DeepMind released a new post today listing a number of areas where AlphaEvolve has contributed value so far and it looks impressive. My experience so far experimenting with these evolutionary techniques doesn’t align with the results I see there which makes me believe I’ve been doing something wrong so far. I’ll dig more into this. Which reminded me that I also need to continue with the rest of the posts in the optimization series that I started.
I was aware of ACE and GEPA as methodologies for improving your prompts in a kind of automated way but apparently there’s another one that I just learned about Combee. From skimming into the Combee paper, it seems that it’s focusing on how to leverage trajectories from parallel agent executions, something the other two don’t do. I need to get into more details but what I’m wondering is if this approach makes it any easier to use these systems outside of a research environment. Having a way to self-improve sounds great but in practice, I found it hard to find a way to incorporate something like ACE in a real production environment.
A practitioner's account of 16 AI-driven optimization experiments on Apache Arrow's select_k_unstable kernel, from 2x to 396x.
Continuum: Efficient and Robust Multi-Turn LLM Agent Scheduling with KV Cache… reminded me about the importance of KV Caches. I was thinking a lot about them a few months ago about the potential of bridging the inference infrastructure with a query engine that can optimize the plan execution to optimize the use of the KV Cache, then with agents I thought that optimizing the KV Cache won’t be as important but I was very wrong. Anthropic and OpenAI are doing a lot of work to optimize the cache based on the workloads and agents are in many respects a different workload on their own. This paper has a smart approach in optimizing the KV cache based on a specific characteristic of the agentic workload related to the back and forth with tool calling. Context engineering at the systems layer!
I made my first Claude Code plugin today to quickly build an agentic application I wanted to test and I realized that it’s a great platform to quickly test new ideas and agentic architectures, especially for multi-agent type of things. It does have it’s issues and limitations, e.g. you can’t really use a non anthropic model right? That’s exactly the reason I find it a great way to quickly put MVPs together to test ideas!
Just tried Anthropic’s Cowork and my first impression is that it indeed brings the experience of Claude Code on the desktop environment, especially if you combine it with the Claude extension for Chrome. It feels more and more that Anthropic is winning over OpenAI when it comes to building products and not only models. (Make sure you understand the security implications before you try Cowork!)
Keystatic is pretty cool. It turns your github repo into a CMS in a very elegant way. The experience of setting up my own GitHub app for oAuth was miserable though, but the keystatic.cloud service they offer is free and it works. Happy that I found this.
Two years ago I started working on something new based on a strong conviction that data was about to become exponentially more important. We built Typedef, an AI-native data engine that unifies inference, search, and data processing into a single system.
Snowflake's 2025 keynote announcement regarding dbt Core clarifies dbt Labs' controversial shift to change dbt Fusion's licensing model. This represents a rational business response to market dynamics.
Snowflake's Cortex AISQL represents a paradigm shift in integrating large language models into data systems as structured, composable functions rather than opaque tools.