Raymond Christopher
@RayChris91
Followers
1K
Following
16K
Media
648
Statuses
14K
Christ-follower AI builder, 🇮🇩 ex-Googler
Bekasi
Joined August 2009
The daily AI Lab battle in X 🔥 You never see something like this elsewhere.
0
0
0
Anti-AI programmers will laugh but reality shocks harder than getting electrocuted ⚡
0
0
0
goated move thanks @Google
I am happy to share that we (the @GoogleAIStudio team) are now a sponsor of the @tailwindcss project! Honored to support and find ways to do more together to help the ecosystem of builders.
0
0
0
Real ones know the gold mine here.
Great change. These numbers make way more sense to me. (Minus Flash 3, ignore that)
0
0
0
You won't realize this is the reason why you are addicted to Claude
github.com
Environment Claude CLI version: 1.0.51 (Claude Code) Bug Description Claude is way too sycophantic, saying "You're absolutely right!" (or correct) on a sizeable fraction of responses....
0
1
0
Open Code @opencode 💪
0
0
0
Anyone questioning this must be new to using AI.
0
0
0
Unpinned and uninstalled @windsurf just because it's 100% no longer relevant.
0
0
0
I've tried something similar recently but there definitely a lot of thought and setup going into this approach. This approach is not for the faintest heart.
There's an AI coding approach that lets you run seriously long-running AI agents (hours, days) that ship code while you sleep. I've tried it, and I'm not going back. It's the Ralph Wiggum approach. Here it is in a nutshell: Run a coding agent with a clean slate, again and again
0
0
1
I'll pay for whatever @thdxr is building here
fyi i'm working on this as usual primitives first - you can bring whichever approach you want: worktrees, docker, cloud machines, whatever it's very hard, esp to do it without requiring a central server and have things work even if your laptop turns off
0
0
4
People rarely say this but the best thing about interfacing with LLM is you don't need exact commands (like -- flag) and it forgives non-ambiguous typos. It saves time to not bother to rewrite or correct your prompt if you can still infer the message as a human.
0
0
1
And this is what the end goal of using coding agent looks like.
Codex just spent like 6 hours debugging CI while I played outside with my kids
0
0
1
The ability to adapt with the latest tech is the real moat nowadays, not your years of experience.
0
0
1
The 4 horsemen of Agentic Coding: - specify clear goals - break down Spec > tasks - deliver review/feedback - give proper contexts throughout The lower the (manual) touch, the higher the reward.
When you see how people use Claude Code/Codex/etc it becomes clear that managing agents is really a management problem Can you specify goals? Can you provide context? Can you divide up tasks? Can you give feedback? These are teachable skills. Also UIs need to support management
0
0
0
The reason why my laptop will never go to sleep again unless not plugged in.
always rough when checking on your agent and it turns out your laptop went to sleep
0
0
0
The moment Codex spawns sub-agents it's gonna be 🔥 Not that you can't achieve what it aims to do now (just spawn a new terminal window), but a mere inconvenience.
you can just ask claude to execute plans using as many task subagents as you need, with clear separation of concerns. tested it with up to 10 subagents and it literally 10x's the execution speed
0
0
1
Active Googlers are allowed to use Claude Code, meanwhile some offices block ChatGPT.
Claude Code built in an hour what took a Google team a year. That part isn’t shocking. What is shocking is that Google allows their engineers to use Claude Code instead of forcing Gemini, Gemini CLI, or Antigravity. Giving engineers access to the best AI coding tool is the best
0
0
1
Accurate representation of Goatsen
not my meme, but still lol'ed!🤣 I like the chesscom logo there also😂😂😂
0
0
0