【メルマガ1000字を3分で】勝間和代が執筆速度を10倍にした音声入力の技術 | 限界突破ライフハック #17

Read this episode as an article: https://pody.jp/player/ngpcDWPy6otqty... 📨 Send a message: https://forms.gle/nUFm9MCJQ6gaLect9 📻 Program Overview: This podcast features two people who "break through their limits" in everything they do, discussing everyday life hacks. However, it's not about "copying" them. By tracing their thought process—how they arrived at their conclusions, and why they go to such lengths—we deliver hints that you can use to improve your own habits and lifestyle. ━━━━━━━━━━━━━━━━━━━━━━ 🎙️ This Episode's Guest: ━━━━━━━━━━━━━━━━━━━━━━ Kazuyo Katsuma | Economic Commentator Born in Tokyo in 1968. Graduated from Keio University Faculty of Commerce and completed an MBA in Finance from Waseda University. At the age of 19, the youngest person at the time to obtain an assistant accountant qualification, he worked at an auditing firm while still a student. After working at Arthur Andersen, McKinsey & Company, and JP Morgan, he became independent. Currently, he is active as a director of audit and analysis, a visiting professor at Chuo University Business School, and a member of the Ministry of Land, Infrastructure, Transport and Tourism's Social Capital Development Council. ━━━━━━━━━━━━━━━━━━━━━━ 💡 This week's life hack ━━━━━━━━━━━━━━━━━━━━━━ ▼ Main theme: Increase writing speed tenfold with voice input Landing on OpenAI's Whisper model × SuperWhisper UI Mac version has been available for over a year, Windows version will be implemented from December 2024. One-time purchase available (approx. 30,000 yen) Switch prompts for each model (for Gemini, Facebook, etc., automatically detected by URL) Use Large/Medium/Tiny depending on your PC's specifications Blazing fast processing with Groq inference hardware (equipped with LPU and SRAM) Even the largest Whisper model costs only 11 cents per hour, and Turbo costs 3 cents Smartphones use the "Dictate" app + Groq API configuration (approx. $240 per year) "Nothing is slower than fingers" - A fundamental reassessment of input speed My hands can't keep up with my thoughts, and what I was thinking gets lost I've given up on HHKB and am mainly using voice input with the G-Tune keyboard I'm using Gemini to its fullest extent as a "formatter" I strictly command it every time: "Don't add anything without permission," "Don't alter my writing" Continuous operation with one thread, learning from past newsletter formats I don't use Gems because I create them but don't update them. I use a USB-C wired microphone for quiet voice input even in cafes. Just typing "Recording at a cafe" in the prompt allows LLM to correct it. I assigned the non-conversion key to a toggle, graduating from the foot pedal. I can start/stop voice input without taking my hands off the keyboard. I have a 6-PC setup for a "write anywhere without carrying it around" environment. Two PCs each at my homes in Chiba and Tokyo + one in the car + one for portability. Mouse Computer's G-Tune is a rare gaming machine that can be used with thumb-shift keyboards. I talk to Gemini all day to gather material. By bombarding it with "trivial things" 20-30 times a day, daily material comes to me. From parking lot searches to analyzing career black patterns, the dialogue itself is input. ▼ Sub-theme: Output philosophy in the AI ​​era "AI handles what's within the framework, humans handle what's outside the framework." Only humans can play with AI. Therefore, input = playing is of utmost importance. Returning to a PC, abandoning the Pixel Fold, and going back to a large screen. ━━━━━━━━━━━━━━━━━━━━━━ 📝 This broadcast ━━━━━━━━━━━━━━━━━━━━━━ A reader of 20 years meets the author for the first time / "We're all Katsuma children" / Researching voice input since 2010 / Whisper x SuperWhisper combination / Prompt switching for each model / Thumb shift and G-Tune / Groq is "an LPU company" / The shock of 11 cents per hour / Fingers are 4 times slower than thoughts / Writing a 1000-character newsletter in 3 minutes / Strict order to Gemini: "Don't add things arbitrarily" / 1-thread operation, I hate Gems / The story of writing an entire book in Hawaii / The theory of double-door refrigerators / The non-conversion key is a toggle / Quiet typing in a cafe / The sensuality of undulating waveforms / Switched from Fold to PC / Owns 6 PCs and all Pixel models / Plays Gemini 20-30 times a day / Humans handle everything outside the frame / You can't survive the AI ​​era without having fun ━━━━━━━━━━━━━━━━━━━━━━ 🎙️ Credits ━━━━━━━━━━━━━━━━━━━━━━ Cast: Kei Tanaka / Kensuke Guest: Kazuyo Katsuma Editing: AL Co., Ltd. Production Management: AL Co., Ltd.