Scalable CTO - AI скачать в хорошем качестве

Scalable CTO - AI

Scalable CTO podcast with Bob Dickinson. Covering AI, including overview of AI/ML and LLMs, properties of LLMs, prompt engineering, grounding, retrieval augmented generation (RAG), prompts as code, and our prompt engineering and tooling, include prompt registry and custom prompt studio, and prompt CI/CD. Usage of Google Vertex AI, including Gemini models, image-to-text models, prediction/embeddings services, and vector indexes in Cloud Firestore. API demos and source code review. Resources referenced: ChatGPT is Bullshit, Hicks, Humphries, & Slater, June 8, 2024 https://link.springer.com/article/10.... Prompt Engineering Guide https://www.promptingguide.ai/ Prompt Decomposition   / prompt-decomposition   Let Me Speak Freely? A Study on the Impact of Format Restrictions on Performance of Large Language Models https://arxiv.org/abs/2408.02442 00:00 - Intro 00:45 - AI overview 04:07 - GenAI hype cycle 06:56 - LLMs 16:44 - Prompt engineering 27:15 - Grounding and RAG 32:05 - How we use LLMs 37:38 - Question answering with embeddings 43:49 - Summarization 47:50 - Prompts as Code 53:23 - Our AI APIs and tooling