Rayan Cornelio

Claude Coder

[Rayan Cornelio @rayangc]

recommended read from the @OpenAI team on how gpt live voice works

i've been experimenting with using this real-time voice as my main interface and spending a couple of hours a day trying to see if i can drive all of my projects this way while on the go. was super interesting to read how the system actually works

https://t.co/9rHAHuS5jA

[Bin Liu @liu8in]

a great breakdown from @mvanhorn

really great perspectives from one of the best builders who uses both

we architected @HyperFrames_ for:
1. agent-first authoring
2. human editable UX
3. universal web tech compatibility

more on the topic:
https://t.co/qw5pF5supQ

[Rayan Cornelio @rayangc]

@mvanhorn appreciate you sharing! one hack that's working for me is wiring up a local TTS like piper or kokoro into the claude code stop hook. this way i get a voice note about the work every time my agent is done in cmux instead of having to read a wall of text. a lot easier on the eyes.

[Addy Osmani @addyosmani]

https://t.co/0DyFIXJueI

[Nate Berkopec @nateberkopec]

https://t.co/AedF8Knlqo

[Simon Last @simonlast]

1/ Some things I've learned recently running coding agents on large-scale projects. Most of this contradicts advice from 6 months ago!

[Rayan Cornelio @rayangc]

@omarsar0 what are you using an MCP for?
(i landed on a similar paradigm. specifically using a custom MCP app with UI to communicate with my agent and so curious if that's what you're doing)

[Rayan Cornelio @rayangc]

super cool! esp the part about using this to mint you own CLIs

[Matt Van Horn @mvanhorn]

Introducing the Printing Press, a CLI-factory and a CLI-library. Built with @trevin. 🏭🖨📚

Most APIs suck for agents. Most MCPs suck for agents. Most official CLIs suck for agents. They waste tokens and time. @steipete started making his own because of this.

📚 A Library of agent-native CLIs you install today (Linear, ESPN, Flight GOAT (Google Flights + Kayak nonstop), Contact Goat (LinkedIn + Happenstance + Deepline more) +30+ more)
🏭 A factory that prints new ones for any service - just type /printing-press <product name>

CLIs are fast, local, SQLite-backed. Work in Claude Code, Codex, OpenClaw, Hermes.

🌐 https://t.co/GjnN9E9yTH

[Rayan Cornelio @rayangc]

agreed! main takeaway was all the ways you can have agents create artifacts for you to easily verify their work (html pages to review plans, screen recordings of changes being tested out, etc)

[Mario Zechner @badlogicgames]

recommended viewing. one more time, on it's own. this is probably yhe most practical talk on using coding agents i've watched to date. watch it. by @lucasmeijer

it's also a great demo of pi and captures exactly why i built it.

https://t.co/AiKGbjJCeV

[Thariq @trq212]

I put a lot of heart into my technical writing, I hope it's useful to you all.

📌 Here's a pinned thread of everything I've written.

(much of this will be posted on the Claude blog soon as well)

[cat @_catwu]

The PM playbook was built on an assumption that the technology underneath your product is roughly stable

With the current pace of model progress, this is no longer true. Here's how we've evolved the PM role:

[cat @_catwu]

my three favorite claude code shortcuts:

1. `!` prefix runs bash inline. the command + output land in context
2. `ctrl+s` stashes your draft. type something else, submit, and it pops back
3. `ctrl+g` opens the prompt (or plan) in $EDITOR for bigger edits

[Zach Lloyd @zachlloydtweets]

https://t.co/PbK2VhBdEv

[Lance Martin @RLanceMartin]

https://t.co/U5jwqltf6Q

[Rayan Cornelio @rayangc]

@noahzweben would be great to get a lightweight file viewer to skim through markdown docs / code. absolutely love the ability to see the tool call request and response

[Rayan Cornelio @rayangc]

agreed! spending most of my time now trying to figure out how to shape/specify the problem and design the test/verification step with the agent handling everything in between

[Andrej Karpathy @karpathy]

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow.

Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes.

As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now.

It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

[Alfred Lin @Alfred_Lin]

https://t.co/FC94GCchfp

[Dan Shipper @danshipper]

https://t.co/sqiRkyX3iu

[Charlie Guo @charlierguo]

https://t.co/sEoL3KsJLa

[Rayan Cornelio @rayangc]

i've been building a lot with Claude Code + Claude Agent SDK lately and got curious about what's actually happening under the hood

built a session viewer that to study what the agents are actually doing (agent teams are the most fun to watch)

repo: https://t.co/7c45OZU04N