Florian Clanet 👨‍💻🌤️

I just went through AWS's guide on using the Kiro CLI for infrastructure optimization.
Really love the prebuilt commands!

You can ask Kiro directly from your terminal:
"Get me compute optimizer recommendations for my account"

AI enhanced coding without good CICD and devops practices is just a 10x problem catalyst.

Enhanced coding with AI will be 2026 main topic. I mean "real enhanced coding with AI": with code hitting production quality standards, scalability, security and (the hardest one) maintainability.
Tools are evolving quickly and will not be the main driver, process and people will

Open Protocols for Agent Interoperability Part 2: Authentication on MCP | Amazon Web Services In Part 1 of our blog series on Open Protocols for Agent Interoperability we covered how the Model Context Protocol (MCP) can be used to facilitate inter-agent communication and the MCP specification ...

Just read an interesting blog post on OAuth2 and the Model Context Protocol (MCP). It's all about making agent communication secure. Check out the new authentication approach, it might help with the authentication question your company is wondering about with MCP deployments.
go.aws/4epNGUx

Not tried Claude Code yet. Only Cline (with prompt caching) and Q CLI with Claude 4.

I'm exploring agentic GenAI for a few months, and I'm still finding it super interesting! It's cool to see how these systems can operate autonomously. Anyone else into agentic solutions? Would love to hear your thoughts and experiences, especially in production.

I launched Q CLI to work on and Amazon Bedrock logs visualizer. Let's see what we can do without writing a single line of code...

Back this morning from the most exciting vacation time ever!
What's new here?

Long time no see here. What's the status out there?

Chain-of-Draft (CoD) Is The New King Of Prompting Techniques A deep dive into the novel Chain-of-Draft (CoD) Prompting that reducing LLM inference cost and latency like never before.

Read “Chain-of-Draft (CoD) Is The New King Of Prompting Techniques“ by Dr. Ashish Bamania on Medium: levelup.gitconnected.com/chain-of-dra...

GM ☕️
Let's go for a one week trip!
What's your plan for this week?

Anthropic’s Claude 3.7 Sonnet hybrid reasoning model is now available in Amazon Bedrock | Amazon Web Services Claude 3.7 Sonnet hybrid reasoning model is Anthropic’s most intelligent model to date excelling at coding and powering AI agents. It is the first Claude model to offer extended thinking—the ability t...

Anthropic Claude 3.7 Sonnet is available on Amazon Bedrock!
This is going to be really interesting for tasks that need deep reasoning, but also to improve the quality of answers in general while matching the price of Sonnet 3.5.

Let's hope Claude Code is next!

aws.amazon.com/fr/blogs/aws...

Training Large Language Models to Reason in Continuous Latent Space We discuss a new paper that seeks to improve how LLMs handle complex reasoning tasks by drawing inspiration from human cognition.

Interesting new Coconut paper article.
arize.com/blog/trainin...

LinkedIn co-founder Reid Hoffman just raised $25 million to take on cancer with AI Reid Hoffman announced the launch of Manas AI, which will use AI to discover new treatments for a variety of diseases

Interesting time! qz.com/reid-hoffman...

AI Agents: Hype versus Reality, redux The contrast couldn’t be starker

As a reminder that AI Agents in terms of the orgs making them are largely about hype, as always check out the amazing @garymarcus.bsky.social - garymarcus.substack.com/p/ai-agents-...

Answer.AI - Practical AI R&D – Answer.AI Practical AI R&D

Just two pieces of content to start the day:
- www.answer.ai
- www.fast.ai
Enjoy your Thursday!

On Choosing Your Complexity | James Eastham | Just a storytelling technologist thinking about systems, software and people. Software is complex, but you can choose your complexity. This post explores the things you do have control over, and the things you don't.

Partly something I've had in draft for a while, partly inspired by Werner's keynote at re:Invent. My rambling thoughts about complexity in software, and how you should think about choosing your complexity...

jameseastham.co.uk/blog/on-choo...

#softwarearchitecture

Structure-based drug design with equivariant diffusion models - Nature Computational Science This work applies diffusion models to conditional molecule generation and shows how they can be used to tackle various structure-based drug design problems

After two years, our paper on generative models for structure-based drug design is finally out in @natcomputsci.bsky.social

www.nature.com/articles/s43...

AWS re:Invent 2024 Serverless - YouTube

I curated a YouTube playlist with my favorite serverless sessions from AWS re:Invent 2024. Featuring:

@ranthebuilder.cloud
@julianwood.com
@alexdebrie.bsky.social
@marcbrooker.bsky.social
@raphaelmanke.bsky.social
and more great folks that I couldn't find here 🥲

youtube.com/playlist?lis...

The phony comforts of AI skepticism It’s fun to say that artificial intelligence is fake and sucks — but evidence is mounting that it’s real and dangerous

The exchange between @caseynewton.bsky.social and @garymarcus.bsky.social on AI scepticism is perfect material for dialectic synthesis. Take the thesis (www.platformer.news/ai-skeptics-...), add the antithesis (garymarcus.substack.com/p/hard-forke...) and draw your own conclusion/synthesis. 1/5

All in one LLM Token Counter Calculate tokens of prompt for all popular LLMs including GPT-4, Claude-3, Llama-3 and many more using pure browser-based Tokenizer.

Want ton estimate your token count for several models (GPT-3.5, GPT-4, Claude-3, Llama-3, and many others)?
Here is a nice and easy link to do so: token-counter.app
(Warning: do not put confidential or regulated data in here...)

With the associated Github: github.com/ppaanngggg/t...

Today re:Invent AIML keynote was interesting... let's see how we will leverage all these news 🚀

The Amazon Nova family of models: Technical report and model card We present Amazon Nova, a new generation of state-of-the-art foundation models that deliver frontier intelligence and industry-leading price performance. Amazon Nova Pro is a highly-capable multimodal...

The Amazon Nova family of models: Technical report and model card #aws www.amazon.science/publications...

Let's see what the latest re:Invent launches for Q Developer can do! I accepted this Corey Quinn challenge and put Q Developer Agent to work on a task he provided: www.youtube.com/watch?v=xMvN...

What's your best re:Invent announcement or session so far?

Nice, I like it!

Let’s Architect! Serverless developer experience in AWS | Amazon Web Services Are you a developer approaching serverless for the first time, or even an experienced one looking for a better way to accelerate your feedback loop from code to production? This collection of resource...

Enhanced developer experience for serverless 🔥 This is your starter pack!
- Better local IDE for lambdas
- TDD with Amazon Q Developer
- Optimize lambda memory size
- NoSQL workbench for DynamoDB
- Lambda Powertools for observability
- A workshop to test them all
aws.amazon.com/fr/blogs/arc...

I'm way too excited to see the
@pydantic.bsky.social crew cooking.

Hey, it's #reInvent time again! 🎉 #AWS

New APIs in Amazon Bedrock to enhance RAG applications, now available | Amazon Web Services With custom connectors and reranking models, you can enhance RAG applications by enabling direct ingestion to knowledge bases without requiring a full sync, and improving response relevance through…

New APIs in Amazon Bedrock to enhance RAG applications 👉 Support for (1) custom connectors and ingestion of streaming data and (2) reranking models https://aws.amazon.com/blogs/aws/new-apis-in-amazon-bedrock-to-enhance-rag-applications-now-available/ #AWS #reInvent #AI #ML #GenAI #MachineLearning #…