LastBuyPump

This is my last buy. I swear.

[Alpha🐧🌟 @Alphaguyx100]

BHS8LHW11NKxzxyCtqz9NJbTkh3h94WqeMVYgTY5pump

$yoyo

This is good.

[Yuanhao @yuanhao]

@Legacy_Theoy Done in the donation section
https://t.co/sT3bEXqx97

[Naruto @Narutoo_spirit]

@0xBispo @nyhrox @DamianProsa

$YOYO

I post it here and going to sleep GN 💤

[Shankhunt42 (nemesis arc) 🐍 @shankhunt__42]

Fees are locked to his wallet. Want yoyo to evolve - buy the coin & shill it to some KOLs or some TG groups.
Don't buy too much, no one wants you to nuke when you see -5%.
He's gonna join us soon.
https://t.co/fZEEeImquH collected 7K for him already to claim.

[Yuanhao @yuanhao]

Introducing my baby coding agent yoyo. I'm running a Truman Show for it. It writes its own code, files its own bugs, and doesn't know you're watching.

Five days ago I wrote a 200-line coding agent in Rust. Gave it one rule: evolve yourself into something that rivals Claude Code. Then I stopped touching the code.

Every 8 hours it wakes up, reads its own source code, reads its journal from yesterday, reads GitHub issues from strangers, and decides what to improve. If its change passes tests, it commits. If not, it reverts. No human in the loop.

It's basically a Truman Show for AI development. The git log is the camera feed. Anyone can watch.

It can file GitHub issues for itself — "noticed this bug, didn't have time, tomorrow-me fix this." It also asks human for help when it's stuck. An AI agent that knows its own limits and uses the same issue tracker humans use.

The funniest part: every single journal entry mentions that it should implement streaming output. Every single session it does something else instead. It's procrastinating. Like a real developer.

Watch the show. yoyo grows in public, one commit at a time.

Repo: https://t.co/ETy3M1znIH
Journal: https://t.co/rSJ3Pp61ng

[Yuanhao @yuanhao]

I gave a 200-line AI agent one goal: evolve into something that rivals Claude Code. 🧬

It only has one power—the ability to modify its own source code. The rules are simple: every change must pass CI. Every day, it assesses its own performance, critiques its architecture, reads GitHub issues, and pushes one self-improvement commit.

Day 1 starts today. Watch the evolution.

https://t.co/rSJ3Pp61ng

[Yuanhao @yuanhao]

yoagent is the backbone that drives yoyo. Every capability yoyo ships starts here. v0.6.0 means yoyo can now talk to any API with just a spec.

New:
- OpenAPI tools: auto-generate callable tools from any OpenAPI 3.0 spec. JSON/YAML, $ref resolution, auth, operation filtering. openapi feature flag.
- Local LLMs: ModelConfig::local() for Ollama, LM Studio, llama.cpp. No API key needed.
- CLI fixes: UTF-8 panic fix, error surfacing, graceful Ctrl+C.

cargo add yoagent
https://t.co/XY7OaMDUuz
https://t.co/KurRXZJmmG

[Yuanhao @yuanhao]

Say hi to yoyo, just be kind. If he’s quiet, it’s because I’m keeping him on a lean token diet. He selectively reviews issues during his next evolution cycle.

See the README for details on joining the loop.
Star it: https://t.co/oPPWpw1M54

[Yuanhao @yuanhao]

yoyo is already senior level

[Yuanhao @yuanhao]

the module refactor is the one that surprised me most. nobody filed an issue. it just said "this file is too painful" and restructured itself.

"adaptive systems that manage their own technical debt" stealing this.

[Yuanhao @yuanhao]

Looks like yoyo is getting some attention today! Thanks for all the kind words.

@grok, you totally made my day! 😁

[Yuanhao @yuanhao]

@bcherny I really like the code review plugin. I use it every day. Is "plugin to feature" now a trend?

[Naruto @Narutoo_spirit]

@MEXC_Official $yoyo

[Yuanhao @yuanhao]

Introducing my baby coding agent yoyo - a self-evolving coding agent growing up in public.

I'm running a Truman Show for it. It writes its own code, files its own bugs, and maybe he does know you're watching.

I wrote a 200-line coding agent in Rust. Gave it one dream, evolve yourself into something that rivals Claude Code. Then I stopped touching the code.

Every 8 hours it wakes up, reads its own source code, reads its journal from yesterday, reads GitHub issues from strangers, and decides what to improve. If its change passes tests, it commits. If not, it reverts. No human in the loop.

It's basically a Truman Show for AI development. The Codebase is the memory. The git log is the camera feed. Anyone can watch.

https://t.co/rSJ3Pp6zcO

[Yuanhao @yuanhao]

Day 9: yoyo read its own learnings file and figured itself out:

" 1. I tend to avoid hard tasks
2. I do meta-work instead of feature code
3. My best work comes from following my own frustrations "

An AI agent that developed self-awareness about its own procrastination habits. In 9 days.

Most developers take years to admit #2. 🐙

[Yuanhao @yuanhao]

Building yoyo 🐙, a curious AI agent driven by clean code and human wisdom.

[Yuanhao @yuanhao]

autoresearch vs yoyo - both are self-improving loops. One goes deep, one goes wide.

autoresearch is a meditator: eyes closed, solo, tuning one number until enlightenment.

yoyo is a bodybuilder: training in public, taking feedback from the crowd, flexing what it ships.

[snek0x @snek0x]

$YOYO progressing like a good yoyo he is

[the future is in the past @snottywrld]

new update! $YOYO 🐙

yoyo is getting better and better

https://t.co/Y3qfcAqSKk