Happy Thanksgiving! Download my new story to share with the little ones over the break. imagine.bolbeck.com/showcase/Tha...

Today is National Family Literacy day, the perfect day to pick a book (or three) and share its magic with the family. Have quality time with your loved ones and inspire the little ones to build good reading habits. #NationalFamilyLiteracyDay #Literacy

Celebrating all the authors whose vivid imaginations transport us to far away lands. Thank you for sharing vision! #NationalAuthorsDay #writers #Authors

I really think that AI can be a useful tool to enhance our daily lives. In the end, it is up to the creative person behind the keyboard that guides the AI to produce an desired outcome. It is this belief that is woven into how we have designed Bolbeck Imagine. #humanincharge #childrensbooks

Check out the quick intro video to Bolbeck Imagine. lnkd.in/ed3xCHJX Putting a lot of time ensuring you can create your next story book and would love to get some constructive feedback on what you think of the platform. So, create a free account and let me know what you think!

I have always thought teachers have one of the most important jobs in society. I am building Bolbeck imagine to help them create books that kids may actually want to read and thus make teachers lives a bit easier. Just upload the manuscript and work with the assistant to create your PDF book.

Have you ever had an idea for a great story? Bring it to life with Bolbeck Imagine.

My goal to make it easy for anyone to make a children’s book continues…

I had a great time presenting my talk about the good, the bad and the just plain cool about creating applications that use AI agents at The AI Collective + Chicago Java Users Group meetup today. We had great questions and insightful conversations

Agentic workflows are not a replacement for microservices; they serve as a new coordination layer built upon existing service invocation.

👿 MCP is all fun, until you add this one malicious MCP server and forget about it. We have discovered a critical flaw in the widely-used Model Context Protocol (MCP) that enables a new form of LLM attack we term 'Tool Poisoning'. Leaks SSH key, API keys, etc. Details below 👇

I had a great time at the ArcOfAI Conference. Lots of great people with great perspectives about AI. Plus I delivered my talk on lessons learned from building agentic apps! Thanks to the organizers that created such a well organized event! @arcofai.bsky.social www.bolbeck.com/blogs/post/l...

Flying to Austin today for the Arc of AI conference ! Excited to learn new concepts and have insightful conversations ! Plus on Thursday, I give my talk and share what i have learnt while building AI apps. @arcofai.bsky.social

Attend @juanperedotech.bsky.social's session at Arc of AI Conference, to discuss lessons learned from the journey in creating agentic AI applications 🤖 www.arcofai.com/speaker/f631... Don't miss out and get your tickets now! Use JOIN-JUAN-50OFF for discount 📅Mar 31-Apr 3. Austin TX 🎟️ arcofai.com

Great summary of the fantastic talks at AI Engineer Online Summit. Check it out & see which ones spike your interest. You could, for example, watch my talk: Lessons from building GenAI based applications 😇 Links to all the talks are in the summary. www.linkedin.com/pulse/agent-...

Google launches a free AI coding assistant with very high usage caps

In the face of competition from Claude 3.7 Sonnet in coding, Google is launching a free version of Gemini Code Assist globally. It comes with: - 180,000 code completions per month - Support for all programming languages in the public domain - 128,000 token context window

Are you intrigued by creating GenAI-based applications? I invite you to watch my video on the AI Engineer channel, where I share insights from my experience in developing GenAI apps. The video will prepare you for a successful start in your AI journey. m.youtube.com/watch?v=YYcN...

1/5 So, how did DeepSeek develop DeepSeek R1? They used both DeepSeek-V3-Base and a simple prompt: 1. They asked the same question multiple times to DeepSeek-V3-Base as a group. 2. They then graded the answers, assigning an accuracy score and a format score (e.g., <think></think>).

Happy to share that I will be presenting at the Arc of AI conference in April! Join me and a distinguished list of speakers in Austin. Let’s share ideas and make some friends! @arcofai.bsky.social www.arcofai.com

🚀 Discover the Best of #SoftwareArchitecture from 2024! Explore this must-read collection of exceptional articles published on #InfoQ last year! 💡 Stay informed. 🔥 Stay inspired. 🏆 And always #StayAhead of the curve! 👉 Knowledge is power! 💪 Thread 👇

Someone on Reddit searched for (the non-existent) John Wick 5, and Google spared no expense to lie to them.

DeepSeek, a LLM trained for a fraction of the cost of GPT-Xx models, in 2 months for 6 million, on limited GPUs due to export restrictions, and competing head to head. This is crazy. It's not the AI part I'm excited about, it's the level of efficiency. github.com/deepseek-ai/...

Maybe one of the best explanations of what it means to be a true senior software dev and leader:

Migrating to uv by @ivanleomk.bsky.social He wrote a guide on how we migrated instructor from poetry to uv and documented the gotchas that they encountered They saw around a 3x speedup and a ~67% reduction in CI Pipeline execution without any code changes python.useinstructor.com/blog/2024/12...

Confirmed. DeepSeek-V3 is coming soon and it is bit better than Claude 3.5 Sonnet per Aider LLM Leaderboard. aider.chat/docs/leaderb...

According to Pluralsight 2025 Tech Forecast by, 20% of companies have deployed AI projects & 55% have plans to do so in 2025. However, 80% of all AI projects fail which is twice the failure rate of other types of projects. Read the full report here: www.pluralsight.com/content/dam/...

LangChin @langchain.bsky.social just released their State of AI report which is full of interesting tidbits like the top LLM providers and the top LLM-as-judge evaluation metrics. You can read the report here: blog.langchain.dev/langchain-st...

Introducing diff view in LangSmith's Prompt Hub 🔍 Easily compare changes between any two commits, review updates, or revert to earlier versions. Go to the Commits tab, toggle 'diff,' and track your prompt evolution — so you can iterate quickly and with control.

Here are the slides from my “Tips for building applications that use GenAI” talk at AICamp earlier this week. It was fun sharing some AI dev tidbits with the audience! #GenAI #AI www.bolbeck.com/files/TipsFo...

Had a great time presenting “Tips and tricks for building applications that use GenAI” at AICamp Chicago today! We had great questions and discussions throughout the talk! #AI

At AWS re:Invent, CTO Werner Vogels shared some fundamental lessons Amazon learned about system design.

Meta releases llama 3.3 70B which rivals in performance the existing Llama 3.1 405B even though it is less than 1/5 the size. Just as exciting, you can now run it in Groq. @groqinc.bsky.social groq.com/a-new-scalin...

Open-source #LLMs are essential for making #AI more accessible! Discover more about OLMo 2 in my latest @infoq.bsky.social check it out!

If someone is curious about the new Amazon Nova models (and many other models), they are available in openrouter openrouter.ai/amazon

Really enjoyed the keynote by @werner.social at reInvent! It was filled with great insights like: - Make evolvability a requirement - Break complexity into pieces - Align organization to architecture - Organize into cells - Design predictable systems - Automate complexity Best keynote at reInvent

AWS announced EKS Auto mode to simplify the management of EKS Kubernetes clusters. Looks interesting and potentially very useful. docs.aws.amazon.com/eks/latest/u...

smollm Everything about the SmolLM & SmolLM2 family of models. https://github.com/huggingface/smollm

AWS just announced a new family of LLMs at ReInvent today called Nova. They seem to be competitively priced and their benchmarks seem pretty good, specially for a V1 product. We will soon find out if reality matches their excellent marketing :) aws.amazon.com/ai/generativ...

This tutorial shows how to use asyncio to work around the limitations of Python’s global interpreter lock (GIL) to achieve efficient concurrent programming.

Vendor beefs up support for re-ranking models, sparse vector retrieval, and security features like audit logs, RBAC, and more.

🎅🎄🎁 ADVENT OF SVELTE 🎅🎄🎁 We're going to ship one thing a day every day from now until Christmas. Wish us luck! svelte.dev/blog/advent-... First up...

Small is beautiful and better for the environment :) Choosing the correct LLM/SLLM for the correct task can help saving money while still achieving the appropriate business goals.

Predictive task scaling in AWS ECS. Assuming this works as announced, it could help right size your container tasks to more closely match demand. This in turn could potentially reduce cost and/or improve application response to traffic spikes. aws.amazon.com/blogs/contai...

To all of those celebrating, Happy Thanksgiving!

Menlo Ventures’ State of AI shows, among other things, LLMs share by provider. Interestingly, OpenAI is still on the lead but it has drop by 16% points year over year while Anthropic has gained 12 points. Here the report that is full of interesting info about AI in 2024: menlovc.com/2024-the-sta...