Tailscale's Lee Briggs on the dangers of MCP for LLMs

“Context Engineer” sounds like a buzzword... because it kind of is. But it does capture a common skill: turning messy, high-dimensional data into a focused 20k-token context through search, reranking, chunking... At ZeroEntropy (YC W25), we just released zerank-1, an Elo-inspired reranker that beats basic hybrid search in real-world RAG pipelines. After getting flooded with requests to share more of our research, we decided to kick off a Discord for folks building at this layer: 🔗 Context Engineers : a space for LLM devs to share research, tooling, and painful lessons about context optimization and retrieval: https://lnkd.in/eGwaMNEQ We’ll be publishing internal research, holding small talks, inviting surprise guest speakers, and sharing experiments that don’t make it into blog posts. First session is this Friday with our CTO Nicholas Pipitone diving into the research behind zerank-1. Who should we invite next?? Tell us in the comments :)

“And the system just works. One of our ML engineers built and launched a full testing framework in a day using the open-source docs from Ray and Dagster. No handholding needed, no custom tools, no waiting around — that’s what good infra unlocks.” High praise for both Ray and Dagster on what you can do with great infrastructure. This is also a fantastic read on how to run ML at scale in Dagster. https://lnkd.in/eV9_YraM

🔧 Designing Observability That Powers Real Reliability A recent academic study argues that true observability in cloud-native systems requires more than just logs. You need distributed tracing, application metrics, and infrastructure metrics working together. At Recursive Loop, we bring those same patterns into your infrastructure: ✅ Tracing to uncover cross-service latency ✅ Metrics to highlight performance and anomalies ✅ Infrastructure visibility to monitor health and scalability Because when your systems aren’t just seen—but understood—your business can be trusted. 🔁 Recursive Loop — Observability Engineered for Reliability #RecursiveLoop #Observability #Tracing #Metrics #CloudNative #InfrastructureHealth

When you're running LLMs at scale, you need more than just theoretical performance data. You're going to want to know how they actually perform under real conditions – throughput, latency, and hardware efficiency all matter. Skipping this step can result in higher costs and wasted resources. Benchmarking early helps you avoid inefficiencies down the line, and it shouldn't take hours to get started. That's where vLLM comes in.

ICYMI: Context engineering anchors AI agents to ground truth. So here's how to upgrade an MCP server to help agents find legit and sourced patterns. By Jon Udell

ICYMI: Context engineering anchors AI agents to ground truth. So here's how to upgrade an MCP server to help agents find legit and sourced patterns. By Jon Udell

Interesting use case for AWS Lambda that we explored at Quesma: sandboxing AI-generated code. We tried WebAssembly first but hit the wall with dependencies and performance. So, we scrapped our experiment for AWS Lambda with Docker containers in an isolated VPC. It gave us predictable execution, network isolation, plus hard  timeouts to prevent running infinite loops. Check out the full writeup from our Piotr Migdał on AWS Fundamentals blog: https://lnkd.in/dnPtab5Q

Lambda has tons of use cases, but one I've missed: using it as some kind of sandbox for running AI-generated code. Lambda's isolation and scaling are a solid fit for this problem. Example: execute generated code inside Lambda, locking it down in a VPC with no public internet. If you need S3 access, it can happen through a VPC endpoint. Lambda’s hard timeouts keep "rogue" code from doing bad things. Dependencies still need to be pinned, sure. But the environment is 100% predictable. Just learned about this through a case study from Piotr Migdał, founding engineer at Quesma: https://lnkd.in/ekWiWCJp