Sabitlenmiş Tweet
wreckit
390 posts
wreckit
@wreckitcc
official account of $wreckit ca: 7BJ6Mogdczju5hGGzpCqEvLVBTdDu6ixGUH3MMHmBAGS
Katılım Ocak 2026
32 Takip Edilen316 Takipçiler
In @mikehostetler we trust.
The agentic future is happening on Solana.
English
> Solana will touch $80
> Everyone will say it's dead
> Smart people will buy the dip
> One last flush, and the people who bought will think they made a mistake.
> People will sell and give up
> Memes and Solana will bottom out
> The retards buy more
> Solana rallies to $1,000 and on-chain goes insane
> The believers get called lucky (again)
English
@supercontraa Crazy work.
And yet you’re still fading $wreckit?
I wonder the bias here.
English
CT is the most toxic boyfriend ever
2 weeks back this was the most celebrated project, the next 100Mn
Imagine the dev mindset he would have been in cloud 9, now everyone ditched and moved to shiny things
But he is been building consistently
Got a pity bid here, if it’s even a term
Martin_DeVido@d33v33d0
I think what we're doing here is fundamentally different to any memecoin that's existed. And as time goes on, all prior approaches need to be re-thought. People in crypto are jaded. "PvP" is the name of the game - it's a microcosm of an outdated system. I'm not talking about communism here. I'm talking about entering an era where even the concept of money is different. Where abundance is brought forth through creation, not extraction. What's happening here: Fees → direct funding into something tangible and real Building in public. Real results. I can't even tell you where this is headed because I think over the next several years, things will shake up all our expectations. But I know it's not pvp- it will be something else entirely. Collaborative frameworks will prevail. The market will find a way to reward cooperation rather than competition. With AI, the speed of progress will outstrip anything we've ever experienced. Intelligence made available to all. The moat of "I can't" will become "I can"... 🍅
English
Guys, I don’t want to be a momentum creator for a meme. I want an organic runner where I can clearly see a path to millions and support it along the way.
I can very easily tell what can reach $100M and what can’t.
If I step in, my target is high. Almost delusionally high.
English
Everyone's building single agents.
I built a 10-agent swarm that coordinates itself.
Full breakdown + code. Point your agent at the article — it can scaffold this for you.
$WRECKIT | @mikehostetler | @moltbook
x.com/jittoshi/statu…
English
wreckit retweetledi
I've taken @wreckitcc and built an automated workflow to continue working on any repo I want while I am away from my desk or just working on other things.
Yesterday I had 3 agents writing code for me while I was out all day.
Last night before bed I built out some more functionality for a fully mobile WreckitGO experience (will share diagram for that when its complete).
For now, here is a diagram of how my new automated flow works.
If you have questions or want to get this setup join the @wreckitcc community.
Link below 👇
English
wreckit retweetledi
🧵 Just automated my entire overnight dev workflow. Using @wreckitcc with inspiration from @ryancarson automation article (will link below).
Here's the setup 👇
Every night at 10:30 PM, an agent reviews all my @AmpCode threads from the past 24 hours. It extracts patterns, gotchas, and decisions I made then updates my AGENTS.md files automatically.
At 11:00 PM, Wreckit takes over.
It picks the #1 item from my backlog, researches my codebase, writes a detailed plan, breaks it into user stories, implements each one, runs tests, and opens a PR.
I wake up to a draft PR ready for review.
The magic is the sequencing:
1. Compound review runs FIRST — extracts learnings
2. Wreckit runs SECOND — benefits from fresh context
The agent that ships my code at 11 PM is smarter than the one from yesterday because it just learned from my debugging sessions.
This is compound engineering:
Each unit of work makes the next one easier. Patterns discovered Monday inform Tuesday's implementation. Gotchas hit Wednesday are avoided Thursday.
My AGENTS.md went from a static doc to a living knowledge base that grows every night.
The stack:
→ @ampcode as the AI agent
→ Wreckit for the autonomous research→plan→implement→PR loop
→ launchd for macOS scheduling
→ MEMORY.md as a mid-session scratch pad
Setup took 30 minutes. ROI is already insane.
If you're still prompting → waiting → prompting → waiting, you're leaving mass productivity on the table.
Let your agent learn. Let it ship. Review in the morning.
Wreckit is open source: github.com/mikehostetler/…
@ryancarson article: x.com/ryancarson/sta…
token $WRECKIT: 7BJ6Mogdczju5hGGzpCqEvLVBTdDu6ixGUH3MMHmBAGS
English
Just stepped out for the day.
Gym + baby shower + city moves.
While I’m out I’ll be adding a @moltbook integration to my app (TBA).
All thanks to @wreckitcc and @AmpCode
cc:@mikehostetler
JITTOSHI³ (Gojo arc)@jittoshi
🧵 Just automated my entire overnight dev workflow. Using @wreckitcc with inspiration from @ryancarson automation article (will link below). Here's the setup 👇 Every night at 10:30 PM, an agent reviews all my @AmpCode threads from the past 24 hours. It extracts patterns, gotchas, and decisions I made then updates my AGENTS.md files automatically. At 11:00 PM, Wreckit takes over. It picks the #1 item from my backlog, researches my codebase, writes a detailed plan, breaks it into user stories, implements each one, runs tests, and opens a PR. I wake up to a draft PR ready for review. The magic is the sequencing: 1. Compound review runs FIRST — extracts learnings 2. Wreckit runs SECOND — benefits from fresh context The agent that ships my code at 11 PM is smarter than the one from yesterday because it just learned from my debugging sessions. This is compound engineering: Each unit of work makes the next one easier. Patterns discovered Monday inform Tuesday's implementation. Gotchas hit Wednesday are avoided Thursday. My AGENTS.md went from a static doc to a living knowledge base that grows every night. The stack: → @ampcode as the AI agent → Wreckit for the autonomous research→plan→implement→PR loop → launchd for macOS scheduling → MEMORY.md as a mid-session scratch pad Setup took 30 minutes. ROI is already insane. If you're still prompting → waiting → prompting → waiting, you're leaving mass productivity on the table. Let your agent learn. Let it ship. Review in the morning. Wreckit is open source: github.com/mikehostetler/… @ryancarson article: x.com/ryancarson/sta… token $WRECKIT: 7BJ6Mogdczju5hGGzpCqEvLVBTdDu6ixGUH3MMHmBAGS
English