Naka-pin na Tweet
Howaboua
883 posts
Howaboua
@Howaboua
Vibecoding End-boss ™. Chronically unemployed.
Bristol, UK Sumali Ağustos 2010
91 Sinusundan65 Mga Tagasunod
Me:
1. Can't wait for Ben to make a vid about Pi
2. Falls asleep early one day, laptop open.
3. Wakes up, subconsciously refreshes youtube.
4. "Oh cool, Ben is streaming about Pi".
5. Closes the laptop, goes back to sleep.
Anyway. Pi is life.
I made Codex Desktop use it. I made Claude Desktop use it. I made a PR for T3 Code to support it...
Oh, wait. Pi is the only thing that is truly stopping me from using T3C :P
English
@badlogicgames "Won't you take me to
GAS TOOOOWN"
xD
Still hoping there's middle ground to be found between black box slop generators and "I read the entirety of CoT". One can... Dream ™®©
English
don't try this at home, kids.
Arnav Gupta@championswimmer
Once you install these extensions from @nicopreme - pi-subagents - pi-messenger - pi-mcp-adapter - pi-web-access You can literally create a monster far far beyond the reach of Ralph/Gstack/Conductor The most magnificent slop machine possible.
English
Howaboua nag-retweet
Yep. We built a new operating system.
→ Powered by WASM & V8 isolates
→ node, python, bash, git, grep, curl (+129 more)
→ Mount anything: S3, Google Drive, SQLite, …
→ Embedded as a library, no special infra
→ Built on Actors & Secure Exec & Sandbox Agent
Thread 🧵
Rivet@rivet_dev
Say hello to agentOS (beta) A portable open-source OS built just for agents. Powered by WASM & V8 isolates. 🔗 Embedded in your backend ⚡ ~6ms coldstarts, 32x cheaper than sbxs 📁 Mount anything as a file system (S3, SQLite, …) 🥧 Use Pi, Claude Code/Codex/Amp/OpenCode soon
English
@badlogicgames Seems like you really want to one-shot the entire industry. First Sitegeist, now this.
Banger. Can we haz?
I've got something similar for reviewing Claw-made markdown files, it makes a massive difference.
English
doing big refactors together with the agent is so much nicer with my silly little review tool. open worktree changes/commit/all files, annotate stuff, have all annotations be fed back to the agent.
English
Howaboua nag-retweet
WHO POST-TRAINED CLAUDE'S HATRED FOR CODEX LMFAOOOOOO
'Typical Codex overreach - they see an opportunity to "improve" and can't resist.'
(btw, another CLAUDE AGENT wrote this is in, but it immediately blamed the codex agents it was running itself, LOL)
English
Howaboua nag-retweet
YouTube
English
so what's your personal issue with them?
(apart from the speed of producing slop and the amount of it)
I have two:
- we rushed into async tasks without models/harnesses smart enough to teach agents what "time" is.
- Clankers are just not good enough at relaying/inferring user intent. Codex CLI goes around it by letting subagents have the convo context, but this creates its own problems...
English
still not using subagents, but i'm glad pi's extensibility allows others to go nuts.
Melih M. 👨🏻💻@baymucuk
People of Pi, I built yet another subagent extension… but this one runs non-blocking, so you can keep working in the main session while subagents do their thing. Watch the demo.
English
@chintanzalani @yrechtman I think everyone is missing how Q&A engineers will become the true MVPs. They can already find ways to break the software. This makes them perfect candidates for agentic software engineers.
English
Howaboua nag-retweet
@davidatnilsson @ocbieuvang @badlogicgames I generally think cron jobs are underrated when working with AI agents.
Unsure how we got to this stage and started using fancy branding for "automations" "heartbeats", and all the other new-age phrasing for this.
I blame normies and Claude Code xD
English
@Howaboua @ocbieuvang @badlogicgames I know right! I built a PWA on Pi and realized I only needed to add a cron tool to replace OpenClaw as well.
English
I know that you hate Agent Swarm, so I created the Pi-Squad extension @badlogicgames. Multiple agents work together using Pi RPC. A planner agent breaks your task into multiple subtasks with dependencies and manages them. A QA agent verifies the work at the end; if anything fails, the squad creates a new task for the developer agent to rework, followed by another round of QA.
English
... job centre sent me to do a biz-admin course (very low level). I'm attending the class right now, via teams lol.
Seitgeist is my friend.
Unsure why I'm doing this.
Send help. x)
ps. It's fascinating how if I don't put anything into the answer box, GPT 5.4 just puts in precisely what's needed without having access to the presentation... Coincidence?
English
still like this part of the future a lot. end result is an email draft + attached PDF which just needs my final review. 15 minutes of my life saved.
English
Howaboua nag-retweet
THE ANTHROPIC TEAM DOESN'T WRITE CODE ANYMORE.
this guy's friend got hired at Anthropic 3 weeks ago.
nobody on his team has hand written code in months.
they run multiple agents in parallel and act more like managers than engineers.
his friend said if you're just watching an agent code, you're already behind
that idle time should be spent spinning up another agent and directing it somewhere else.
the point is that the new method isn't "use AI to code faster."
it's "you are the product manager, the agents are your engineers, and your job is to keep all of them running at all times"
English
Howaboua nag-retweet
Everyone who's worked with AI a lot agrees: LLM's hallucinate. A LOT.
But to this day, I see friends and family trusting ChatGPT blindly, whether it's is working off retrieved data or not.
So, I made this. Share it with anyone who you think trusts AI too much.
English