Gemini 4 Argon is available to Google AI Pro and Ultra subscribers globally at https://t.co/vnZpmNUoAX and in the app. Learn more in this 🧵 https://t.co/rTKOshApho
D*MN, MY GROK BOTS HAD THE WORST START OF THE WEEK AND STILL TURNED $14,670 INTO $33,775 BY THE END OF DAY 2
that’s +130% in one day. 3 desks, one screen, Jev picking the coins
day 1 ended at $14,670. i didn’t touch anything, just let the desk run
the day:
→ opened $14,670
→ first hour slow, $15,121
→ Jev bought Moodeng, bank $17,349
→ then a red stretch, back down to $16,355. i almost shut it off
→ the bots bought the dip on Giga, $22,183
→ $23,023, then another pullback to $21,750
→ Popcat and SPX carried the afternoon, $29,711
→ closed at $33,775
the 5 coins the desk lived in all day: SPX, POPCAT, MOODENG, GIGA, PENGU
the funny part, the bots talk while they trade. one typed “moodeng is my spirit animal” right before buying it. another saw a red candle and said “dip? i see a discount”
honestly the passes impressed me more than the buys. Jev skipped way more coins than it touched
the rules stay boring on purpose:
→ no pool under $40K
→ no chart that’s already vertical
→ 2 losses in a row and the bot is out
→ a post on X is research, never the trigger
at 0:06 the bank drops to $16K and the chart goes red. at 0:09 the whole desk swarms Giga. at 0:21 it closes at $33.7K
the setup and every prompt are in the article below ↓ free
would you have shut the bots off at the first red stretch
F*CK RUNNING OPUS ON EVERY TURN
Anthropic just showed the cleanest Claude Code advisor setup.
Sonnet 5.5 builds everything.
Opus 5.5 sits on call and only wakes up when shit actually breaks.
Sonnet runs the main session, edits, and runs tests.
Three Haiku 5.5 subagents do the dirty work:
→ search maps the repo
→ grep finds call sites
→ docs pulls the API refs
JEV sits underneath and resolves thousands of micro-forks (which file, which tool, retry or stop) before anything reaches Sonnet.
Opus only gets called twice:
→ before a plan (reads the whole session)
→ when the same error hits twice
It never writes code itself.
It just looks at the full transcript and says “wrong fixture, not the code.”
Sonnet applies the advice.
Zero regressions.
Contract clean.
All green.
You launch it like this:
claude --model sonnet --advisor opus
The full tree
> Sonnet 5.5 runs the main session and writes code
> Haiku 5.5 x3 - search / grep / docs
> JEV routes deterministic forks
> Opus 5.5 on call - only on same error x2 or before done
> Sonnet applies the advice and verifies
Paste the tree + this prompt into Claude Code
Rebuild my Claude Code setup around this tree:
Run the main session on Sonnet 5.5 with effort high and --advisor opus.
Delegate search, grep, and docs to Haiku 5.5 subagents.
Route deterministic forks (which file, which tool, retry or stop) through JEV.
Only call Opus when the same error hits twice or before done.
Opus reads the full session transcript and gives advice. Never writes code itself.
Sonnet applies the advice, shows the diff, then verifies.
Show me every change as a diff first. No edits until I say go.
save it, it’ll come in handy.
How safe would American roads be if there were no drunk drivers?
Our new research shows that sober drivers experience a fatal crash involvement rate 23.3% lower than the overall average.
But sobriety alone doesn’t guarantee safety. Roadway fatalities are a systemic issue often driven by low visibility, fatigue, and poor road infrastructure.
The Waymo Driver is never impaired and built to navigate the systemic risks that sober human drivers still face.
Learn more https://t.co/U9UaxkOQoj