The Sandra Kublik Podcast

In this episode, I unpack the three most common blockers I keep bumping into when working with businesses on their internal AI adoption + share example techniques to address them, so you can start swapping "I'll wait till it's perfect" for hands-on wins. If you're interested in a collab, reach out at: team [at] sandrakublik.com

📌 Free resources: AI Workflow Discovery Template https://sandrakublik.kit.com/bf4f1634a0 👋 Say hi

Twitter/X: https://x.com/itsSandraKublik

LinkedIn: https://www.linkedin.com/in/itssandrakublik/

Video version of this episode on YouTube: https://youtu.be/-f8rDW7xFF8

In this episode, I'm sharing 4 field-tested tools I use to establish baseline AI usage in any organization: usage stats, user surveys, one-on-one chats, and diary studies. They are a powerful combination that lets you gather and analyze data to identify training needs, understand blockers, and optimize AI tool adoption across different departments.

Want to work together? If you're interested in chatting about a project, please reach me at: team [at] sandra.kublik.com

📌 Free resources: AI Workflow Discovery Template: https://sandrakublik.kit.com/bf4f1634a0

👋 Say hi:

X: https://x.com/itsSandraKublik

LinkedIn: https://www.linkedin.com/in/itssandrakublik/

Chapters:

00:00 What gets measured gets managed

00:35 Establishing your baseline AI usage

01:04 Looking at your usage stats

03:13 User surveys: beyond analytics

04:59 Going deeper with 1-on-1 user chats

08:03 Diary studies: a day in the life

10:28 Additional resources

♬ Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

Shopify published an internal AI memo blueprint that says: "AI is now mandatory". I've annotated the memo section by section and today I'll share takeaways I recommend to copy, to question, and to personalize for your own AI adoption playbook.

Thank you for hanging out with me today!

Resources:

Full text of the Shopify CEO internal AI memo: https://x.com/tobi/status/1909251946235437514

Watch this episode on YouTube: https://youtu.be/LmUmc_1SRZ8

About me: www.sandrakublik.com

Say hi via X/Twitter and LinkedIn @itsSandraKublik

For collaboration, reach out at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In this episode, I talk with Swyx, co-host of the Latent Space podcast, co-founder of the AI Engineer World's Fair conference, and founder of the AI News newsletter.

Here's what we cover:

• The AI Engineer Persona: What defines this new role and how it keeps evolving
• AI Startup Moats: What it takes to create a competitive edge as an AI startup today
• AI News Newsletter: Swyx walks me through his latest venture, a popular newsletter that many in the AI space rely on for industry updates.
• The Future of AI Engineering: Impressions from AI Engineer World's Fair 2024 and the ever-changing landscape of AI space

Resources:

Watch all the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff

AI Engineer World's Fair YouTube channel: @aiDotEngineer

Latent Space Podcast YouTube channel: @LatentSpaceTV

Swyx Twitter: @swyx

Swyx Talk from AIEWF 2024: https://www.youtube.com/live/5zE2sMka620?si=sRtP5o_A1HHeYvBl&t=29923

My Talk from AIEWF 2024: https://youtu.be/vNssL4u5jb8?si=r65xPlZoU0dlpUZp

The Rise of AI Engineer Essay: https://www.latent.space/p/ai-engineer

AI News Newsletter: https://buttondown.email/ainews#subscribe-form

How to Hire AI Engineers — with James Brady & Adam Wiggins of Elicit episode: https://www.latent.space/p/hiring

Heralds of the AI Content Flippening — with Youssef Rizk of Wondercraft.ai: https://www.latent.space/p/wondercraft

About me: www.sandrakublik.com

You can keep in touch with me via IG, TikTok, X, LinkedIn @itsSandraKublik

For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In today's episode, I chat about retrieval-augmented generation (RAG) with Naomi White, Engineering Program Manager at Cohere. Naomi shares her unique journey from starting as an Executive Assistant at DeepMind to becoming a key player in Cohere's cutting-edge RAG development team.

In this episode, we cover: - Naomi's inspiring journey from DeepMind to Cohere and building a career in AI as a non-technical person; - Strategies for navigating the imposter syndrome; - Beginner-friendly tips for learning LLM concepts and staying up to date with AI advancements; - A behind-the-scenes look at developing LLM tool use with RAG techniques;

Resources:

* Jay Alammar's walkthrough "Building LLM Agents With Tool Use": https://youtu.be/5drn2DO7gNY?feature=shared

* Notebook On Multi-Step Tool Use With Cohere And LangChain: https://colab.research.google.com/github/cohere-ai/notebooks/blob/main/notebooks/Data_Analyst_Agent_Cohere_and_Langchain.ipynb

* GitHub repo with open-source chat interface optimized for RAG: https://github.com/cohere-ai/cohere-toolkit

* Learn more about Command R+, Cohere model optimized for RAG use cases: https://txt.cohere.com/command-r-plus-microsoft-azure/

* Try it out: https://cohere.com/coral

Watch podcast episodes on YouTube: https://youtu.be/CzcfrY1aKho&list=UULFjG6QzmabZrBEeGh3vi-wDQ&ab_channel=SandraKublik

I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228

You can keep in touch with me via socials:

IG, TikTok, X, LinkedIn @itsSandraKublik

Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik

Website: www.sandrakublik.com

For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In this episode, I chat about rerankers and representation models with Sylvie Shi, Rerank Team Lead at Cohere. We cover the essentials of non-generative large language models, focusing on how reranking and representation models are upgrading search and advanced RAG systems.

Sylvie shares her journey into machine learning and her work on semantic search and rerankers. We chat about common use cases, good practices, and potential pitfalls when using reranking models, highlighting their application in advanced RAG systems.

• Guide On How To Build Advanced RAG With Command R+ and Rerank: https://cohere.com/blog/rag-chatbot
• Text Representation, Embeddings, Reranking - LLM University course resources: https://docs.cohere.com/docs/intro-large-language-models https://docs.cohere.com/docs/intro-text-representation https://docs.cohere.com/docs/intro-semantic-search https://docs.cohere.com/docs/reranking-2
• Reranking Best Practices: https://docs.cohere.com/docs/reranking-best-practices
• Learn more about Rerank 3: https://cohere.com/blog/rerank-3
• Try it out: https://docs.cohere.com/docs/rerank-2

Watch all the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff

I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228

You can keep in touch with me via socials:

IG, TikTok, X, LinkedIn @itsSandraKublik

Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik

Website: www.sandrakublik.com

For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In today's episode, I chat about retrieval-augmented generation (RAG) with Ola Piktus, Member of Technical Staff at Cohere. With her roots in software engineering and significant contributions to RAG through her research, Ola has an interesting perspective on the evolution of RAG and the language technology landscape.

We cover a range of topics, including the potential of RAG to address model hallucination, the transition from theory to real-world application, and the balance between creativity and accuracy in AI models. Ola's journey from software engineering to state-of-the-art research in AI is a rare glimpse into the development of language tech.

• Ola's research paper "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks": https://arxiv.org/abs/2005.11401
• Ola's research paper "Improving Wikipedia Verifiability with AI": https://arxiv.org/abs/2207.06220
• Learn more about Command R+, Cohere model optimized for RAG use cases: https://txt.cohere.com/command-r-plus-microsoft-azure/
• Try it out: https://cohere.com/coral

Watch the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228 You can keep in touch with me via socials: IG, TikTok, X, LinkedIn @itsSandraKublik Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik Website: www.sandrakublik.com For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In today's episode, I chat about the latest in Large Language Models research with Sara Hooker, VP of Cohere For AI research lab. Sara shares her experiences, from growing up in diverse cultures to leading cutting-edge AI research. We delve into fundamental ML questions, the importance of diverse linguistic representation, Aya, the open questions of multilingual models and language models in general, and LLM projects in 2024.

• Sara's interview with Sam Charrington on TWIML: https://youtu.be/MIvfCKwLnq8?si=QnnPo46ebQiIsyOT
• Learn more about C4AI research: https://cohere.com/research
• Learn more about Aya project: https://cohere.com/research/aya

Watch the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228 You can keep in touch with me via socials: IG, TikTok, X, LinkedIn @itsSandraKublik Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik Website: www.sandrakublik.com For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In this episode, I explore the depths of AI safety and bias with Seraphina Goldfarb-Tarrant, Head of AI Safety at Cohere. We discuss mitigating AI risks, the impact of large language models on society, and the future of responsible AI development. Seraphina shares insights on participatory design, combating misinformation, and proactive steps toward safety for businesses deploying LLMs.

We also cover Seraphina's academic journey, her transition from neuroscience to studying ancient civilizations, becoming a professional sailor, her connection with nature and Zen practice, and her perception of language, mind, and AI.

Seraphina's Enterprise Guide on AI Safety: https://txt.cohere.com/the-enterprise-guide-to-ai-safety/

Watch the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228 You can keep in touch with me via socials: IG, TikTok, X, LinkedIn @itsSandraKublik Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik Website: www.sandrakublik.com For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/

In this conversation, I sit down with Elaine Gao, Product Manager at Cohere, to explore chat interfaces and building LLM-based products. We talk about creating interfaces that users love, the ethical implications of LLM products, and what the future holds for our dynamic space. Elaine also shares her journey from business school to becoming a leader in managing LLM products. In our conversation, we cover: - Navigating career transitions - The culture of Silicon Valley and its impact - The art of user-centric design in technology - Balancing innovation with ethical considerations - Future trends in language technology and AI Watch the episodes on YouTube: https://www.youtube.com/playlist?list=PLhBwB1lBTwJWR19WkJw87EMh1JU9O7Uff I'm a co-author of a manual on GPT-3 and OpenAI API (Packt, 2023): https://www.packtpub.com/product/gpt-3/9781805125228 You can keep in touch with me via socials: IG, TikTok, X, LinkedIn @itsSandraKublik Sign up for my Substack newsletter for the personal scoop on the topics covered today: https://substack.com/@itssandrakublik Website: www.sandrakublik.com For brand collabs, reach out to me at [email protected]

Music by: Acid Jazz by Kevin MacLeod is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/by/4.0/