mager

Just keep iterating, it will always get better every turn

@_LuoFuli It’s a creativity explosion 🎨

Imagination is the ceiling of productivity in the new era. Inspiring imagination is the core of management in the age of Claw.

@dontbesilent @garrytan I need a skill that gives me da direct window to Garry’s brain

1、这是我迄今为止用过的最好的 skill 2、无论我们多么大力宣传这个 skill，大部分都只会围观，最多安装一下，不会真的去在自己的上下文里面去推进 3、所以，这让这个 skill 的使用体验更稀缺了

@trq212 Great conversation today with @hudaman! How do Anthropic devs iterate and test a skill locally in the same session?

CC Livestream x.com/i/broadcasts/1…

Are you building a software factory yet? Start small, and keep iterating mager.co/blog/2026-03-1…

@hudaman I created ME.md: loooom.xyz/me I can just paste this in to any clean Claude session to get instant context: loooom.xyz/me/mager/raw

Next step after you get a SOUL.md injected into your agent is to create the HEART.md and GUT.md and so on

I let AI analyze every first round March Madness matchup 🏀 mager.co/blog/2026-03-1…

@whatdotcd @clawcon @msg @jwegener Have fun, wish I could be there

I'm gonna be hosting @clawcon LA with @msg and @jwegener come hang! luma.com/clawconla

Using Skills well is a skill issue. I didn't quite realize how much until I wrote this, the best can completely transform how your team works.

Agentic software engineering adoption is on fire at @Uber. 1,800 code changes per week are now written entirely by Uber's internal background coding agent, and 95% of our engineers now use AI every month across all the tools we track. This is a real reset moment for engineering; it's one of the most exciting times to lead. This shift requires builders to be curious and hands-on. I’m incredibly lucky to be surrounded by a team that’s doing exactly that. The best part is that the strongest adoption isn’t being pushed top down from leadership announcements; it’s coming from engineers who are quietly experimenting, quietly shipping, and quietly pushing things forward. I love spending time with those engineers because there’s no substitute for being close to the work. Over the last few months, we leaned in hard, and the results have been phenomenal. The bigger shift: going agentic. 84% of AI users are now working with agent-style workflows, not just tab completion. Claude Code usage nearly doubled in 2 months (32% → 63%), while IDE-based tools have largely plateaued. Engineers are moving from accepting suggestions to delegating tasks. Even within traditional IDEs, ~70% of committed code is now AI-generated. Background agents are writing code autonomously. Our internal background coding agent went from <1% of all code changes to 8% in just a few months. There is zero human authoring. Engineers review and approve, but the code is written entirely by AI agents. The role of the engineer is shifting - from writing every line to architecting systems and reviewing AI-generated code. More to come from the @UberEng team in the coming days.

It’s linked github.com/mager/claude-t…, check out the post

Build your own Claude terminal interface in 10 min mager.co/blog/2026-03-1…

@RealProductGirl @Lovable Fast design iteration. And you can continue chatting to improve the design. It outputs a working app/prototype.

@mager @Lovable Thank you! Let's see how far I get. I've heard good things about it. Why is it one of your favorite tools?

Day 0 of learning to vibe code as a PM with ZERO software background. Lines of code written in my life: 0 First tool: @Lovable Documenting every win, fail, and everything in between. Follow along to see my daily struggles 🚀 Wish me luck Builders! 👊

@garrytan Are you using promptfoo?

Full evals system coming to GStack shortly. LLM evals are the only way you can make fully agentic systems that are repeatably better as you improve the workflow, context engineering and prompts.

stop spending money on Claude Code. Chipotle's support bot is free:

@pmarca @RalphQuintero Share your top pods

My information consumption is now 1/4 X, 1/4 podcast interviews of the smartest practitioners, 1/4 talking to the leading AI models, and 1/4 reading old books. The opportunity cost of anything else is far too high, and rising daily.

@karpathy Let me know when I can start training on my Mac mini :)

nanochat now trains GPT-2 capability model in just 2 hours on a single 8XH100 node (down from ~3 hours 1 month ago). Getting a lot closer to ~interactive! A bunch of tuning and features (fp8) went in but the biggest difference was a switch of the dataset from FineWeb-edu to NVIDIA ClimbMix (nice work NVIDIA!). I had tried Olmo, FineWeb, DCLM which all led to regressions, ClimbMix worked really well out of the box (to the point that I am slightly suspicious about about goodharting, though reading the paper it seems ~ok). In other news, after trying a few approaches for how to set things up, I now have AI Agents iterating on nanochat automatically, so I'll just leave this running for a while, go relax a bit and enjoy the feeling of post-agi :). Visualized here as an example: 110 changes made over the last ~12 hours, bringing the validation loss so far from 0.862415 down to 0.858039 for a d12 model, at no cost to wall clock time. The agent works on a feature branch, tries out ideas, merges them when they work and iterates. Amusingly, over the last ~2 weeks I almost feel like I've iterated more on the "meta-setup" where I optimize and tune the agent flows even more than the nanochat repo directly.

You should get an extra allotment of tokens from Fri - Sun