Insights for running a business that doesn’t need you.
Practical operations and AI insights to help your business become easier to run, and less dependent on you.
Page 4 of 10
Lovable Moved Its Own Site Off Next.js. The 98.5% Figure Is the Story.
Lovable has moved its own 42-million-visitor website onto the same server-rendered hosting as its 70 million AI-built apps. The migration is engineering news; the 98.5% lift in traffic from AI tools after pre-rendering is the number owners should look at.
8 min readYour Agent Doesn't Need a Memory Layer. It Needs Its History Back.
A plain retrieval baseline over raw conversation history beat the tested agentic memory systems in six of seven benchmark settings, and took a fraction of the ingestion time. Measure the cheap baseline first.
6 min readOne Box, Five Models, and a Licence Problem
LocalRouter runs chat, speech, image and video on one NVIDIA DGX Spark with a sub-second voice loop. The router is Apache 2.0. The models are not, and the licences decide what a business can ship.
8 min readThe expensive model should stop typing: what the handoff research actually found
Escalating a cheap model's failed work up to a strong one recovers less than half the quality gap and multiplies cost. The favourable direction is the other way round: strong model plans, cheap model executes.
7 min readAn Agent Wrote a TypeScript Compiler in Three Weeks. The Reason Why Matters More Than the Speed.
Cloudflare and VoidZero have taken on tsrs, a Rust port of the TypeScript compiler written almost entirely by AI agents. The interesting part is not the benchmark. It is why the agents could do it at all.
7 min readThe Console Under the TV Now Runs AI: Inside the PS5 Local Model Server
A jailbroken PlayStation 5 runs language models natively on its own GPU at up to 207 tokens a second and serves them to the whole network. What it proves, and what it cannot do.
5 min readThe Harness Beat the Model: What Google's Self-Improving Agents Actually Proved
Google open-sourced a framework where the setup around a model improves itself. A weaker model with a self-improved harness beat a stronger one, and that changes what an owner should check before buying a bigger model.
5 min readMeta's Muse Spark Open: What an 816B Download Actually Gets You
Meta released Muse Spark Open's weights and called it the most capable US open-weight model to date. Independent index data agrees it is strong, and shows it is not the leader: GLM-5.3 and Kimi K3 still score higher.
6 min readYour Software Roadmap Is Now a Free Spec. Redis Proved It in One Week.
When Redis changed its licence, a backed open-source replacement existed in eight days. Roadmaps are now public specifications, and that changes which tools a small business can safely build on.
6 min read