The Evolution and Challenges of AI Coding Agents: From Dial-Up to Highway

2025-09-22
The Evolution and Challenges of AI Coding Agents: From Dial-Up to Highway

The rapid advancement of Large Language Model (LLM)-powered AI coding agents has brought unprecedented productivity gains, but also immense infrastructure challenges. Drawing an analogy to the dial-up internet era, the author describes the evolution of AI coding agents from early inefficient and unreliable states to their current widespread use, while still facing high latency and cost issues. The author argues that higher tok/s (tokens per second) speeds are key and predicts the future will see more advanced, less manually-intensive AI coding workflows, and more flexible pricing models to cope with peak loads.

Read more
Development

AI Inference Costs: Not as Expensive as You Think

2025-08-28
AI Inference Costs: Not as Expensive as You Think

This article challenges the narrative that AI inference is prohibitively expensive and unsustainable. By calculating the costs of running AI inference on H100 GPUs, the author demonstrates that input processing is incredibly cheap (fractions of a cent per million tokens), while output generation is significantly more expensive (dollars per million tokens). This cost asymmetry explains the profitability of some applications (like coding assistants) and the high cost of others (like video generation). The author argues that this cost disparity is often overlooked, leading to an overestimation of AI inference costs, which may benefit incumbents and stifle competition and innovation.

Read more