3D toàn diện bằng cách sử dụng MONAI để phân đoạn lá lách trên bộ dữ liệu Decathlon Task09 Phân đoạn y tế. Chúng tôi làm việc với máy quét CT thể tích, áp dụng hình ảnh y tế…
‘Nghe này, đó không phải mục đích tôi đến đây.’ | Hình ảnh: Apple Thử nghiệm ban đầu của chúng tôi đã cho thấy rằng Siri AI biết khi nào nên im lặng và điều đó phụ thuộc rất nhiều vào thiết kế. Trong một cuộc phỏng vấn với Mostly Human do MacRumors phát hiện, Crai…
Các tính năng tổng quát trong ứng dụng Ảnh mới của iOS 27 sẽ thêm các pixel giả vào một số ảnh của bạn, nhưng Jon McCormack của Apple cho biết công ty không sử dụng AI “vì lợi ích của AI”.
Preply sử dụng OpenAI để khởi chạy các bản tóm tắt bài học do AI tạo ra, cung cấp phản hồi được cá nhân hóa và các bài tập học ngôn ngữ.
<p>Sau hai ngày trải nghiệm <a href="https://simonwillison.net/2026/Jun/9/claude-fable-5/">Claude Fable 5</a>, tôi nghĩ cách tốt nhất để mô tả nó là <strong>chủ động không ngừng</strong>. Nó biết rất nhiều…
Nghiên cứu sâu hiện nằm trong Perplexity Computer, chia các câu hỏi khó thành các nhiệm vụ phụ và định tuyến trên hơn 20 mô hình biên giới. Bài đăng Sự bối rối chuyển nghiên cứu sâu vào máy tính, định tuyến các nhiệm vụ nghiên cứu phụ trên 2…
Thị trường trong thiết bị đầu cuối của Grok Build bao gồm các kỹ năng, tác nhân, hook và máy chủ MCP với xác minh cam kết-SHA trên mọi plugin từ xa. Bài đăng xAI vận chuyển Grok Build Plugin Marketplace với MongoDB, Vercel, Sentry, Ch…
Thibault Sottiaux đã giúp biến mã hóa AI trở thành một trong những doanh nghiệp phát triển nhanh nhất của OpenAI. Hiện anh ấy đang giám sát quá trình đại tu toàn diện ChatGPT.
Bài đăng này trình bày một quy trình xử lý tài liệu thông minh bao gồm cả tùy chọn suy luận theo yêu cầu và suy luận hàng loạt trên Amazon Bedrock để mang lại sự linh hoạt về thời gian xử lý tài liệu và…
Deezer introduced a tool that scans playlists from Spotify, Apple Music, and other platforms to identify AI music.
Deezer now offers a free AI music detector that lets users on any major streaming platform check whether AI-generated songs are hiding in their playlists. The article Free Deezer tool lets users on any streaming service…
Today, we’re excited to announce two new capabilities that make Quick Sight dashboards even more expressive and business-aligned: sparklines and custom sort for controls. In this post, we walk through both features, wha…
Pool's new app automatically sorts screenshots into personalized collections, tracks down the original links behind saved content, and helps you rediscover products, recipes, travel ideas, and other things you meant to…
<p><strong>Release:</strong> <a href="https://github.com/simonw/datasette/releases/tag/1.0a33">datasette 1.0a33</a></p> <p>This alpha is a significant step on the road to a stable 1.0, finally extending the <code>?_extr…
Blueprint instruction optimization is a BDA feature that automatically refines your extraction instructions to address this challenge directly. You provide three to ten example documents with expected values, and BDA re…
The new chatbot, called Ask DoorDash, allows users to search the app for what they're looking for in their own words instead of having to scroll through restaurants and stores to build a cart.
Anthropic has apologized for stealthily throttling its new AI model, Claude Fable 5, with hidden guardrails that undermine both researchers and rivals using it to develop competing systems. The company says it is revers…
The Hermes Agent dashboard now builds complete agent profiles in one flow, replacing multi-step CLI setup for users. The post Nous Research Ships Hermes Agent Profile Builder: Identity, Model, Skills, and MCP Servers in…
Deezer will now scan your playlists on other streaming platforms to detect AI-generated music. Deezer was the first of the big streaming services to start labeling AI-generated music. It even offered its tech to other p…
<p><strong>Release:</strong> <a href="https://github.com/simonw/asyncinject/releases/tag/0.7">asyncinject 0.7</a></p> <p>I built this utility library to support an <code>asyncio</code> dependency injection pattern a few…
The decision comes as India emerges as the world’s largest GCC market.
If founders and other business leaders weren't already envious of Dario Amodei, who sits atop one of the world's fastest-growing AI companies, they're going to be seriously envious now.
Frontier teams are not just using AI to code faster. They’re redesigning how software gets built. The result is 4.5x productivity gains, in some cases more than 10x.
OpenAI supports the EU Code of Practice on AI content transparency, advancing provenance standards and tools to help people understand AI-generated content.
Profiling in PyTorch (Part 2): From nn.Linear to a Fused MLP
Discover how astrophysicist Chi-kwan Chan uses Codex to build black hole simulations, helping scientists study extreme physics and test Einstein’s theory of general relativity.
Learn how BBVA scaled ChatGPT Enterprise to 100,000 employees and partnered with OpenAI to accelerate AI-powered banking transformation worldwide.
<p><strong>Release:</strong> <a href="https://github.com/datasette/datasette-agent/releases/tag/0.2a0">datasette-agent 0.2a0</a></p> <p>Highlights from the release notes:</p> <blockquote> <ul> <li>Tools can now ask the…
We implement an instrumented workflow for Microsoft SkillOpt end to end. We set up the repository, connect OpenAI-compatible model access, and configure the optimizer and target models. We evaluate the original seed ski…
<p><strong><a href="https://blog.google/innovation-and-ai/technology/developers-tools/diffusion-gemma-faster-text-generation/">DiffusionGemma</a></strong></p> Last May Google briefly released an experimental Gemini Diff…
Diffusion AI is most common in image generation, but it can make text outputs much faster.
A car pulls up to the curb. The app says, “Your ride is here.” No one’s in the driver’s seat. For people who live in one of the dozens of cities now hosting robotaxi services, this is already a reality. The robotaxi ind…
Anthropic just released Claude Fable 5, calling it the most powerful AI model it has ever made widely available and praising its skills in biology, among others. But the model won't answer basic biology questions - the…
New college graduates around the country have been booing and heckling commencement speakers who hype up AI. Microsoft would like everyone to talk it out. In a blog post running more than 3,100 words, Microsoft vice cha…
Anthropic's security team found that its Mythos Preview AI model can turn security patches for Firefox and the Windows kernel into working exploits within hours, for a few thousand dollars and no specialized knowledge.…
A group of independent musicians is suing Google claiming it trained Lyria on their uploads. | Image: Cath Virginia / The Verge If you've uploaded a song to YouTube, Google almost certainly considers your video fair gam…
Anthropic released Claude Fable, its first Mythos-class AI model, yesterday and it's already causing concerns inside Microsoft. Sources tell me that Microsoft is limiting the use of Claude Fable 5 for employees because…
Google is making some changes to how it saves your interactions with Search. In an email sent to users, Google says it will save the images, files, audio, and video you use to search under a new "Search Services History…
New research suggests that AI memory systems can degrade model performance and encourage sycophantic tendencies.
Cybersecurity researchers are complaining that Anthropic's new model Fable has guardrails that are too strict for any cybersecurity work.
Today, we’re announcing the Neuron Agentic Development capabilities: a collection of AI agents and skills that make this possible for developers building on AWS Trainium and AWS Inferentia. In this post, we explain how…
<blockquote cite="https://twitter.com/jeremyphoward/status/2064595816875217362"><p>Easy solution to slow down recursive AI self improvement:</p> <ul> <li>The lab with the top-ranked model must agree THEY must not use it…
In this post, you build an AI-powered equipment repair assistant using Amazon Bedrock AgentCore that helps farmers and field technicians diagnose equipment problems, identify required parts, and access manufacturer-appr…
The ACLU is suing two Florida police departments over the arrest of a Fort Myers man in a child-abduction case, saying officers treated a flawed face-recognition match as a near-certain ID.
Anthropic has released Claude Fable 5, the first model in its new Mythos class. It leads nearly every benchmark, including SWE-bench Verified at 95 percent, but costs twice as much as Opus 4.8 at 10 or 50 dollars per mi…
Decart is launching Oasis 3, a real-time world model that generates photorealistic driving environments for autonomous vehicle testing, now available via API for developers to build on.
Google is giving NotebookLM a major upgrade. The research tool now runs on Gemini 3.5 Flash, has its own cloud computer for code execution, and can find sources on its own via Google Search. In internal tests, the new s…
Software development has changed. Engineers no longer type most code by hand. They describe intent, and AI agents do the work. Modern tools plan tasks, edit across files, run tests, and open pull requests. Many now ship…
The Argentine national team will be Google’s test bench and technological showcase during the World Cup.
Claude Fable 5 ships generally available with classifiers; Mythos 5 stays limited, cyber safeguards lifted, through Project Glasswing. The post Anthropic Releases Claude Fable 5 and Claude Mythos 5: Same Underlying Mode…
In this tutorial, we work with NVIDIA's Nemotron-Pretraining-Code-v3 dataset as a large-scale metadata index for code pretraining research. We stream the dataset instead of downloading it, inspect its schema, and build…
<p><strong><a href="https://jonready.com/blog/posts/claude-fable5-is-allowed-to-sabotage-your-app-if-youre-a-competitor.html">If Claude Fable stops helping you, you'll never know</a></strong></p> Jonathon Ready hig…
See how LSEG uses OpenAI to scale trusted AI across its global business, accelerating insights, shrinking release cycles, and empowering 4,000 employees.
<p>I didn't have early access to today's <a href="https://www.anthropic.com/news/claude-fable-5-mythos-5">Claude Fable 5</a> release, but I've spent the past ~5.5 hours putting it through its paces. My initial impressio…
Siri, are you there? Parents want one thing, and one thing only, out of AI: to add a list of soccer games or "spirit week" theme days from an email or a poorly formatted flyer onto their calendar in one shot. And I have…
<p><strong>Release:</strong> <a href="https://github.com/simonw/llm/releases/tag/0.32a3">llm 0.32a3</a></p> <p>Almost entirely written by the new Claude Fable 5, see <a href="https://simonwillison.net/2026/Jun/9/claude-…
<p><strong>TIL:</strong> <a href="https://til.simonwillison.net/llms/agentsview-custom-model-price">Setting a custom price for a model in AgentsView</a></p> <p>I've been really enjoying <a href="https://agentsview.io/">…
Microsoft AI CEO Mustafa Suleyman says it's "really, really dangerous" for Anthropic to speculate about Claude's consciousness inside its "constitution," or the instructions that tell the model how to behave. During an…
Can Voice Agents Handle Bilingual Customers? Benchmarking Frontier ASR on Code-Switched Speech
New frontier model refuses cybersecurity, biology, and chemistry queries.
<blockquote cite="https://twitter.com/karpathy/status/2064409694761054332"><p>I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand fo…
Voice translations preserve speaker's tone, pacing, pitch—with SynthID watermarks for security.
Anthropic ships two new models, Claude Fable 5 and Mythos 5, that claim to blow past the current Opus generation, especially in coding and research. Fable 5 finished a code migration for Stripe in one day that would hav…
Apple primarily made the case for an improved experience with its longstanding Siri assistant, which like most other announcements had a hefty helping of AI.
Gemini 3.5 Live Translate streams speech-to-speech translation across 70+ languages. It generates audio continuously, staying a few seconds behind the speaker. The model reaches developers via the Gemini Live API, plus…
Google releases Gemini 3.5 Live Translate, an audio model for real-time translation across more than 70 languages. The system translates continuously without waiting for a sentence to end and claims to preserve the spea…
Anthropic is releasing Claude Mythos 5 to trusted organizations and Claude Fable 5 to the public, a version it says can’t be used for cyberattacks.
Anthropic just announced Claude Fable 5, a new AI model it said is the most powerful model it has ever made widely available. According to the company, Fable 5 "shows exceptional performance in software engineering, kno…
Anthropic is releasing Claude Fable 5, its first Mythos-class model available to the public. The model comes with guardrails that block responses in high-risk areas like cybersecurity and biology.
In this post, we demonstrate how a hands-free FNOL intake system combines agents built with the Strands Agents SDK for domain reasoning with Amazon Bedrock AgentCore Browser Tool for live portal interaction. This approa…
Apple’s feature showcase at WWDC 2026 didn’t flag which if these “photographs” are real or created with its new AI fakery. | Images by Apple / compiled by The Verge Apple used to question whether generative AI-powered e…
This post shows engineering teams how to apply that principle to one of the most time-sensitive workflows in engineering: incident triage. You will build a custom incident triage assistant agent using Amazon Quick that…
With SpaceX, Anthropic, and OpenAI all eyeing massive public debuts, the tech industry may soon have a new class of corporate overlords — and a new acronym to match. Say goodbye to FAANG and hello to MANGOS.
Introducing North Mini Code: Cohere’s First Model For Developers
Gemini 3.5 Live Translate brings near real-time, natural speech translation to Google AI Studio, Google Translate and Google Meet.
Microsoft AI head Mustafa Suleyman is walking back his statement about AI automating jobs done by white-collar workers, including lawyers, accountants, and project managers. During an episode of Decoder on Monday, Suley…
Apple kicked off its annual developer conference with bold promises about AI. The company, CEO Tim Cook said, would be "introducing new technologies and innovations that push the limits on what's possible." But its slew…
Most of Apple's current AI ideas are roughly the same as everyone else's AI ideas. A chatbot you can ask questions; quick ways to create or summarize text; bizarre, borderline creepy image-generation tools. The company…
Some models run in Google's cloud, but without giving Google any kind of access.
At WWDC 2026, Apple showed off a rebuilt version of Siri. The assistant runs on foundation models developed with Google. For complex queries, it taps Nvidia GPUs. The article Apple Intelligence gets a second shot with h…
OpenAI is backing away from fully autonomous AI research by 2028, now talking about a "tandem" between humans and machines. Altman and Pachocki also call for an international body that could slow frontier development if…
In 2019, Alex Vindman testified during President Trump’s first impeachment trial–a decision that ended his military career. Now he wants to challenge the president from the halls of Congress.
As adoption of AI agents looks set to surge by as much as 300% in the next two years, leadership teams are carefully considering the implications of a hybrid human-AI workforce. Unlike existing enterprise-level automati…
At SXSW London last week I gave a talk called “Five things you need to know about AI,” in which I shared what I think are the biggest themes in AI right now. I pulled a few things from our first AI10 list, an annual gui…
A new Harvard and Perplexity paper uses matched-pair sessions to compare an autonomous agent with a search assistant. It finds large gains in autonomy, time, and cost, plus broader scope of work attempted. The post A Ne…
Can Apple's new AI glow-up put to bed accusations that it's losing an all-important industry race?
<p>Given how badly burned anyone who took Apple's <a href="https://simonwillison.net/2024/Jun/10/apple-intelligence/">2024 WWDC Apple Intelligence announcements</a> at face value was, I'm holding to a strict "I'll belie…
Apple is trying to solve one of Safari's biggest weaknesses with AI. Safari has long lacked the robust library of extensions that its rivals have, mainly due to the stringent development requirements from Apple. But now…
New features coming this fall alongside two-tiered, Google-powered AI model overhaul.
Apple’s WWDC 2026 event kicked off this morning at 10 a.m. PT at Apple Park, starting a week full of expected announcements around Siri, iOS 27, Apple Intelligence and more, along with developer events and demos. This y…
NotebookLM is getting a big upgrade, but it's only for AI Ultra and enterprise accounts right now.
Apple is adding new AI-powered features to Safari, Shortcuts, and Password apps.
Shortcuts gets an AI upgrade, letting you describe the workflow you want in a prompt.
Apple's AI image generator is getting a makeover that could make it more competitive.
A new spatial "Reframe" feature will let users use AI to adjust perspectives.
73 packages run self-replicating stealer as soon as they're opened by an AI agent.
Siri is finally getting its own app.
"If you're grabbing a bite with friends and point your iPhone at the bill, then [you can] select what you ordered to split the tab with Apple Cash," said Apple VP of Software Sebastien Marineau-Mes.
From a stand-alone app to a Google Gemini partnership, here’s everything you need to know from WWDC 2026 about Apple’s upcoming overhaul of Siri.
The idea behind the new "Siri AI" is to turn the assistant from a voice controlled assistant into an AI companion that can do a lot more.
Amazon is expanding its print-on-demand features to AI-generated designs created using Alexa for Shopping for products like T-shirts, water bottles, and hoodies. Shoppers can use text prompts to generate images that are…
Two years after first revealing its plans for Apple Intelligence and a smarter Siri that never fully materialized, at WWDC, Apple just revealed a new set of AI features and a smarter, more personalized Siri. Apple calls…
The code WIRED identified is gone from the latest version of Meta AI, the companion app for the company’s smart glasses. Meta won’t say why or whether it’s coming back.
Amazon Bedrock AgentCore Runtime gives each agent session its own isolated microVM with a persistent workspace, secure tool access through Gateway, and built-in observability—so you can run Claude Code, Codex, Kiro, and…
In this post, we introduce mathematical optimization, explain how it fits within the broader AI landscape, and showcase real-world success stories where the Innovation Center has partnered with customers to deliver conc…
This blog has previously discussed FHE for ML inference in the post Enable fully homomorphic encryption with Amazon SageMaker endpoints for secure, real-time inferencing, but this post goes a little further. That previo…
In this post, we cover the structure of Amazon Quick ARNs and provide a practical mental model for working with them. By the end, you can look at an ARN and immediately understand what it means for your migration strate…
Google is rolling out "across the board" updates to NotebookLM. The AI-powered note-taking app now uses Google's upgraded Gemini 3.5 model, which will allow it to respond with "more accurate and reliable information," a…
Only 26 percent of companies have full visibility into their AI costs, a KPMG survey finds. The article Most companies are flying blind on AI spending appeared first on The Decoder.
OpenAI confirms a confidential S-1 submission to the SEC and has not yet determined timing for further action.
Today I’m talking with Mustafa Suleyman, the CEO of Microsoft AI. And I’m actually going to keep today’s intro short — I’m working from my wife’s family farm this week, as you’ll see in the video, but also this is a rea…
Results from a randomized controlled trial show the potential of Gemini’s Guided Learning feature to boost engagement and accelerate learning.
Meta has put a number on the security breach in its AI support chatbot for Instagram for the first time: at least 20,225 accounts were compromised. For nearly seven weeks, the system sent password reset links to arbitra…
Machine learning has its limits—how is it being used?
Apple's biggest event of the year is nearly here. The company's Worldwide Developers Conference will spotlight updates to iOS, macOS, and all of Apple's other operating systems, and this year's event could also include…
Moms are outsourcing tedious household tasks to ChatGPT and selling courses teaching others to do the same. Where are all the dads?
Microsoft AI has released MAI-Transcribe-1.5, the second iteration of its in-house speech-to-text family. The model covers 43 languages, adds keyword (entity) biasing for domain-specific terms, posts a 2.4% Word-Error-R…
A year ago at London Tech Week, NVIDIA founder and CEO Jensen Huang and U.K. Prime Minister Keir Starmer made a declaration: the U.K. would be an AI maker, not an AI taker. At this year’s event, NVIDIA and its partners…
A vision for the future of AI, focusing on access, safety, and shared prosperity as OpenAI works to ensure AGI benefits everyone.
OpenAI launches the Economic Research Exchange to study AI’s impact on jobs, productivity, and the economy. Applications are now open for selected research projects.
<p><strong>Release:</strong> <a href="https://github.com/datasette/datasette-agent-edit/releases/tag/0.1a0">datasette-agent-edit 0.1a0</a></p> <p>I'm planning several plugins for <a href="https://agent.datasette.io/">Da…
NVIDIA and Doosan Group are expanding their collaboration to advance new opportunities across physical AI, robotics and AI factory infrastructure, spanning Doosan Robotics, Doosan Bobcat, Doosan Enerbility and Doosan Co…
We're likely to see more price increases as the big AI companies plan to go public.
Amazing Digital Dentures (a failed project)
Mythograph Atelier #1 - Abstract Art That Means Something to You
Notion's head of product said he was "astonished" at “the amount of people RT-ing this."
In this tutorial, we use GEPA as a reflective prompt-evolution framework to improve how a small language model solves multi-step arithmetic word problems. We start from a weak seed prompt, build a deterministic benchmar…
"Chat is dead" — at least, according to a senior OpenAI employee.
Aitana Lopez, AI avatar by creative agency The Clueless. | Image: The Clueless This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on AI confusion, follow Robert Har…
Sponsors especially OPENAI CODEX voucher usage for codex - openAI challange
OpenAI is planning the biggest overhaul of ChatGPT since its launch. The chatbot will become a "superapp" bundling coding tools, AI agents, and partner apps like Canva and Booking.com. "Chat is dead," the company says i…
How accurate does an AI system need to be?
Perplexity's new "Search as Code" architecture dumps rigid search APIs and lets AI models write their own search routines in Python. By letting the agent handle its own filtering and deduplication inside a sandbox, the…
OpenAI's new Lockdown Mode for ChatGPT disables web access, Deep Research, and Agent Mode to make data theft through prompt injection attacks harder. The mode doesn't fully prevent such attacks, it only blocks the final…
Low-code and no-code AI platforms now turn a prompt into a working app, agent, or model. This guide compares 21 tools across app builders, automation, AI agents, and machine learning platforms, each linked to its offici…
At GTC Taipei at COMPUTEX last week, NVIDIA unveiled RTX Spark, the superchip that reinvents Windows PCs for the era of personal AI agents. On the heels of this announcement, NVIDIA founder and CEO Jensen Huang headed t…
UIUC and Chroma's Harness-1 is a 20B retrieval subagent trained with reinforcement learning inside a stateful search harness. The harness maintains the bookkeeping — candidate pool, importance-tagged curated set, eviden…
Even with Lockdown Mode, ChatGPT could be still vulnerable to prompt injections, but the goal is to reduce the likelihood that sensitive data gets shared in the process.
Five labs, five minds: building a multi-model finance drama on small models
Apple's WWDC nears: Here's what you can look forward to.
President Donald Trump said he's discussing deals "where the American people can benefit from the success of AI."
An AI-generated image of the royal family featuring two Queen Elizabeth IIs. | Image: Meta AI Facebook has long been filled with feeds of clickbait articles. Now, Meta is making its own clickbait articles with AI. The s…
Our first glimpse of the new AI Siri came all the way back at WWDC 2024. Apple has been on its back foot, AI-wise, for the past few years. But in a strange way, playing from behind might not be such a bad move. At WWDC…
Elon Musk's xAI used Anthropic's Claude to train its own coding models for months and kept going even after Anthropic cut off access, using private accounts and the Blackbox AI service. Meanwhile, xAI's pretraining team…
Plus: Hackers use Meta’s AI bots to hack Instagram accounts, Anthropic helps NSA hackers, a decades-long GPS satellite mystery may have been solved, and more.
NVIDIA released Nemotron 3.5 ASR, a cache-aware 600M streaming model transcribing 40 language-locales in real time from one checkpoint. The post NVIDIA Releases Nemotron 3.5 ASR: A 600M-Parameter Cache-Aware Streaming M…
Alibaba's Qwen team has released Qwen3.7-Plus, a multimodal agent model that combines visual perception, GUI operation, and coding in a single agent loop. In a demo, an agent built on the model autonomously developed a…
<p><strong>Release:</strong> <a href="https://github.com/simonw/micropython-wasm/releases/tag/0.1a2">micropython-wasm 0.1a2</a></p> <p>I added a CLI to <code>micropython-wasm</code> (<a href="https://github.com/simonw/m…
<p>I've been experimenting with different approaches to running code in a sandbox for several years now, but my latest attempt feels like it might finally have all of the characteristics I've been looking for. I've rele…
<p><strong><a href="https://help.openai.com/en/articles/20001061-lockdown-mode">OpenAI Help: Lockdown Mode</a></strong></p> OpenAI first teased this <a href="https://openai.com/index/introducing-lockdown-mode-and-elevat…
Set up Qualcomm AI Hub Models to run MobileNet-V2 inference, YOLOv7 detection, and compile models on real devices. The post A Hands-On Coding Tutorial on Qualcomm AI Hub Models for Classification, Object Detection, and…
Applications for Startup Battlefield 200 officially close on June 8, 11:59 p.m. PT. Don't wait any longer. Secure your shot at competing on the Disrupt Stage at TechCrunch Disrupt 2026 this October at San Francisco's Mo…
Compare Gemma 4 edge formats: BF16, Q4_0 QAT, and mobile QAT, on published memory numbers and design tradeoffs. The post Google DeepMind Releases Gemma 4 QAT Checkpoints: Q4_0 and a New Mobile Format Cut On-Device Memor…
We're now deep into developer conference season, and one of the themes so far is the relentless conviction from Big Tech companies that AI is going to change everything about how we do everything. Nvidia's Jensen Huang…
The Air succeeds as a minimalist, reliable fitness tracker, but Google's AI Health Coach feels unnecessary.
Microsoft CEO Satya Nadella has sharply criticized an internal memo proposing to make users "addicted" to the company's new AI agent Scout. "Not sure who is writing and leaking this nonsense," Nadella wrote to about 50…
Microsoft’s AI products aren’t selling, and Github’s been plagued with troubles. WIRED spoke with VP Scott Hanselman about whether the company is in catch-up mode.
"The whole conversation shifted from tokenmaxxing and 'go fast' to 'we need guardrails, how do we control this?'"
Here are Google’s latest AI updates from May 2026
When Quilty hit the industry trades earlier this year, the AI startup promised that its tool could accurately predict a film's success just by reading the script. When people actually got a chance to experiment with Qui…
Microsoft sells its LLM training approach as different from other AI companies. It isn't. The company trained its new MAI models partly on unlicensed web data like Common Crawl, despite claiming they used only "clean an…
Anthropic has reportedly stationed about half a dozen engineers directly at the NSA to adapt its Mythos AI model for offensive cyber operations. The model could be used to break into networks in China or Iran. That fits…
<blockquote cite="https://ladybird.org/posts/changing-how-we-develop-ladybird/"><p>We will no longer accept public pull requests. [...]</p> <p>A substantial patch used to imply substantial effort, and that effort was a…
NVIDIA Dynamo Snapshot checkpoints and restores vLLM inference workers on Kubernetes using CRIU and cuda-checkpoint tools. The post NVIDIA AI Releases Dynamo Snapshot: A CRIU-Based Fast Startup System for AI Inference o…
Perplexity AI announces a hybrid local-server inference orchestrator for Personal Computer, automatically routing AI tasks between on-device and cloud models. The post Perplexity AI Introduces Hybrid Local-Server Infere…
A hands-on guide to running Microsoft Fara in Colab, testing the browser agent loop with a mock endpoint. The post Microsoft Fara Tutorial: Run a Browser-Use Agent in Google Colab with a Mock OpenAI-Compatible Endpoint…
On June 5, 404 Media reported that attackers had been using Meta’s AI customer support agent to steal Instagram accounts. Their approach was simple: They asked the agent to link the accounts to email addresses that they…
AI companies are using serif to project humanity. Critics are calling it “tasteslop.”
Anthropic is sharing internal data showing how much Claude is speeding up its own AI development: more than 80 percent of production code now comes from Claude, and engineers are shipping eight times as much code per da…
Home to cutting-edge sovereign AI infrastructure and robotics innovators, as well as one of the world’s most passionate gaming communities, South Korea is one of the world’s centers of AI. NVIDIA founder and CEO Jensen…
In the current environment, remaining heads down has diminishing returns; at some point, you have to make some noise just to remind the market you exist.
<p><strong><a href="https://charitydotwtf.substack.com/p/ai-enthusiasts-are-in-a-race-against">AI enthusiasts are in a race against time, AI skeptics are in a race against entropy</a></strong></p> Charity Majors neatly…
The Airbnb CEO said last year it hasn't struck an LLM partnership because existing products weren't quite ready.
This tutorial walks through a complete NLP pipeline for research-level mathematics. Using the ResearchMath-14k dataset, we extract field-specific keywords with TF-IDF, generate sentence embeddings, visualize the problem…
Robot demonstrations can distort public perceptions of robotic capabilities.
Poke, the startup that lets people use AI agents through simple text messages, has become the first AI agent approved for Apple’s Messages for Business platform.
Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI
Bot traffic now outpaces human traffic on the internet, Cloudflare CEO Matthew Prince says, years ahead of his late 2027 forecast. He blames AI agents for the surge. His conclusion for the future of the web: "Clearly it…
Deploy NVIDIA Nemotron 3 Ultra on Amazon SageMaker JumpStart. Get 5x faster inference and 30% lower cost for agentic AI workloads with this frontier reasoning model.
ChatGPT's updated "Dreaming" memory system now builds coherent user profiles from conversations instead of saving scattered bullet points. OpenAI says the success rate for keeping information current jumped from 52.2 pe…
<blockquote cite="https://www.404media.co/google-employees-internally-share-memes-about-how-its-ai-sucks/"><p>After this story was published Google's spokesperson reached out and asked us to publish a slightly different…
Creators often have to parse through charts and dashboards to understand their performance, but with the new AI assistant, they can get quick answers to questions like "When should I post?" and "What are people saying i…
Apple's WWDC nears: Here's what you can look forward to.
According to a Bain survey of 951 companies, almost 40 percent achieved less than 10 percent in AI cost savings, even though most had targeted 11 to 20 percent. One alleged reason is that only 7 percent actually run ful…
Taiwan Semiconductor Manufacturing Co. - the world's biggest semiconductor-maker - is struggling to meet demands from American customers even with its factory buildout in the US, according to reports from Reuters and Bl…
Today on Decoder, I’m talking to Ryan Mac, a technology reporter at The New York Times and coauthor of the excellent book Character Limit: How Elon Musk Destroyed Twitter, which came out in 2024. I can’t recommend it en…
OpenAI CEO Sam Altman outlines the next phase of AI products: a "proactive AI" that runs constantly in the background and acts on its own instead of waiting for user prompts. Companies are also wrestling with spiraling…
June’s forecast with GeForce NOW: 100% chance of gaming. GeForce NOW is lining up new adventures for the month, from big-name blockbusters to quirky indies ready for the spotlight. Members can dive into fresh worlds, sq…
How to Fine-Tune Nemotron 3.5 ASR for Your Language, Domain, or Accent
Nobody should be subjected to seeing shrimp Jesus all over their social feeds. | Image: Cath Virginia / The Verge, Getty Images It's almost impossible to avoid seeing AI-generated content online, but it doesn't have to…
EVA-Bench Data 2.0: 3 Domains, 121 Tools, 213 Scenarios
Some of the AI industry's biggest rivals have put their many, many grievances aside for a common cause: making it harder for people to use their technology to develop biological weapons. In an open letter to US lawmaker…
Learn how Endava is using AI agents, ChatGPT Enterprise, and Codex to accelerate software delivery, automate workflows, and build an AI-native culture across the enterprise.
The design hasn’t changed much from the original Proteus, which was announced in 2022. | Image: Amazon Amazon has announced a new version of its fully autonomous warehouse robot, Proteus, that will interact using langua…
ChatGPT introduces a new memory system to better remember preferences, keeping context fresh and relevant across conversations.
xAI has released "grok-imagine-video-1.5-preview," an image-to-video model that turns still images into cinematic videos at up to 720p based on text prompts. Multiple clips can be stitched together into longer scenes. T…
The model handles long documents, images and audio in a single context window, expanding the design space for agentic and retrieval-heavy workloads.
Leading AI labs, executives, and scientists are sending a letter to lawmakers urging them to improve tracking of synthetic DNA sequences that could be used for bioweapons.
Designing the hf CLI as an agent-optimized way to work with the Hub
An action plan for AI-powered biological resilience
Lovable and Google signed an expanded multiyear deal that involves a 5x expansion of Lovable's footprint on Google Cloud, and expanded access to Anthropic Claude.
Google must change AI Overviews after claiming users don't want "lots of sources."
In this post, we introduce Amazon Bedrock Ops Alert, a three-layer automated monitoring solution that proactively detects operational issues, dynamically adjusts alarm thresholds, classifies alarms by category, automati…
We build a document intelligence backend with iii by registering modular functions and reusing them across multiple triggers. The post How to Build a Document Intelligence Backend with iii Using Workers, Functions, and…
Gemma 4 12B uses a new encoding scheme and token prediction to punch above its weight.
Dreambeans is a curated list of AI-illustrated "stories" culled from the personal data in your Google account.
Critics say Trump plan to test AI models is short-sighted, performative.
Spencer Huang, Nvidia’s robotics lead, tells WIRED that the new bot combines the best of both worlds.
In this post, we show you how to get started with NEXUS on Amazon SageMaker JumpStart, walk through the deployment process, and demonstrate how to run predictions against your enterprise datasets.
This week we've got tandem hands-ons with Google's new Gemini AI agent - Spark - from my colleagues David Pierce and Jay Peters. Their takeaways are similar: It's so effective that it's scary. Spark knew that David's do…
In this post, we look at how to use SOCI on publicly available Deep Learning AMIs and Containers, when to use the various SOCI modes provided by the tool, and how to quickly and efficiently use this tool in your workloa…
Amazon's updated search bar will now show you AI-generated images of products as you describe them. For now, the in-app feature only surfaces AI images of clothing and home goods, allowing you to tap on the image that b…
In this post, you learn how to use Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) together to improve the tool-calling accuracy of a small language model (SLM). The example uses Amazon SageMaker A…
Amazon will use visual search and AI to show AI-generated product images that match your search queries. The retailer says it will help guide users to products.
What makes a robot gripper useful isn’t that it can pick up one object — it’s that it can pick up the next one, and the one after that, with a tool it’s never held before. What makes an autonomous vehicle system safe is…
At CVPR, NVIDIA is unveiling new physical AI agent skills that help researchers and developers speed the development of autonomous vehicles, robots and vision AI systems. The core challenge in physical AI research isn’t…
Doubts linger over whether Meta can close the gap with rivals.
GPT-Rosalind advances life sciences research with enhanced biological reasoning, medicinal chemistry expertise, genomics analysis, and experimental workflow capabilities.
Uncover second-hand scores with AI tools in Google Search and Shopping.
Direct Preference Optimization Beyond Chatbots
<p><strong><a href="https://www.bloomberg.com/news/articles/2026-06-02/uber-caps-usage-of-ai-tools-like-claude-code-to-cut-costs">Uber Caps Usage of AI Tools Like Claude Code to Manage Costs</a></strong></p> I wrote <a…
See how Wasmer used Codex with GPT-5.5 to build a Node.js runtime for the edge, accelerating development 10x to 20x and shipping in weeks instead of months.
Hermes Desktop is a no-terminal GUI sharing one agent core, skills, and memory with the Hermes Agent CLI. The post Nous Research Releases Hermes Desktop: A Native Cross-Platform Front End for Hermes Agent v0.15.2 with S…
NVIDIA released Cosmos 3, open omnimodal world models pairing an autoregressive VLM reasoner with a diffusion generator for physical AI. The post NVIDIA Releases Cosmos 3: A Two-Tower Mixture-of-Transformers Foundation…
Adding MCP Tools to Reachy Mini
<p>Microsoft <a href="https://microsoft.ai/news/building-a-hillclimbing-machine-launching-seven-new-mai-models/">announced two new text LLMs</a> this morning - <strong><a href="https://microsoft.ai/news/introducing-mai-…
Accelerated computing has revolutionized industrial engineering, compressing simulation times from weeks to hours. Today’s remaining challenges sit in the end-to-end workflow surrounding the simulations: computer-aided…
Microsoft missed the boat on apps, so get ready for agents.
<p><strong>Release:</strong> <a href="https://github.com/datasette/datasette-agent-micropython/releases/tag/0.1a0">datasette-agent-micropython 0.1a0</a></p> <p>I want <a href="https://agent.datasette.io">Datasette Agent…
<p><strong>Release:</strong> <a href="https://github.com/simonw/micropython-wasm/releases/tag/0.1a1">micropython-wasm 0.1a1</a></p> <p>Fixes for some limitations that emerged while I was trying to use this to build <cod…
The agentic AI moment has arrived, but delivering on its promise requires more than good models. It also takes fast hardware, secure runtimes, a responsive data layer and models tuned for long-running reasoning. NVIDIA…
International Mathematical Union endorses warning about tech industry influence.
<p><img src="https://static.inaturalist.org/photos/671786719/large.jpg" alt="California Brown Pelican"></p><p>California Brown Pelican, in Fort Mason, CA, US</p><p>I'm at the <a href="https://build.microsoft.com/">Micro…
Google's June Android feature drop includes more scam detection, more AirDrop, and yes, more AI.
Fine-tuning for domain-specific tasks means improving performance in one area without degrading the model’s general capabilities, and getting that balance right is harder than it looks. This post walks through how to na…
In this post, we'll walk through implementing object detection with Amazon Nova 2 Lite. You'll learn how to deploy an object detection application using Amazon Bedrock, AWS Lambda, and Amazon API Gateway. You'll also le…
This post walks through how Baz built their Spec Review agent using Amazon Bedrock and Amazon Bedrock AgentCore. We'll cover the architecture decisions, implementation details, and the business outcomes they achieved by…
Holo3.1: Fast & Local Computer Use Agents
Travelers built an AI-powered Claim Assistant with OpenAI to guide customers through filing claims, provide 24/7 support, and scale operations during peak demand.
The global health care sector is under increasing strain. Decades of chronic underinvestment and constraints in recruitment have coincided with a surge in demand for services for aging populations. Gaps in provision are…
This article is from Making AI Work, MIT Technology Review’s limited-run newsletter examining how to apply LLMs across industries. To receive it in your inbox,sign up here. From accounting to design to market research a…
Financial institutions have spent years building AI: fraud models, credit models, recommendation engines and risk systems. While this sprawl of task-specific models has been effective, it’s also constrained by siloed sy…
<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/pasted-file-editor">Pasted File Editor</a></p> <p>I really like how you can paste a large volume of text into <a href="https://claude.ail">claude.ai</a>…
Agentic AI is getting physical. At COMPUTEX on Tuesday, NVIDIA announced NVIDIA JetPack 7.2 and NVIDIA NemoClaw support on NVIDIA Jetson. JetPack 7.2 brings agentic AI skills, Yocto project support, NVIDIA CUDA 13 on NV…
Learn how Googlers used AI to produce Google I/O 2026.
Introducing Mellum2: A 12B Mixture-of-Experts Model by JetBrains
Beyond LLMs: Why Scalable Enterprise AI Adoption Depends on Agent Logic
An updated agentic model improves multi-step tool use and reliability on long tasks, a focus area as agent workloads move toward production.
We used Google AI Studio to vibe code a quiz about our top I/O 2026 announcements.
Watch 9 videos showing the capabilities of Gemini Omni and Gemini 3.5, announced at Google I/O 2026.
University of Waterloo students develop AI prototypes like sign language tutors to reshape the future of education and work.
Here are 12 of the biggest Google I/O 2026 keynote moments, including news about Gemini Omni, Gemini 3.5 Flash and more.
It is one thing to say AI will change the world. It is another to expect the class of 2026 to applaud it. In fact, when former Google CEO Eric Schmidt told University of Arizona graduates that their task is to help shap…
Amid rapidly growing adoption of enterprise-level AI agents, there’s a disconnect emerging between ambition and execution. Although 85% of organizations say they want to be agentic within the next three years, 76% say t…
Artificial intelligence has not so far produced a clean story of mass unemployment. Aggregate employment in developed countries remains broadly stable, and recent assessments have found limited evidence that AI has shif…
Haven’t you heard? White-collar jobs are going away, decimated by AI. Waves of layoffs in the tech sector (most recently at Coinbase and Meta and Cisco) are said to presage what will soon come for all of us knowledge wo…
A recap of the 2026 I/O Dialogues, where leaders discuss the future of AI, quantum computing, robotics and creativity.
For a quarter century, the Google search box has been one of the most recognizable interfaces in computing: a thin white rectangle, a blinking cursor, a few typed words, and a list of blue links. On Tuesday, Google will…
Biologists use Co-Scientist to find novel factors that successfully rejuvenate human cells.
We’re expanding access to Google AI Ultra subscribers globally and introducing a new capability powered by Street View.
A collection of science tools and experiments to expand the scale and precision of scientific exploration.
We're expanding our tools to help you understand how content was created and edited across the web.
Google DeepMind and Singapore partner to apply frontier AI to address complex challenges across health, education, and sustainability and more.
.apr-fig { text-align: center; margin: 1.35em 0; line-height: 1.4; } .apr-fig--wide img { display: inline-block; width: 100%; max-width: 100%; height: auto; vertical-align: middle; } .apr-fig--wide-0-8 { max-width: 80%;…
.grasp-results-table table { font-size: 0.875rem; line-height: 1.35; width: 100%; } .grasp-results-table th, .grasp-results-table td { padding: 0.35rem 0.5rem; } /* Consistent whitespace between major sections (this pos…
--> Understanding the behavior of complex machine learning systems, particularly Large Language Models (LLMs), is a critical challenge in modern artificial intelligence. Interpretability research aims to make the decisi…
Salesforce on Tuesday launched an entirely rebuilt version of Slackbot, the company's workplace assistant, transforming it from a simple notification tool into what executives describe as a fully powered AI agent capabl…
An encoder (optical system) maps objects to noiseless images, which noise corrupts into measurements. Our information estimator uses only these noisy measurements and a noise model to quantify how well measurements dist…
In this post, I’ll introduce a reinforcement learning (RL) algorithm based on an “alternative” paradigm: divide and conquer. Unlike traditional methods, this algorithm is not based on temporal difference (TD) learning (…
What exactly does word2vec learn, and how? Answering this question amounts to understanding representation learning in a minimal yet interesting language modeling task. Despite the fact that word2vec is a well-known pre…
.modal { display: none; position: fixed; z-index: 9999; padding-top: 50px; left: 0; top: 0; width: 100%; height: 100%; overflow: auto; background-color: rgba(0,0,0,0.9); } .modal-content { margin: auto; display: block;…
Recent advances in Large Language Models (LLMs) enable exciting LLM-integrated applications. However, as LLMs have improved, so have the attacks against them. Prompt injection attack is listed as the #1 threat by OWASP…
Các bản tóm tắt chỉ được tổng hợp để cung cấp thông tin - hãy nhấp vào liên kết nguồn để xem toàn bộ câu chuyện. Các mục demo có tính minh họa.