it comes with its own temporary friction, torch.export is still maturing for backward/training graphs and all effort over the past two years was put on inference maturity
CODE
-
torch.export: Compile-time autotuning for training without runtime JIT
By
–
if you want to avoid guards and do autotuning at compile-time, have no jit at runtime, consider torch.export; thats what you'd want.
in fact, @ezyang has been pushing it for some training workloads that dont want any jitting at runtime and ones that need a bunch of sharding info -
AI threatens coding bootcamps: build instead of learn
By
–
This is gonna kill the coding bootcamp industry lmao, why learn when you can just build
-

Optimizing AI Coding Workflows with System Instructions
By
–
Daily reminder : ~/.codex/instructions.md or ~/.claude/CLAUDE.md make a huge difference! My instructions:
– Don't catch errors, I prefer to raise them and fix them myself
– Go simple
– For GPT-5 in codex: don't abbreviate names (model has a tendency to make confuse -
Custom MCP Connectors Enable Advanced AI Agent Integration
By
–
Huge! and custom MCP connectors too, that's so awesome.
-
AI Tool Access Control and Performance Risk Management
By
–
Most of the time I wouldn't be forking the repo, and I'm personally not comfortable with letting the AI delete my workflow run logs yet. Having these tools made available to the AI all the time creates unnecessary risk – and poorer performance.
-
Claude Code Tool Configuration Requires Manual JSON Editing
By
–
Unfortunately, the solution in Claude Code now is to manually edit the json settings file, and specify the tools I want into a list there; quite a painful experience to have to copy-paste across the 91 tools GitHub has made available.
-
Selective Tool Configuration for Model Context Protocol
By
–
This MCP is great: ton of power tools from adding comments to pending reviews, getting discussions threads, forking the repo, and more. However, in practice, I only want a subset of these tools depending on the task I'm working on.
-

Managing Large MCP Servers Token Usage Efficiently
By
–
Some of these MCP servers are huge! GitHub's server comes with 46.3k tokens out of the box, with a ton of power tools. Here's how I'm managing it: