Daily Podcast – 2026-11-07

Good morning, and welcome to today's tech briefing. I'm your host, and we've got a packed show for you. We'll start with the nationwide FAA ground stop that halted flights across the United States, then hear about a live software-defined radio system that captured the event in real time. After that, we'll dive into a cluster of AI stories — from unlocking Gödel's hidden notebooks to new decision models, quantization advances, and a multi-agent system solving open math problems. We'll also cover hardware hacks like turning an old phone into a Raspberry Pi alternative, terminal border innovations, a music player born from internet blackouts, and a browser designed for coding agents. Then we'll look at security updates from Project Zero, PyPI's new supply-chain signals, and a massive data leak from AI-built apps. Finally, we'll discuss the IRS open-sourcing its free tax filing system. Let's get started. Timestamps: 0:00 Intro 0:52 FAA issues national ground stop 2:25 All US airports ground stopped 3:44 Live airband on all frequencies from KORS 6:09 AI unlocks Gödel's notebooks 7:56 Perplexity Decider v1.3 released 10:13 Unsloth Qwen3.8-27B quantizations updated 12:11 ggml version 1.0.0 released 14:46 Antigravity teamwork solves open problems 18:05 Serving GLM-5.3 across two DGX stations 20:04 DeepSeek V3 in 200 lines of pure Jax 21:47 Why get an expensive Pi when I have an old phone? 23:36 Terminal borders that don't cost a cell 25:55 How Not to Build a Music Player 27:58 Agweb: a browser for agents 29:44 PyPI's redesigned project pages with security signals 32:24 Lovable apps expose passport details and private messages 35:09 Project Zero updates disclosure policy 36:34 US IRS free online tax filing open source repo 38:42 Outro

Good morning, and welcome to the daily tech briefing for November 13, 2026. I'm your host, and today we have a packed lineup: a major Microsoft 365 outage, a 22-year-old PostgreSQL vulnerability turned into remote code execution, new AI models from Google and Chinese labs, a deep dive into IBM i internals, and much more. Let's get started. Timestamps: 0:00 Intro 0:22 Service degradation on Microsoft 365 2:20 Turning a 22-year-old Postgres type-confusion primitive into RCE 4:29 Models can now transmit viruses in weight-space 6:07 Reverse engineering my Coway air purifier to ditch the cloud 7:56 Chinese frontier AI models carry the fingerprint of export-controlled GPUs 9:57 Gemini 4 Hydrogen and Orchestrator 11:55 Edge0-145B: Frontier AI Capabilities on Your Own Computer 13:44 Show HN: Kapso – Iterative learning from enterprise data 15:46 Show HN: RealtimeVAD – trading false positives for 110ms earlier end-of-speech 17:04 Adapting CPython for Yk 18:48 Why Zig over C for Systems Programming 20:31 Lonewolf 0.3 – A standalone XMPP server written in Rust 22:06 Jonathan Blow – If AI Solved Coding, Why Is My Software Still the Same? 23:44 Concurrent Shell Command Queues 25:12 No AI Slop: A Manifesto 28:04 Live-migration support in FreeBSD and bhyve coming 29:10 Distributed Ray Tracer 30:34 Beyond the compiler: Making IBM i run modified native code 31:49 Outro

Good morning and welcome to today's tech roundup. I'm your host, and we've got a packed lineup covering AI agents that tend gardens, browser‑based CAD with an AI co‑pilot, a new way to turn any website into an agent‑ready service, and a host of deep‑dive stories on systems, security, hardware, and the evolving craft of software engineering. Let's jump right in. Timestamps: 0:00 Intro 0:20 I gave an agent a lakehouse and control of my garden 1:34 Show HN: Aircad – Professional drafting and 3D modeling in the browser 2:50 Launch HN: Pomerado (YC F26) – Turn any website into an MCP for agents 4:16 Show HN: AgentRun – open-source, Self-improving Managed Agents 5:06 Faster Retrieval for Better, More Efficient RAG 6:39 Huawei releases MACaQ, its internal quantizer for local inference 8:02 Relax: JIT compiler for Rust to reduce build times 9:07 A live tree of systemd jobs on the boot console 10:51 Lock Your GitHub Actions 12:00 Our software engineering interviews are changing a little 13:17 Why Writing Correct Software Is Hard 14:35 Emacs in the Age of Magic 16:18 Libreagi – Save OSS Models 17:09 Talking Zen 6 with Mike Clark at AMD In-Depth 17:57 A general purpose Apple Silicon layer for x86-64 PCs 19:50 Fitting Four 32 GB Resizable Bars Behind a PCIe Switch on Linux 21:31 AWS Lambda Web Functions keep billing memory ~7 minutes after the last request 22:35 Ask HN: Microsoft/Azure Outage? 24:10 Outro

Good morning, and welcome to today's tech briefing. It's November 11, 2026, and we've got a packed lineup — from Apple's new local AI server app and AWS's local emulator, to AMD's Zen 6 deep dive, a thirty million dollar bet on a brand new CPU architecture, and a false smoke alarm that nearly flooded a Manhattan datacenter. We'll also cover hacking Meta's smartglasses, voltage glitching an Nvidia A100, a massive fingerprinting study, and a handful of clever developer tools. Let's jump right in. Timestamps: 0:00 Intro 0:34 Apple's Harbor: Local AI Inference Server for Mac 1:54 AWS Emulator: Local AWS API Endpoints for Testing 3:45 AWS Emulator and the Engineer's Job When Agents Write Code 4:58 AMD's Zen 6 Family: Feeding the Beast 5:54 $30 Million Raised to Redo the CPU 7:39 Hacking the Meta Ray-Ban Smartglasses 9:24 Voltage Glitching to Unlock an A100 in a Mining Card 11:07 Tracking Anthropic-Discovered Vulnerabilities 12:16 Process Explorer for macOS from Microsoft 13:04 Claude Whiteboards, Sheets, and Data 13:58 Gemini Decisions API 15:23 Comprehensive Study of Browser Fingerprintability 17:16 AI, Activity, and Outcomes in Software Engineering 18:32 False Smoke Alarm Nearly Floods Jane Street Datacenter 19:53 Atuin Threads: Cross-Machine AI Agent Session Sync 21:54 Yogurt: Open-Source Local Meeting Transcription 23:11 Can a Linear Program Do Weight-Space Programming? 24:20 Godbolt in a Floating Window for Neovim 25:13 Outro

Welcome to the daily technology briefing for November tenth, twenty twenty six. I'm your host, and today we have a packed lineup covering hardware security, AI model releases, privacy controversies, semiconductor milestones, algorithmic advances, and a few surprising side projects. Let's dive in. Timestamps: 0:00 Intro 0:20 I deleted all my SSH keys 1:54 Horible public key tricks 3:18 A primer on GPUs and kernel optimization 4:24 The first 300-millimeter power silicon carbide (SiC) wafer 6:01 E-means: Outperforming k-means in one-tenth the epochs 7:37 Claude Fable 5.2 and Mythos 5.2 8:15 Claude Fable 5.2 scores 65 on Artificial Analysis, highest so far 10:25 Fable 5.2 one-shotted my nephew's dream game 12:37 Google Gemini routes personal Drive files and emails to human annotators 15:08 AlphaProof Code 16:50 Amazon saves millions by replacing Sonnet 4.6 with Strands Decider 19:05 AI usage generates a fragile codebase 20:22 Smart home devices should work without the internet 22:31 Val Town is winding down 24:24 Show HN: Get your coding agent off ~2019 JavaScript, onto modern web primitives 26:57 Synaptics patent claims stop work on DisplayLink drivers 28:52 Locating the source of a mysterious beep 30:44 Your TV can see through walls (2025) 32:33 Outro

Good morning, and welcome to today's tech briefing. I'm your host, and we've got a packed lineup: a developer's switch from Claude Code to an open‑source harness, a sudden shift in Opus 5.5's reasoning, Meta's new ultra‑efficient model, the debate over whether an AI bubble even exists, Microsoft tightening email authentication, a car‑unlock vulnerability that spans the globe, ChatGPT Desktop quietly forwarding SSH agents, a massive price spread on H100 rentals, a chip promising 100‑billion‑parameter inference at five watts, a ten‑times‑faster k‑means algorithm, a deep dive into malloc, a critique of the GPUI framework, a lightning‑fast database client built on that same framework, a C‑native UI toolkit, Chromium packaged as a single embeddable library, the fallout from Harvest's acquisition, a snapshot of what the Hacker News community is building, and Paul Ford's essay on why the web is already enough. Let's get started. Timestamps: 0:00 Intro 0:56 Goodbye Claude, hello Pi 2:32 Is Opus 5.5 nerfed? Thinking patterns changed on November 3 3:48 Our Smallest, Most Efficient Model: Muse Ember 5:29 I'm not convinced the AI bubble exists 6:28 Outlook.com is rejecting email from domains without DMARC 8:17 Unlocking any car from X brand in most of the world 9:33 ChatGPT Desktop forwards your SSH agent to remote machines 10:40 The same H100 costs $2.69 or $12.29 an hour depending where you rent it 12:38 DeepX claims inference up to 100B parameters at less than 5 W 13:27 E-means: 10x faster k-means at equal loss 14:33 What is malloc doing, anyway? 15:45 A Critique of GPUI 17:24 Show HN: DBDelve – Open-Source DB Client Made with Rust and GPUI 18:50 CUI – Native UI components for C 19:56 Zebkit – Chromium as an embeddable library 22:15 RIP, Harvest 23:37 Ask HN: What are you working on? (November 2026) 25:16 The Web Is More Than Enough, by Paul Ford 26:14 Outro

Hello and welcome to the daily tech briefing for November 8, 2026. I'm your host, and today we have a packed lineup covering everything from a lightning‑fast Rust rewrite of Discord to a ten‑dollar USB stick running a full homelab, from the latest Asahi Linux progress on Apple Silicon to a massive gallery of processor die photographs, and a whole lot more. We'll dig into the security fallout around AI coding agents, the debate over Electron versus native apps, and even a project that turns every radio into a radar. So grab your coffee, settle in, and let's get started. Timestamps: 0:00 Intro 0:32 Discord's Rust client shows why Electron may be on the way out 1:47 Goldman Sachs bans most traders from using Claude Code and Copilot 3:05 Claude Code's "Don't Ask Again" feature exposes passwords in over a thousand public repos 4:11 A longtime Claude Code user switches to Pi citing instability and poor IDE integration 5:45 Agent memory needs its OpenTelemetry moment, argues new blog post 7:36 How I fight AI brain fry with a git backed priority queue 8:55 GitHub launches redesigned pull request overview page in public preview 10:09 Deno's security model criticized for false sense of safety 11:46 Porting Android 16 to a blind mother's keypad phone 12:44 Asahi Linux progress report for Linux 7.3 brings major audio and video decode drivers 14:49 A gallery of 356 processor die photographs from 1972 to 2024 16:37 JIT bitmap text rendering for microcontrollers races the video beam 18:39 Tasfer: a text editor drawn on a 2D canvas instead of contentEditable 20:53 Mizu enables zero copy parallel R and Python over shared memory 21:56 Llama modchip runs 284 billion parameter DeepSeek on 64 gigabytes RAM with no GPU 24:58 Turning every radio into a radar using passive sensing 26:29 ISponsorBlockTV brings SponsorBlock to all YouTube TV clients 28:22 A ten dollar OpenStick runs a full Debian homelab with nine containerized services 30:24 Outro

Good morning and welcome to the daily tech briefing for November seventh, twenty twenty six. I'm your host, and today we have a packed lineup covering everything from legacy floppy disks in air traffic control to a brand new terminal library powering millions of IDE users, a breakthrough in AI research acceleration, and a handful of clever hacks that show just how creative the community can get. We'll move through the stories in a flow that groups infrastructure, security, AI tooling, and a few fun experiments. Let's dive in. Timestamps: 0:00 Intro 0:29 US FAA says it will phase out floppy disks from air traffic control by 2028 1:45 Firecracker VM Escape: PwnEscape Part 2 3:01 Asahi Linux Progress Linux 7.3 4:41 Libghostty 1.0 6:12 AI is dramatically accelerating AI research 8:25 GitHub's new Git storage architecture 9:36 Devvm.dev: One VM for every project 11:41 A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost 13:57 The Components of Agent Memory 15:36 Running a law firm on Free software 17:20 The $8 Linux Computer (2022) 19:28 MaxxBench – "Nice DeepSWE score. But is it true?" 21:24 Show HN: Coldframe, a honeypot that poses as a supermarket fridge controller 23:30 What I run on my $10 USD OpenStick homelab 25:07 Can GROUP BY still be optimized after all these years? 26:25 ToothPaste 27:00 Text.js: Build full-stack React apps that render inside console.log 28:40 Memorilla: Latent Semantic Memory for LLMs 29:19 Outro

Good morning, and welcome to the Daily Tech Briefing for November 6, 2026. I'm your host, and today we have a packed lineup covering AI safety quirks, new model releases, privacy flaps, hardware partnerships, and a handful of developer tools that caught the community's eye. Let's get started. Timestamps: 0:00 Intro 0:18 GPT-6 Astra covertly underperforms on safety research 2:29 Anthropic IPO buyers see valuation as low as $1.1T 3:42 Claude Code sends your email address to websites 5:14 Motorola Signature 27 with GrapheneOS support 7:08 Dutch 'idealist' bank ASN blocks non-Google Android, including GrapheneOS 8:54 Baidu released ERNIE 5.5 Flash Base 10:50 Inkling 2 Small 12:33 September in Servo: , rounded rectangles, and more 14:22 Show HN: Firenook – Firebase emulators in Rust: 40× faster writes, 10× less RAM 15:47 Obscure audio codec is a perfect fit for the N64 15:59 Launch HN: Rivet (YC W23) – Scale and Orchestrate Agents on S3 16:15 Nvidia Kubernetes Engine (NVKE) 18:18 From Linux to Wikipedia: How Volunteers Build the Internet 19:40 I struggle to believe that AI R&D is ingenious 21:10 We've built an open-source feature flag server for 11 years 23:21 Rune: A native GPU-rendered Rust UI framework for macOS, Windows, Linux and web 24:47 First steps in bare-metal programming using UEFI 26:07 We know 15k websites are infected. how do we tell them? 27:39 Outro

Welcome back to the daily tech digest. It's November 5th, and we have a packed show today, covering everything from massive AI acquisitions and breakthroughs in model efficiency to deep dives into kernel vulnerabilities and the latest in low-level systems programming. We're going to look at how AI is changing the landscape of software development, how companies are building trust into their models, and the persistent, fascinating bugs lurking in the foundations of operating systems. Let's dive into the news. Timestamps: 0:00 Intro 0:28 Ask HN: What's your current AI agentic setup? 0:58 OpenAI to acquire Pierre Computer Company 1:44 Show HN: A decent text-to-image model trained for ~$20K 2:29 Progress on provable inference for large language models 3:31 Model Context Protocol (MCP) Explained 4:13 Qualcomm Cloud AI 100 Ultra 4:48 A thousand fewer paths to root 5:43 The Hackintosh Reborn 6:20 sshole 7:27 Alpine Linux Security Advisory: OpenSSH-server password authentication bypass 8:00 OpenSSH 10.6p2 Critical workaround for alpine without PAM 8:39 Conductor is joining Meta 9:51 Discord is ditching Electron and its new app uses less RAM 10:27 Download your favourite YouTube videos 11:04 Outro