Ideas for building
better software.

This Week in Spring - August 4th, 2026

Vectorize - Vectorize indexes now support up to 20 million vectors

You can now store up to 20 million vectors in a single Vectorize index, doubling the previous limit of 10 million vectors. This enables larger-scale semantic search, recommendation systems, and retrieval-augmented generation (RAG) applications without splitting data across multiple indexes. Vectorize continues to support indexes with up to 1,536 dimensions per vector at 32-bit precision. Refer to the Vectorize limits documentation for complete details.

Semantic memory search for AI agents

Your AI agent handles a long onboarding conversation. The next day, it asks the same user for their name. That's not a bug. A language model keeps no memory of earlier calls, so without an external memory layer, each request starts fresh and the agent...

When does the A2A protocol actually matter?

If you're building multi-agent systems, someone has probably asked whether you're "doing A2A yet," with the implication that you should be. When teams actually reach for it, most can't say why they need A2A over MCP. A more useful question: do your ag...

New Pass-ta-key attacks let malware hijack Google-synced passkeys

Security researchers have discovered three attacks that allow malware on already-compromised Windows devices to abuse Google Password Manager's synced passkeys to take over accounts, bypass user verification, and extract passkey private keys. [...]

Building a company in 2026: What I’m learning along the way

I have spent the last 18 years working in technology, including around six years as a CTO in startups. Architecture, cloud infrastructure, engineering teams, incidents, and delivery deadlines are familiar to me. This is also my second time as a co-founder, but my role is different now. At CatechLabs, I work across product, sales, customer discovery, positioning, partnerships, and strategy. That shift made something clear: knowing how to build software is not the same as knowing how to build a co

Episode 5 — Who Gets to Flip the Switch

Week 3. "The artifact is sitting in the registry. Somebody still has to actually run it. Who, and how?" Developer ↓ Runner ↓ Cache ↓ Artifact Today ↓ Deployment Junior Engineer: So the image is built, tagged, sitting in the registry. What's the simplest possible way to actually get it running in production? Senior Engineer: The simplest way is also the most dangerous way. Want to guess it first? Junior Engineer: Stop the old container. Start the new one. Senior Engineer: That's

Designing a Backend System That Handles 100K Requests/Second (Without Melting Your Database)

TL;DR Architecture At high level: Global Traffic Routing (Geo DNS / Anycast) Edge Layer (CDN + WAF + Rate Limiting) Load Balancers (L4 + L7) Stateless App Tier (autoscaled microservices) Multi layer Caching (edge, distributed, local) Data Layer (sharded DB + replicas + queue based writes) Async Processing (Kafka / stream workers) Observability + Auto healing Never start with boxes and arrows. Start with numbers. Assume: 100K RPS peak 95th percentile response target: < 150ms Read heavy workloa