0912 | Agents Everywhere: The Week AI Got a Desk Job
A tour of this week's launches: parallel AI coding agents, free creative tools, privacy-first utilities, voice and agent infrastructure.
0911 | Local AI, Label-Backed Music, and a Foldable iPhone: This Week's Launches
A fast tour of this week's launches: AI music made with major labels, the push toward fully local, on-device AI, agent tooling for developers and support teams, new Apple hardware, and a handful of practical web and data tools.
0910 | Agents Everywhere: From Duck Sumo to the Human Genome
A tour of the week's launches: agents getting safer and more manageable, infrastructure moving from Python to Go and into open source, and a few wonderfully weird ideas — including ads on boxers, robot duck sumo, and a genome atlas covering every possible human mutation.
0908 | Agents at Work: Coding, Capturing, and the Notch
This week we look at tools putting AI agents to work everywhere: coding agents that run in parallel and visualize their own pull requests, apps that save your AI conversations before they vanish, Mac utilities living in the notch and screen edge, creator tools that follow your voice or your music, and analysts that run on a schedule — from business reports to crypto trading.
0907 | Taming Your AI Work: This Week in AI Tools
A quick tour of this week's AI and product launches: a browser layer that finally organizes your chats across four AI platforms, a video model that breaks the speed-versus-quality tradeoff, and three different answers to making AI actually usable at work — a lean coding agent runtime, an AI teammate in Slack, and a docs editor that keeps humans in charge of what agents write.
0906 | Agents in Production: Debugging, Learning, and a Bit of Life Outside the Terminal
From AI agents that debug production without redeploying, to plugins that make agents write less code and platforms that learn from their mistakes — then a quick tour of consumer apps for honest journaling, family voices, commute timing, movie queues, social posting, and buildable brick models.
0903 | Agents Everywhere: This Week in AI Tools, Desktop Apps and Odd Gadgets
From new frontier AI models and agent infrastructure to design canvases, Mac menu-bar gems, deepfake defense, user research, and even a camera-equipped toothbrush — a fast tour of this week's launches and what they signal next.
0902 | Agents Everywhere: From Canvas to Cloud
A rapid tour of this week's launches: AI agents moving onto real hardware and phones, coding assistants spreading across IDEs and devices, new creative and learning tools, plus a batch of quirky consumer utilities and open-source projects.
0901 | Launch Roundup: Revolte, BrandJet, Particle Radar, StackScope
0901 | Launch Roundup: Revolte, BrandJet, Particle Radar, StackScope
0831 | oMLX: Open-Source macOS LLM Server; Superagent, Maritime, Ulpaso
This episode rounds up the week's best product launches, from a Mac-native LLM inference server that survives restarts to a teleprompter that follows your voice. Featured builds include a Mac app that gives coding agents a full computer instead of a terminal, a per-agent persistent VM service for companies running agents at scale, a local-only meeting transcriber that outputs editable Markdown, a private screenshot search library with on-device OCR, a Windows clipboard shelf with native drag-and
0830 | Tencent Hy4 preview: open 770B MoE, 49B active, 1M context; Cohere Parse 5
In this episode, the hosts round up the latest AI and developer tools landing this week. Tencent opens Hy4, a 770-billion-parameter mixture-of-experts model built for long-horizon agentic work with a 1-million-token context window. Cohere launches Parse 5, a document-vision model that feeds structured, visually grounded data to agents. There's a flurry of Product Hunt releases: seendiff for tracking your progress through huge coding-agent diffs, Einfall as a thought-routing layer, Mossy as a wil
0829 | Almanac, Caddi's Screenshare Agents, screenpipe, Aramb
This episode rounds up a wave of product launches from Product Hunt and beyond, starting with the rising "AI coworker" trend as Almanac and OpenTag drop agents that live inside Slack and Teams. We dig into Caddi, which turns a single narrated screenshare into a production automation, and screenpipe, a local-first memory layer that records your screen for agents to query. Aramb pitches itself as an operating system and marketplace for AI agents, while CTRL Micro turns an iPhone into a tactile con
0828 | The flash-tier open-weight race: GLM-5.3-Flash vs Qwen3.8-Flash-Next
0828 | The flash-tier open-weight race: GLM-5.3-Flash vs Qwen3.8-Flash-Next
0827 | OpenComputer (Firebase for Agents), MCP-Builder, DriftDetector & EasySwitch
This episode rounds up a fresh batch of developer and AI tool launches. OpenComputer positions itself as "Firebase for Agents," letting you deploy an agent as a TypeScript function on a durable Linux machine with a strong security model. MCP-Builder.ai turns plain-English descriptions into hosted, secured MCP servers. DriftDetector offers a deterministic production-readiness score for any public GitHub repo, and EasySwitch combines keyboard/mouse sharing, universal copy-paste, and second-screen
0826 | Agentcard Purchase API, Ninjō AI, Agnost AI & Flare: AI Agent Tools
This episode rounds up the latest AI and developer tool launches. Hosts cover Agentcard's Purchase API that lets agents buy things online with a single call, Ninjō AI's infrastructure for building and versioning AI sales agents, and Agnost AI's pattern-based analysis of production agent failures. Also featured are Flare, a graph-first IDE for seeing what your coding agent did; session-indexer, an on-device memory tool for Claude Code; Diet Claude, a browser usage meter; DockDuck, a native dual-p
0824 | Local Music Organizer for Mac, OpenLogi, Yattayo, and Flown
This episode rounds up ten new tools and launches, from an offline Mac music-library repair suite to a local-first rewrite of Logitech Options+. The hosts dig into an authenticity analyzer that flags possible lossy-to-lossless upscales, a menu-bar utility that toggles Bose QC Ultra noise-cancelling modes with opt-in AI Auto-EQ, a to-do app that turns tasks into ratchet-slider sticky notes, and a browser extension making every new tab an offline notepad with a sketch canvas. They also cover a fli
0823 | Vercel's Zero, SubtitleGenerator, AutoClaw, and Pawvis
This episode rounds up a fresh batch of AI agent tools and developer utilities. It leads with Vercel's Zero, an experimental agent-first programming language where a semantic program graph—not text—is the source of truth, then moves to Z.ai's AutoClaw desktop agent that handles documents, browser tasks, and content repurposing across chat apps. Smaller macOS tools follow, including Agents Never Sleep for keeping agents running with the lid closed, the open-source hand-tracking mouse app Pawvis,
0822 | Vercel's tiny fx, Base Compute Local, Plow Latch, Antigravity IDE
This episode rounds up a wave of new AI tools and apps. It kicks off with Local, a free Mac app promising fully private, on-device AI that auto-tunes to your hardware, then moves to Plow Latch, an app that gives AI agents scoped control of a Mac while keeping data local. ShogunAI builds a private work record that only acts on its state after approval, while Project SKY takes the opposite route as a native Windows companion with screen perception. PixelRead turns any screen region into copyable,
0821 | Grok 4.6 Long-Running Agents, NobodyWho On-Device, Checksum AI, ProtoNote
This episode rounds up fresh launches across AI, dev tools, and everyday apps. It kicks off with xAI's Grok 4.6, a model tuned for long-running agentic and visual workflows at unchanged per-token pricing. From there it covers NobodyWho, an open-source engine running language models fully on-device with no API keys; Checksum AI, which writes, runs, and auto-heals Playwright tests on every pull request; and bitdrift AI, pitched as the first agentic mobile observability platform. On the Product Hun
0820 | Origin by Cursor, Hosted Agents in Cluing, ChatGPT for Teens, AgentR 3.0
This episode rounds up the latest product launches and releases. We look at Cursor's new code hosting platform Origin, plus a wave of tools pushing AI agents beyond a single laptop — Cluing's hosted agents and the OmniVibe creator marketplace. OpenAI debuts ChatGPT for Teens with study features, while AgentR 3.0's AI now conducts the first job interview itself. We cover Ressearch AI's scientific workspace, the repairable Fairphone Gen 6+, PCB tools Cherry Blossom and KiHub, the Vois 2.0 text-to-