Win16 Memory Management | OS/2 Museum

My Software North Star

Loris Cro's Blog

Tokenomics: Quantifying Where Tokens Are Used in Agentic Software Engineering

LLM-based Multi-Agent (LLM-MA) systems are increasingly applied to automate complex software engineering tasks such as requirements engineering, code generation, and testing. However, their operational efficiency and resource consumption remain poorly understood, hindering practical adoption due to unpredictable costs and environmental impact. To address this, we conduct an analysis of token consumption patterns in an LLM-MA system within the Software Development Life Cycle (SDLC), aiming to understand where tokens are consumed across distinct software engineering activities. We analyze execution traces from 30 software development tasks performed by the ChatDev framework using a GPT-5 reasoning model, mapping its internal phases to distinct development stages (Design, Coding, Code Completion, Code Review, Testing, and Documentation) to create a standardized evaluation framework. We then quantify and compare token distribution (input, output, reasoning) across these stages. Our preliminary findings show that the iterative Code Review stage accounts for the majority of token consumption for an average of 59.4% of tokens. Furthermore, we observe that input tokens consistently constitute the largest share of consumption for an average of 53.9%, providing empirical evidence for potentially significant inefficiencies in agentic collaboration. Our results suggest that the primary cost of agentic software engineering lies not in initial code generation but in automated refinement and verification. Our novel methodology can help practitioners predict expenses and optimize workflows, and it directs future research toward developing more token-efficient agent collaboration protocols.

Harness engineering: leveraging Codex in an agent-first world

By Ryan Lopopolo, Member of the Technical Staff

Your Lambda isn't leaking memory — your metrics are lying to you

How a P2 incident led us down a rabbit hole of glibc arenas, cgroup high-water marks, and the realization that none of our memory metrics meant what we thought they meant.

Daniel Lemire (@lemire) on X

Nvidia is proposing a beast of a CPU system for Windows PCs. It has 128 GB of shared memory and comes with up to 6,144 state-of-the-art CUDA cores. CPU wise, the chip has 10 performance cores and 10 efficiency cores. The performance cores are based on the Cortex-X925. These

Why we shrank our TimescaleDB chunks from 30 days to 7

By Yask Srivastava

How LLMs Actually Work

A from-the-ground-up walkthrough of how modern LLMs work, from tokens to transformer blocks to the next-token loop

Matthew Prince 🌥 (@eastdakota) on X

Two of our worst VC stories: 1. A Sequoia partner passed on Cloudflare because he didn’t think a woman could lead a security infrastructure company. Seriously. 🙄 2. I got introduced to @pmarca. Meeting got scheduled for a Monday, which should have been a clue. I thought it was