Tony Bai's tech world (tonybai.com). Not satisfied with just "knowing how", we strive for mastery. Focused on Go language internals, high-quality engineering practices, and cloud‑native architecture, exploring cutting‑edge intersections of Go and AI. Gophers who pursue technology are welcome—follow me and evolve with Go.
At a three‑day Thoughtworks summit in Engelberg, senior engineers agreed that AI‑driven code generation has become abundant, while the true bottleneck now lies in verification, governance, and maintaining human judgment across legacy modernization, token economics, and organizational practices.
The Model Context Protocol (MCP) released its most extensive revision on July 28, 2026, removing all stateful handshake and session mechanisms, introducing multi‑round‑trip requests, aligning authentication with OAuth 2.0/OIDC, deprecating legacy features, and seeing Anthropic’s Claude be the first product to adapt the new spec.
The Go Collections working group’s umbrella proposal for Go 1.28 adds seven generic collection components—including hash‑based Set, ordered tree Map, and a revamped generic heap—addressing long‑standing gaps in the language’s standard library and outlining design trade‑offs and future extensions.
When AI writes most of the code, the promised speed gains mask a deeper issue: current programming models, trained only to pass tests, fail to learn maintainability, leading to declining PR quality, rising bugs, and the collapse of "lights‑off" software factories.
Two years after its debut, GoMLX has grown from a proof‑of‑concept into a production‑ready Go machine‑learning framework with a modular compute engine, four core abstractions, multi‑backend support (XLA, pure Go, DarwinML), ecosystem bridges to HuggingFace and ONNX, and new features such as KAN, VNN, gradient checkpointing and experimental distributed training.
Anthropic engineers used Claude Fable 5 and Claude Opus 4.8 to migrate massive codebases—Bun from Zig to Rust and a Python project to TypeScript—in under two weeks, cutting memory usage by 91%, reducing binary size by 19% and boosting performance by up to 5% while outlining a reusable six‑step workflow‑centric methodology.
The article analyzes how AI agents can generate code at massive scale, creating "dark" software factories that accumulate comprehension debt, and argues that verification, back‑pressure limits, and a human‑owned outer loop are essential to keep the system reliable.
Buzz is Block’s open‑source, Nostr‑based collaboration workspace that gives each AI agent a cryptographic identity, is model‑agnostic, integrates chat, code repositories and automated workflows, and demonstrates how identity and coordination become the primary bottlenecks when scaling AI agents in teams.
Filippo Valsorda proposes a PHC‑style one‑line Passkey record format that reuses WebAuthn authenticator data, eliminates the need for cross‑account Credential ID uniqueness checks, and includes a draft stateless crypto/passkey Go 1.28 API detailing registration and login flows.
The article analyzes the rapid transition from Loop Engineering to Graph Engineering in AI agent development, explains why simple feedback loops fail, outlines four failure modes, shows how graph structures address them, and provides a detailed 14‑step roadmap with practical examples using Claude Code.
