2️⃣ Second Day 1 of re:invent is in the books. I updated my All things (Gen) AI article over the morning coffee.

🎡 Very exciting updates including a new entire Model family Amazon Nova with multi-modal understanding models and content generation models for video and images, multi-agent support and model-distillation support and automated reasoning checks in guardrails.

🔇 All the updates in the article are in italic letters for now so that you can have an incremental read(best solution I could find with limited capabilities of LinkedIn articles).

First version of my re:invent summary on all things Gen AI. Trying to keep up-to-date, putting announcements into perspective and taking you along on this journey. I hope to take the time to update this daily - Let’s see how it goes. Let me know if this is helpful.

👏 Kudos to Ioana Blaut for compiling this nice overview post!

💡 Indeed it has been an awesome event. Good to re-connect with customers, partners and colleagues and getting to know some new - very interesting - people. Still reflecting on the numerous insightful and deep discussions.

🧳 Super excited to hear from customers who started their initially impossible appearing transformation journey and making great progress. Obviously #GenAI was also a topic. Great customer products and partner solutions showed clearly visible that we are moving uphill on the slope of enlightenment. Another journey I’m thrilled to continue with our customers and partners.

Is Yoko Ono an early prompt engineer?!

Sounds like a random question, but I think she is. She is prompting the audience to interact and participate in her art works.

So basically she has been often provided simple text prompts, to engage viewers to become agents, happily embracing the non-deterministic nature of the overall system . Sample prompts: “Observe three paintings carefully. Mix them well in your head”, “Draw a map to get lost”. Earlier version of the later which has some step by step instructions is in one of the attached pictures.

Sharing some positive energy from my morning run and the nice progress in preparation for next week’s MEDIENTAGE MÜNCHEN.

I’m planning to be at the event for all 3 days - happy to meet you there. I’ll be running two session there:

Oliver and I will kick-off on Wednesday with the masterclass “Von der Idee zur Realität” where we dive deep into various use cases for GenAI in the Media&Entertaining industry and how you can build those on AWS. [German Session]

Now that we have Generative AI and the power of LLMs, we can forget about classic recommender systems, right?!

Maybe not, we will see. Nensi Hakobjanyan and I take you through a journey of understanding what role LLMs can play in personalisation today. Spoiler - seldom in life we need to e.g. decide between taking a red or a blue pill, but often a good solution is composed by multiple approaches 🙂.