Скачать с ютуб OpenAI клонирует голоса, а StableDiffusion генерирует музыку / GPT в науке / AIA Podcast #31 в хорошем качестве

OpenAI клонирует голоса, а StableDiffusion генерирует музыку / GPT в науке / AIA Podcast #31

Ребятушки, там в выпуске как всегда ооочень интересно, но времени на написание описание сегодня не было. Вы пока так послушайте, а в следующем выпуске подвезем и ссылки и таймкоды и вот это вот всё. Приятного времяпрепровождения! 👉 AW.club: https://bit.ly/awclub-ru 👉 Наш чатик в ТГ: https://t.me/aiapodcast 👉 Каталог ИИ-сервисов: https://awclub.github.io/catalog НАВИГАЦИЯ 00:00:00 Solving the mystery of the "character sequence" from the previous episode 00:03:13 A joke about data vectorization 00:05:09 Introduction and greetings 00:06:01 Thanks to subscribers and invitation to the Telegram chat 00:07:11 Information about the podcast sponsors: Victor, CLW, and Engineering Excellence 00:08:28 Information about the podcast's service catalog and blog 00:09:40 Offer for paid visits from the hosts to listeners (joke) 00:10:01 A joke about artificial intelligence 00:10:08 Possible teaser about new friends of the podcast 00:10:16 Information about the podcast's friends: Engineering Excellence and Vitalik Vishnevsky 00:11:27 Transition to the "Big Fish" section 00:11:47 OpenAI and the April Fool's joke (or not?) about free access to GPT-3.5 00:16:11 OpenAI released a voice cloning model - Voicebox 00:17:57 Discussion of multimodal models 0:18:49 OpenAI and Microsoft plan to build a $100 billion data center called "Stargate" 00:22:47 Stability AI: Ed Mostack's (mastak's) departure and the launch of a new service - Harmonai 00:25:24 Harmonai - a competitor to Jukebox from OpenAI 00:28:03 Jukebox v3 - music generation in Russian 00:29:11 Infinite radio from Boomy 00:30:03 Apple will teach Siri to "see" the screen and context 00:32:56 X.AI announced Bard 1.5 00:37:10 DeepMind presented a new model - "Mixture of Experts" 00:41:34 Anthropic talked about a new vulnerability in large language models 00:44:41 Databricks presented a new open model - Dolly 2.0 00:49:48 JetBrains AI Assistant is now integrated into IDEs 00:55:05 Intel confirms that Microsoft Copilot will be able to run locally on computers 01:00:00 Discussion of the impact of gamers and the porn industry on technology development 01:00:10 Interview with the creator of adult comics using AI in the podcast "We are doomed" 01:01:42 Transition to the "Science and Technology" section 01:01:54 Neuralink: demonstration of the first patient with an implant playing Civilization VI 01:13:54 New season of the podcast "Naked Mole Rat" and an experiment with Ilya Kolmanovsky 01:18:07 Open Interpreter OS - an Open Source project that has grown into an operating system 01:22:12 Apple closed Project Titan (electric car) and may switch to robotics 01:29:23 The "Law and Order" section is empty 01:29:50 "Services and Links" section: 01:30:03 Commander - a large language model with 132 billion parameters 01:32:19 Devy - an Open Source analogue of DevMode from Google AI 01:35:40 S-Agent - an Open Source project from Princeton University that writes code and pull requests 01:38:01 CodeWP - a commercial service for code review 01:40:05 VOC - Zero-shot Text-to-speech engine 01:44:11 Discussion of the podcast contest and its completion date 01:46:06 Ethics: the earthquake in Taiwan and the world's dependence on chip production in that country 01:56:07 Conclusion and farewell ГДЕ НАС СМОТРЕТЬ И СЛУШАТЬ 🔸 YouTube:    • OpenAI клонирует голоса, а StableDiff...   🔸 Google Podcasts: https://bit.ly/aia-google 🔸 Apple Podcasts: https://bit.ly/aia-apple 🔸 Spotify: https://bit.ly/aia-spotify 🔸 Скачать mp3: https://aia.simplecast.com/episodes/31 Сервисы из выпуска (ну хоть их успели добавить) 🔹https://github.com/stitionai/devika - опенсорсный девиан 🔹 SWE-agent - https://github.com/princeton-nlp/SWE-... 🔹 https://coderabbit.ai - авторевьювер коммерческий 🔹 Command R+ https://txt.cohere.com/command-r-plus... 🔹 https://github.com/jasonppy/VoiceCraft