Тема редкая для этого блога, но иногда поднималась. Недавно испытал организаторско-каталогизаторский, простите, оргазм, натолкнувшись на logseq (website, youtube). Очередно приложение для ведения заметок, а также аутлайнер, календарь, анки, канбан, управление проектами, TODO и многое другое (в том числе с помощью плагинов). Причём, всё это сделано на текстовых файлах с лёгкой разметкой (поддерживает markdown и org) и двунаправленных ссылках (что позволяет использовать logseq для цеттелькастена).
Само по себе это не новость, есть очень популярный Obsidian (youtube), например. Но в отличие от Обсидиана, logseq вдохновлялся emacs с org-mode и имеет с ними почти полную совместимость. Вот человек использует logseq вместе с emacs, org-mode и org-roam. То есть, если нужно добавить коротенькую заметку, можно написать её в десктопном или мобильном logseq, а для больших литературных форм всегда можно запустить emacs со всей его мощью.
Советую ли я logseq? Если с компьютером на ты и хотелось когда-нибудь попробовать emacs, то разумеется. Если оба ответа "нет", но есть желание посмотреть на современное приложение по ведению заметок, то рекомендую попробовать Obsidian. Вдруг зайдёт.
***
Но это ещё не всё, что я хотел сказать. Пока игрался с logseq и интересовался последними достижениями в увеличении продуктивности, наткнулся на то, как для этого используют AI.
Вот Тим Феррис, известный на Западе инфоцыган бизнес-ангел и коуч, беседует с неким Сэмом Коркосом, который рассказывет, почему Loom важнейший бизнес-инструмент последних пяти лет.
1. **Introduction to Loom**: The speaker discusses the importance of Loom as a business enablement tool for recording screens and creating asynchronous messages. Loom allows for easy sharing of recorded content without lag time (00:00-00:29).
2. **Scaling Content**: The speaker talks about how content can scale infinitely using tools like Loom, unlike in-person interactions or phone calls (00:29-01:50).
3. **Example of Sharing Internally**: The speaker shares an example of an internal meeting with 40-50 people and how using Loom for such meetings can be more efficient than traditional methods (01:50-03:00).
4. **Use Cases for Loom**: The discussion covers various use cases for Loom, including giving verbal feedback, creating summaries, and challenging perceptions about its effectiveness (03:00-04:38).
5. **Recording Tips**: The speaker emphasizes the importance of recording without overthinking and doing only one take to avoid unnecessary repetition (04:38-05:34).
6. **Async Communication**: The concept of using audio and video memos during asynchronous communication and the initial discomfort it may cause (05:34-06:55).
7. **Searchability of Loom Content**: The discussion about how to make Loom content searchable, mentioning future improvements in AI-based search (06:55-10:37).
8. **Using Notion for Organization**: How the speaker's team uses Notion databases to organize tasks and link them to Loom videos for easy reference (10:37-12:42).
9. **Workflow Recording for Delegation**: The recommendation to record workflow processes to improve delegation and the surprising simplicity of replicating tasks through pattern matching (12:42-14:19).
via Harpa AI
Аналогичным образом Артём Кирсанов (студент computational neuroscience в МГУ, ведущий свой блог на Ютуб по-английски) обсуждал свою стратегию эффективного потребления информации, в т.ч. суммаризацию видео на Ютубе с помощью AI, а потом кто-то взял и прогнал его же видео через Harpa AI, получив суммаризацию:
00:00 📚 Effective Reading Strategy Introduction - The key to effective reading is starting with a summary of core ideas and gradually adding details. - Book summaries are valuable tools for more efficient reading, especially for non-fiction books. - Summaries provide a quick rundown of main ideas, helping prime the brain for incoming information. 01:57 📖 Importance of Book Summaries - Humanity has written about 150 million books; being selective is crucial. - Book summaries help filter information overload and ensure you invest time wisely. - Reading summaries before diving into content enhances understanding and engagement. 03:47 📘 Utilizing Book Summaries Effectively - Reading summaries side by side with the book enhances comprehension and note-taking. - Summaries remain useful even after reading the entire book for referencing and recalling content. - Book summaries serve as a tool for linking ideas and understanding the overall structure of a book. 04:42 📚 Shortform Platform Overview - Shortform is a go-to platform for detailed and holistic book summaries. - One-page summaries on Shortform are rich in information, offering key ideas in a concise format. - Full guides on Shortform provide a chapter-by-chapter breakdown of book content. 06:50 🗣️ Content Augmentation on Shortform - Shortform's content augmentation includes notes added by guide authors for additional context. - Shortform's book guides go beyond summarizing, providing background information, perspectives, and drawbacks. - The platform covers books from various fields, but not every book is available due to the effort required in creating guides. 08:25 📚 Integrating Shortform with Obsidian - Shortform integrates with Obsidian, allowing automatic syncing of highlights and notes. - Highlights made in Shortform sync to Obsidian via the Readwise API, creating corresponding nodes. - Refactoring notes in Obsidian helps organize and link ideas from book summaries effectively. 10:31 📘 Refactoring Book Highlights into Atomic Ideas - Refactoring highlights into atomic ideas in Obsidian follows a Zettelkasten note-taking philosophy. - Refactoring allows the creation of individual idea notes linked together for a comprehensive knowledge network. - Highlights from book summaries contribute to the development of a personal knowledge base. 11:56 📖 Extending Book Notes into Atomic Ideas - Transforming book summaries into atomic ideas enhances understanding and retention. - The Zettelkasten approach facilitates the organization of ideas from various books into a cohesive network. - Atomic ideas become valuable references for future work, essays, or personal notes. 12:39 🤖 Shortform Browser Extension for AI Summaries - Shortform's browser extension leverages AI to generate summaries for arbitrary content. - The extension simplifies the process of summarizing articles or YouTube videos without the need for complex prompts. - Integrating the extension with information consumption enhances efficiency and provides condensed versions of content. 15:23 🎓 Applying AI Summaries to Lecture Recordings - Lecture recordings serve as Triple A sources-accessible, aggregate, and accurate. - AI-generated summaries of lecture recordings provide condensed information with added context. - Skimming through AI summaries helps identify worthwhile lectures before investing time in watching them. 16:45 ⚙️ Limitations of AI Summaries - AI summaries may occasionally confuse uncommon words or names due to transcription errors. - Despite minor quirks, the AI tool performs well in providing condensed and augmented information. - Awareness of the tool's limitations ensures a mindful approach when using AI-generated summaries. 17:53 🛒 Conclusion and Shortform Offer - Shortform's AI browser extension simplifies summarizing any webpage content, making it a valuable tool. - Using Shortform in conjunction with Obsidian enhances the efficiency of information consumption and note-taking. - The video concludes with an offer for viewers to try Shortform with a referral link for a discount and free trial.
Made with HARPA AI
В принципе, то же самое можно повторить и локально. OpenAI отдала свою модель распознавания речи Whisper в открытый доступ, можно напустить на свою коллекцию видеокурсов и посмотреть, что будет. Как минимум это будет полезно для полнотекстового поиска и написания заметок (редактировать текст всё же быстрее, чем писать с что-то нуля, а Whisper может в дополнение к языку оригинала заодно транскрипт и перевести). В идеале же можно попробовать получить суммаризацию видео, как это делает Harpa AI. Если, конечно, найду какой-нибудь подходящий бесплатный инструмент для того. Пока беглый поиск не очень в этом помог.
Что до онлайн AI инструментов и сервисов, то их уже больше 1000:
Похоже, генеративный AI уже не хайп, как я думал, а реальная штука.
P.S. Случай с AI, кстати, наглядно демонстрирует, как американцы будут бороться с китайцам с помощью технологий (the third offset[1][2]): сначала ничего нет, ничего нет, потом какие-то фундаментальные результаты, а через год-два целая прикладная экосистема.