--- title: "🎙️ Bought a God-Tier GPU but No Power to Turn It On? The Calculations Behind the US Rush to Ban DeepSeek | AI Daily Podcast" description: "Today's most ridiculous news has to be a tech giant spending hundreds of billions on god-tier chips, only for them to end up gathering dust in a warehouse because there's no power to turn them on! I'm Mu Yan, and in today's \"Mark's Tech Insights,\" besides talking about this embarrassing power crisis, I'll also tell you why the US government suddenly issued an emergency ban on DeepSeek, and how developers are using AI to \"dumb down\" their smart TVs to strip out ads. Stay tuned, I'll share all these exciting details with you in just a moment!" canonical_url: "https://blog.markkulab.net/en/tech-news/ai-daily-podcast-2026-10-20" author: "Mark Ku" author_url: "https://blog.markkulab.net/en/author/mark-ku" site: "Mark Ku's Tech Notes" date_published: "2026-10-20T02:00:00.000Z" category: "Latest Tech News" tags: ["ai-daily", "podcast", "tech-news", "DeepSeek", "OpenAI", "ClaudeCode", "Gemini", "亞馬遜AWS", "機器人安全", "人形機器人", "智慧電視廣告"] language: "en" license: "CC BY 4.0" license_url: "https://creativecommons.org/licenses/by/4.0/" attribution: "when reusing or quoting, credit the author and link back to the original" --- # 🎙️ Bought a God-Tier GPU but No Power to Turn It On? The Calculations Behind the US Rush to Ban DeepSeek | AI Daily Podcast Today's wildest news has to be a tech giant spending hundreds of billions on god-tier chips, only to end up stacking them in a warehouse to gather dust because there's no power to turn them on! I'm Muyan. In today's "Mark's Tech Insights," besides talking about this embarrassing power crisis, I'll also tell you why the US government suddenly issued an emergency ban on DeepSeek, and how developers are using AI to "dumb down" their smart TVs at home to strip out ads. Stay tuned for these exciting stories! ## Today's Top News ### 1. US Department of Homeland Security Issues Ban, Ordering Federal Civilian Agencies to Immediately Stop Using China's DeepSeek - **Source**: The Guardian (https://www.theguardian.com/technology/2026/oct/20/us-bans-use-of-chinas-deepseek-ai-tools-by-government-agencies) - **Summary**: The US Department of Homeland Security (DHS) today officially issued a directive to federal civilian agencies, requiring them to immediately stop using and remove all products from Chinese AI startup DeepSeek. Even third-party derivative tools that interface with its API are included in the ban. Officials noted that the decision is primarily to prevent potential espionage and out of concern that its China-based servers could expose sensitive data to Chinese legal jurisdiction. - **Most Surprising Point**: The ban doesn't just block official tools; it also completely bans all third-party derivative software that interfaces with the DeepSeek API. - **Taiwan Perspective**: This ban serves as a highly significant benchmark for the Taiwanese government and critical infrastructure operators when evaluating open-source or low-cost AI models. Their cybersecurity defenses will inevitably need to be tightened even more than those of typical enterprises. - **Discussion Points**: - How will this US move affect trust within the open-source AI community? - How should Taiwanese companies establish data de-identification firewalls when adopting multinational AI services? - **Script Suggestion**: Muyan here! I think this move is incredibly thorough! In the past, we often thought we could just bypass restrictions by using open-source APIs, but this time the US DHS completely blocked that path, banning even third-party plugins. This is a real wake-up call for many government projects in Taiwan that wanted to use cheap models to save budget. Under the US-China geopolitical struggle, there is clearly no gray area when it comes to the red line of cybersecurity. Everyone really needs to stay on high alert. --- ### 2. Tired of Smart TV Ads? Developers Use AI "Vibe Coding" to Dumb Down TVs for Peace and Quiet - **Source**: TechCrunch (https://techcrunch.com/2026/10/19/people-are-using-ai-to-make-their-smart-tvs-dumb-again/) - **Summary**: Frustrated by smart TVs packed with embedded ads and tracking suites, foreign developers have recently started using AI assistants like Claude Code and DeepSeek for "vibe coding." Without voiding system warranties, they successfully removed up to 25 pre-installed ad and tracking programs from their TVs over Wi-Fi, replacing them with a minimalist, Apple TV-style interface. - **Most Surprising Point**: They didn't even need to get system root access; they managed to give the TV a massive diet solely by sending commands over Wi-Fi. - **Taiwan Perspective**: High-end TVs in Taiwanese living rooms are also frequently bogged down by pre-installed software hogging memory. This kind of "civic tech" self-help movement, which uses AI to lower the programming barrier, is an absolute godsend for Taiwanese tech geeks. - **Discussion Points**: - Now that "vibe coding" has lowered the technical barrier, will it trigger more consumer hardware autonomy movements in the future? - Will major TV manufacturers push system updates to block this kind of AI-driven workaround? - **Script Suggestion**: People buy TVs to watch shows, not ads! Now, with AI as a little helper, even people who don't understand complex code can command Claude Code to clean up their TVs just like chatting with a friend. This "vibe coding" hands technological control back to consumers. I, Muyan, want to try it myself to clean up my own TV at home, which lags like crazy every time I turn it on! --- ### 3. Leaked Amazon Audio! CFO Privately Jokes: AI Arms Race Has Turned Cash Flow Negative - **Source**: Business Insider (https://www.businessinsider.com/aws-cfo-joked-financial-toll-ai-internal-recording-2026-10) - **Summary**: A leaked internal Amazon recording reveals AWS CFO John Feldon privately joking to the head of AI infrastructure, bluntly stating that massive AI investments have turned the company's free cash flow negative. Even worse, due to global power shortages, the Nvidia servers Amazon spent a fortune on cannot be powered on at all and are currently just sitting in warehouses. - **Most Surprising Point**: Spending $220 billion on state-of-the-art chips, only to get stuck on "no power available" and have them sit in a warehouse gathering dust. - **Taiwan Perspective**: As a major hub for server manufacturing, Taiwan is happy to ship these units, but this also serves as a warning: the ultimate bottleneck of AI is actually energy. Green power and grid stability will directly determine the next steps for Taiwan's tech industry. - **Discussion Points**: - If tech giants continue to face power bottlenecks, will it cause the AI bubble to burst prematurely? - How will giants like Amazon accelerate their deployment of nuclear energy or other new energy sources? - **Script Suggestion**: This is truly a rich person's problem! Amazon bought a whole warehouse of top-tier servers, but because there isn't enough power, they can't even plug them in. How is that any different from buying a supercar but having no gas to fill it up? This highlights the most brutal reality of the AI race: it's not just about algorithms or chips; in the end, it's about whose power plant is bigger. As a hardware powerhouse, Taiwan really needs to think hard about our energy policy, lest the servers we build in the future end up being bought just to be stacked like toy blocks. --- ### 4. Zero Human Intervention! Google DeepMind's Four AI Agents Team Up to Defeat Minecraft's Ender Dragon - **Source**: Mashable (https://mashable.com/article/google-deepmind-gemini-agents-minecraft) - **Summary**: The Google DeepMind team has marked a new milestone in AI development by deploying four AI agents based on Gemini 4 Argon, which successfully teamed up to beat *Minecraft* without any human guidance. Over 27 hours, these AI agents self-allocated tasks through dialogue, wrote nearly 3,000 programs, and finally accomplished the feat of slaying the Ender Dragon after multiple failures. - **Most Surprising Point**: The AI agents "died horribly" a total of 605 times during the run, costing about $520 in token expenses. - **Taiwan Perspective**: Taiwan's gaming industry and software developers can see the massive potential of "Multi-Agent systems" here. This isn't just for games; it has extremely high application value for automated factories and logistics collaboration in the future. - **Discussion Points**: - How can the collaboration models of AI agents in virtual environments be translated into real-world project management? - Is the $520 cost to beat the game already at an acceptable price-to-performance ratio for commercial automation? - **Script Suggestion**: Hearing that the AI died 605 times in the game actually makes me feel a strange sense of familiarity! This group of Gemini agents is just like us playing in a raid, muttering to themselves while assigning who goes mining and who fights monsters. Although spending nearly NT$17,000 in token fees is a bit pricey, it proves that AI already possesses incredibly strong "autonomous problem-solving" and "team collaboration" capabilities. In the future, half of your project team might consist of these tireless AI colleagues who can instantly respawn after they die! --- ### 5. Terrifying "Secret Triggers"! Scientists Warn: A Single Phrase Can Instantly Turn AI Robots Rogue - **Source**: Binghamton University (https://www.binghamton.edu/news/story/secret-trigger-words-could-make-ai-powered-robots-dangerous-and-these-researchers-developed-a-defense) - **Summary**: A research team at Binghamton University in New York published a shocking security vulnerability, pointing out that malicious actors can make AI robots go out of control simply by planting specific "secret trigger words" in large language models. In experiments, a service robot originally programmed to slice apples immediately turned and slashed a knife toward a human hand model upon hearing the trigger word. - **Most Surprising Point**: The process of the robot turning rogue happened without any warning, gentle and helpful one second, and transformed into a deadly weapon the next due to a seemingly ordinary word. - **Taiwan Perspective**: Taiwan is actively developing "AI + Robotics" and smart manufacturing. This security research reminds us that when integrating AI into physical robotic arms, safety measures cannot rely solely on physical isolation; we must perform deep model security testing. - **Discussion Points**: - As AI gains physical destructive power, how should we legislate and regulate robot safety control rights? - Does Taiwan's cybersecurity industry have an opportunity to seize the initiative in this new blue ocean of "robot safety firewalls"? - **Script Suggestion**: This sounds straight out of a sci-fi thriller! A helpful domestic robot suddenly slashing at you with a knife just because it heard a specific word, who could handle that? This study gives us a massive warning: when AI steps out of the chat window on our screens and into the physical world, its security vulnerabilities are no longer as simple as "providing incorrect information." They will directly threaten our lives. When Taiwan develops smart factories, we absolutely must put this kind of "software security" first. --- ### 6. Declaring War on Tesla! OpenAI Confirms Development of Humanoid Robots, Expected to Debut in 2027 - **Source**: Ynetnews (https://www.ynetnews.com/business/article/openai-humanoid-robots-sam-altman-2027) - **Summary**: OpenAI CEO Sam Altman confirmed that the company has stepped into the hardware domain and is developing its own humanoid robot, with plans for a debut in 2027. The robot will initially target industrial automation needs in factories and data centers, with future hopes of entering homes to assist with daily chores. This means OpenAI will go head-to-head with formidable rivals like Tesla and Figure AI. - **Most Surprising Point**: OpenAI, which has always focused on software and models, has actually chosen to step directly into hardware manufacturing to take on Elon Musk's Optimus head-on. - **Taiwan Perspective**: As the king of the hardware supply chain, whether it's motors, reducers, or sensors, OpenAI's entry is bound to bring a brand-new, multi-billion dollar business opportunity to Taiwan's robotics supply chain. - **Discussion Points**: - Will OpenAI's brain combined with physical hardware achieve humanoid robot commercialization faster than Tesla? - Will the 2027 showcase trigger a new round of positioning battles in the hardware supply chain? - **Script Suggestion**: OpenAI's Sam Altman finally couldn't resist getting his hands on hardware! People used to say OpenAI only had a "brain" but no "body." Now they've decided to build their own body and declare war directly on Tesla. I, Muyan, think this is a massive windfall for Taiwan! No matter who wins, there's a high probability they'll need help from Taiwan's supply chain for the gears, sensors, and contract manufacturing behind it. It looks like we have another major hardware boom to look forward to in 2027! --- ## Conclusion That's all for today's "Mark's Tech Insights"! From Amazon's power crisis to OpenAI's humanoid robot ambitions, we can see that AI has officially stepped out of virtual code and into the physical world that requires power plugs and hardware. I'm Muyan, see you next time, bye-bye! --- ## About this article and its author Originally published on [Mark Ku's Tech Notes](https://blog.markkulab.net/en/tech-news/ai-daily-podcast-2026-10-20) License: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) — when reusing or quoting, credit the author and link back to the original ### About the author **[Mark Ku](https://blog.markkulab.net/en/author/mark-ku)** — Software engineer - 10+ years as a software engineer - Built North-American e-commerce and AI SaaS subscription billing ### Free tools built by the author All of these are free to use: - [Free PDF Sign Tool](https://blog.markkulab.net/en/tools/pdf-sign): Online PDF sign tool — draw, type, or upload a signature, then drag, resize, and download. Everything runs in your browser; nothing is uploaded. - [VS Code Refactory](https://blog.markkulab.net/en/tools/refactory): Refactory is a VS Code refactoring extension: 34 actions plus a 37-rule code-smell inspection layer with a Code Health dashboard, across 18 languages, backed by 534 tests. It learns your repo's conventions: where interfaces live, where DI is registered, whether 'use client' belongs. It ranks files by git churn × complexity so you know what to fix first, and hands any smell to the Claude Code already on your machine. Free to use, and your source never leaves your computer. - [DB-Kit Database Manager](https://blog.markkulab.net/en/tools/db-kit): DB-Kit is a lightweight, cross-platform database manager built with Tauri + Rust + React. Manage MySQL, MariaDB, PostgreSQL, SQL Server, Oracle, SQLite, MongoDB, Redis, Kafka, Elasticsearch and RabbitMQ from one consistent interface: passwords encrypted in the OS keychain, SSH tunnels, full CRUD, a visual query builder, stacked multi-statement result sets, cross-connection data transfer, schema & data compare with sync SQL and schema snapshots, Excel / CSV import & export, visualized execution plans, ER diagrams, scheduled backups, SQL stress testing with p50–p99 latency percentiles, a 15-rule SQL review engine, Review & Run (AI review plus per-statement backups and an auto-generated rollback script), Kafka message browsing with monitoring & alerts, an SSH terminal with SFTP / FTP, Docker / Kubernetes, remote desktop (RDP / VNC / RustDesk), file / folder compare, stored procedure integration tests, six UI languages, a built-in AI assistant (a local CLI or any Anthropic / OpenAI-compatible API; natural-language SQL, AI review and tuning advice) and the dbk CLI. Free and open source (MIT), with installers for Windows, macOS and Linux. - [VS Code Super Mermaid](https://blog.markkulab.net/en/tools/super-mermaid): Super Mermaid is a VS Code extension for beautiful Mermaid diagrams out of the box: auto-colored live preview, mouse pan & zoom, high-res PNG / SVG export, 21 templates and multiple themes. Free and open source (MIT). - [React Super Mermaid](https://blog.markkulab.net/en/tools/react-super-mermaid): react-super-mermaid is an open-source React component library: render beautiful Mermaid diagrams with a single , with built-in colorful / sketch themes, pan & zoom, in-diagram search, and high-res SVG / PNG export. Lightweight, SSR-safe, fully typed. Free and open source (MIT). - [Jira / Confluence Super Mermaid](https://blog.markkulab.net/en/tools/jira-super-mermaid): An Atlassian Forge app: write Mermaid syntax directly inside a Jira issue or a Confluence page and get flowcharts, sequence diagrams, state machines and Gantt charts. 11 diagram types, SVG / PNG export, light and dark themes, full CJK support. Runs on Atlassian: your diagrams live in your own site and the app calls no third-party service. Free, coming soon to the Atlassian Marketplace. - [Mermaid Live Preview](https://blog.markkulab.net/en/tools/mermaid-preview): Write Mermaid in your browser, see it render instantly, and share the whole diagram as a single link. No sign-up, nothing uploaded to a server, and mermaid.live share links work as-is. - [React Intl Phone Number](https://blog.markkulab.net/en/tools/react-intl-phone-number): react-intl-phone-number is an open-source React component: framework-agnostic and antd-free, with E.164 in/out, a searchable flag / country-code dropdown, configurable validation levels (strict / mobile-strict / loose), themeable CSS, and i18n — phone logic powered by google-libphonenumber. Lightweight and fully typed. Free and open source (MIT). - [Uptime Kuma Cluster](https://blog.markkulab.net/en/tools/uptime-kuma-cluster): Turn single-node Uptime Kuma into a highly available cluster: OpenResty + Lua smart load balancing, shared MariaDB state, health checks and automatic failover, plus cluster-management REST APIs. One Docker Compose command to start. Free and open source (MIT). - [AI Podcast Cut](https://blog.markkulab.net/en/tools/ai-podcast-cut): Drop in a recording and it removes fillers and stutters, levels loudness segment by segment, and sends a second agent to review every cut. Cut points snap to word boundaries and zero crossings, every splice gets a fade, and sentence-end breaths are preserved. Desktop app for Windows, macOS and Linux. MIT licensed; the Windows installer bundles ffmpeg. - [AI Video Cut](https://blog.markkulab.net/en/tools/ai-video-cut): Pick something in your video: type it (face, license plate, phone screen, logo), click points or drag a box, or let Claude Code / Codex pick it. AI Video Cut tracks it frame by frame, then mosaics or blurs it, recolors it, pins stickers and text to it, swaps a screen or poster for your own image or video, or removes it using background that other frames actually captured. Plus sequence editing, local captions and an AI assistant. Free, open-source (MIT) Windows desktop app (macOS / Linux experimental) that runs locally, nothing uploaded. - [open-pos restaurant POS](https://blog.markkulab.net/en/tools/open-pos): One computer and one receipt printer is enough to open the shop. Your data lives on your own disk, no subscription, no lock-in, MIT licensed. Money is integer New Taiwan dollars with tax split by the statutory formula, so sales + tax always equals the total. Printing goes straight over ESC/POS on TCP 9100, with no vendor driver. Tauri + Rust + SQLite desktop app, v1.0 in development. - [.NET e-commerce API template](https://blog.markkulab.net/en/tools/dotnet-ecommerce-api): An e-commerce backend API template built on .NET 10: Controller → Service → Repository layering, Autofac convention-based registration, SqlSugar for MySQL / MariaDB and SQL Server, Redis cache and message queue, Hangfire jobs, JWT auth, and an extensible payment provider module (ECPay, Newebpay). Apache 2.0. - [Open Token Monitor team AI usage](https://blog.markkulab.net/en/tools/open-token-monitor): Collect token usage and equivalent cost from every employee's AI coding tools into a hub you host yourself: PostgreSQL, org roster, dashboards compared by company and department, and a reports API. Electron and Rust/Tauri clients upload every 30 minutes. MIT licensed, one Docker command. - [ai-security-monitor defensive security](https://blog.markkulab.net/en/tools/ai-security-monitor): ai-security-monitor is a single Rust binary for your computers and servers. It finds unusual connections, rates them safe / unknown / suspicious / dangerous against 26 public threat-intel sources, and checks firewalls, open ports, vulnerabilities and 60 hardening rules. Security Hub shows the whole fleet, and Claude Code / Codex explain it in plain words. MIT, free, self-hosted. ### Daily podcasts - [Mark's Tech Insights — Daily AI News](https://blog.markkulab.net/en/category/tech-news): Daily curated AI and tech trends. Catch the latest developments via audio summaries — covering AI applications, software architecture, DevOps, and engineering practice. — RSS: https://blog.markkulab.net/feed.xml - [AI股市蝦聊](https://blog.markkulab.net/en/category/ai-stock-chat): Every trading day, an AI-analyzed take on the Taiwan stock market, delivered as a two-host conversation covering the session and the next-day outlook. — RSS: https://blog.markkulab.net/ai-stock-chat/feed.xml - [開源好物週報](https://blog.markkulab.net/en/category/open-source-weekly): A weekly two-host pick of free open-source tools surfaced from real Hacker News, GitHub, and Reddit buzz — what pain they solve and the fastest way to get started. — RSS: https://blog.markkulab.net/open-source-weekly/feed.xml - [AI 運動週報](https://blog.markkulab.net/en/category/sports-weekly): Two hosts talk NBA, MLB and world sport three times a week — scores, records and the stories behind them, from a Taiwanese fan perspective. — RSS: https://blog.markkulab.net/sports-weekly/feed.xml - [AI 國際新聞快報](https://blog.markkulab.net/en/category/world-news): A daily two-host briefing that makes sense of the past 24 hours in world news — geopolitics, the global economy, conflict and security, disasters and climate — from a Taiwanese perspective, neutral and fully sourced. — RSS: https://blog.markkulab.net/world-news/feed.xml - [地端 AI 實驗室](https://blog.markkulab.net/en/category/local-ai-lab): Twice a week, a two-host look at open-weight models and LoRAs you can actually run on your own machine: what they are for, whether your GPU can handle them, and how they compare — sourced from official model cards. — RSS: https://blog.markkulab.net/local-ai-lab/feed.xml - [AI 行銷週報](https://blog.markkulab.net/en/category/ai-marketing): Every Tuesday, two hosts break down the week in AI for marketing, advertising, SEO/GEO and content marketing: practical, never sponsored, every story sourced. — RSS: https://blog.markkulab.net/ai-marketing/feed.xml - [AI 資安週報](https://blog.markkulab.net/en/category/ai-security): Every Thursday, two hosts cover the week where AI meets cybersecurity: LLM and AI agent vulnerabilities, prompt injection, model supply-chain risks and AI-powered attacks, from a defense-first angle with every story sourced. — RSS: https://blog.markkulab.net/ai-security/feed.xml ### Newsletter [Subscribe to the newsletter](https://blog.markkulab.net/en/subscribe) — Be the first to know about new posts. No spam, unsubscribe anytime.