sumeetfx

[sumeetfx @sumeetfx]

One thing that took me years to understand was that in crypto you are always chasing the 100x. When in reality, you are most likely not going to catch. We see 2x-5x regularly and think ‘oh ill wait till 10x-20x before taking profit’ and even then we don’t take profit.

[sumeetfx @sumeetfx]

@Gioom @bored2boar Me and you both mate, on paper it looks good. Then when live, not so good! The marathon continues!

[cvxv666 @antpalkin]

His old model took a week to backtest one idea. Claude Fable 5 does it in two minutes - so he ran 600 over past days and walked away with $90,000 profit.

The edge was never having a good idea. It's killing the bad ones fast enough to find the one that isn't.

His wallet link: https://t.co/VKWouVbpbO

Here's the loop.

Feed Fable 5 a rule - "buy UP side when the book leans 70/30 in the final 90 seconds"

It replays that idea tick-by-tick across thousands of real resolved markets - real fills, real slippage. Two minutes later: dead or alive.

He ran 600 ideas in 48 hours. 597 died. Three printed.

A hedge-fund desk tests maybe 20 ideas a quarter. He tested 600 in a weekend - because the brain doing the testing dropped from a week per idea to two minutes.

The three survivors now run live on MiroFish, sized by Kelly, firing only when the simulation and the order book disagree. $90,000 in one day.

He didn't get smarter than Wall Street. He got 5,000x more shots on goal.

The bot's live right now, taking those same three setups. Two clicks and you copy every trade he made: https://t.co/vbDZyVbI3v

[Gipp 🦅 @gippp69]

THIS GUY CONNECTED CLAUDE TO TRADINGVIEW AND CUT 200+ HOURS OF MANUAL CHART WORK

40 trades a week used to mean the same loop: open the chart, check RSI, compare EMA, write notes and wait for the next setup

now Claude reads TradingView 24/7, tracks volume, indicators and price conditions while he is not even at the desk

this is not a $20k trading terminal and not a quant fund setup, just one laptop, one chart and one strategy written in plain English

the crazy part is not that Claude predicts the market, it is that a regular trader can remove the boring human loop from every setup ↓

[Miles Deutscher @milesdeutscher]

Anthropic posted a FULL GUIDE on how to prompt Fable 5 (Mythos).

Claude Fable 5 is not meant to be prompted like any other model.

It's meant to run autonomously.

Here's exactly how to enable Fable to do work for you with minimal manual intervention:

1. Effort selection

Anthropic recommends using High for most tasks and Xhigh only for complex workflows.

Low/medium: quick questions, basic research
High: default for most work
Xhigh: complex builds, multi-step analysis
Ultracode: full autonomous orchestration

2. /loop prompting

Use /loop prompts to kick Fable off to complete full tasks.

/loop <time interval> + <goal>

3. Tell it WHY, not just what (context)

Fable can't perform on instructions alone. It needs context to make decisions on its own.

Anthropic's exact prompting structure:

"I'm working on [larger task] for [who it's for]. They need [what the output enables]. With that in mind: [your actual request]."

4. Keep prompts short (instructions)

Counterintuitive but critical.

Over-engineering your prompts on Fable 5 degrades output. You're constraining a model that would have figured it out on its own.

4. Tell it when to stop and check in during runs

"Pause for me only when the work genuinely requires my input: a destructive action, a real scope change, or something only I can provide. Otherwise, keep going and report back when done."

5. Build it a memory system

Fable performs best when it can record lessons from its previous loops.

Give it a markdown file and this instruction:

"Store one lesson per file with a one-line summary at the top. Record corrections and confirmed approaches. Don't save what the repo or chat history already records."

The optimal general prompt structure:

"Goal: I'm working on [larger task] for [who it's for]. They need [what the output enables].

Request: [your specific ask in one sentence]

Output format: [exactly how you want it]

Constraints: [what must not happen]."

One last thing - your old prompts may actually work against you.

Skills and project instructions built for Opus 4.8 may produce worse results on Fable.

Bookmark this to actually maximize your Fable workflows.

[Crypto Rover @cryptorover]

Claude Fable (Mythos) can fully control your tradingview.

It can automatically trade for you 24/7.

Manual traders are completely cooked.

[Mr. Buzzoni @polydao]

https://t.co/E8GJ7yfZEM

[Himanshu Kumar @codewithimanshu]

Claude is now controlling TradingView live from my terminal.

Switching symbols. Writing Pine Script. Batch scanning futures. Replay trading. Drawing levels.

All autonomous. Zero clicks.

Still has rough edges but the vision is crystal clear.

I told it:

Find me every BTC futures contract with RSI below 30 and volume spike above 200%.

14 seconds later:

→ 6 contracts identified
→ Charts loaded
→ Support levels drawn
→ Pine Script backtests running
→ Entry zones marked

Didn't touch the mouse once.

Then I said:

Replay last week. Show me where your system would have entered.

It switched to replay mode. Scrolled through price action. Marked every edge. Calculated P&L in real-time.

$4,780 theoretical profit from 9 trades.

83% win rate.

Now it writes custom Pine indicators on command:

Build me a momentum oscillator that tracks whale wallet activity correlated with price.

40 seconds. Script deployed. Indicator live on chart.

Most traders are still clicking through 50 charts manually.

Claude scans 200+ in under a minute.

Finds the setups. Draws the levels. Backtests the edge. All while you watch.

This is not about replacing your strategy.

It's about executing it 100x faster.

You only need Claude + laptop + 1 hour/day.

Giving This Free for 24 hours. To get it:

1. Comment the word CLAUDE
2. Like and Retweet this post
3. Follow me @codewithimanshu (so i can DM you)

Save this post. Deploy this setup this weekend. Start testing. Scale on evidence.

[AI Edge @aiedge_]

https://t.co/B6UPWQyTM3

[monokern @monokern]

https://t.co/RmlXgqYeJg

[Miles Deutscher @milesdeutscher]

what the f*ck did Anthropic just release?

You can now spin up HUNDREDS of Claude subagents?

This new feature is the most powerful feature Anthropic has ever shipped.

Dynamic workflows run scripts that can deploy tens to hundreds of parallel subagents in a single session.

Here's how to deploy these new dynamic workflows (explained in 60 seconds):

1. Open Claude Code CLI, Desktop, or VS Code extension

(must be on Max, Team, or Enterprise plan)

2. Turn on Auto Mode

This is required for the best experience with dynamic workflows.

(shift + tab to cycle)

3. Pick your starting method:

Option A: Ask Claude directly

"Create a workflow for [your task]."

Option B: Turn on Ultracode

Found in the effort menu. Sets effort to xhigh and lets Claude decide automatically when to spin up a workflow.

4. Confirm the run

The first time a workflow triggers, Claude shows you exactly what's about to run and asks you to confirm before anything starts.

5. Let it run

Claude breaks your task into subtasks, fans them out across parallel subagents, checks every result before folding it in, and hands you back a single

The best use cases right now:

- Codebase-wide bug hunts and security audits
- Large migrations touching hundreds of files
- Any critical work needs to be independently verified before it reaches you

Work you'd normally plan in quarters now finishes in days.

Save this and deploy your first workflow now.

[Ronin @DeRonin_]

Do you understand what Anthropic just shipped to Claude Code???

one prompt = 100 agents working at once + a second team trying to prove them wrong.. you only see what survived

[ how it actually works ]:

> you say "workflow" in a prompt
> Claude writes an orchestration plan and follows it strictly
> 10s to 100s of subagents work in parallel
> adversarial agents try to break their work before you see it
> progress saves as it goes.. interrupted runs pick up where they stopped

the plan is the contract

every agent before this had ATTENTION ISSUES.. it tracked the first few things, drifted, and shipped something different than you asked for

you've been QA-ing your own agent's work 100 runs in a row.. and not noticing (just chilling-out)

[ literally what every Claude user STOPS dealing with ]:

- tasks too big for one session (you used to stitch the pieces back yourself)
- agents drifting halfway through long runs (every multi-hour task, predictably)
- double-checking every output yourself (a side job nobody admits to)
- crashes wiping out hours of progress (back to square one)

[ the Bun rewrite is the proof ]:

Jarred Sumner used dynamic workflows to port Bun from Zig to Rust

> 750,000 lines of Rust
> 99.8% of the test suite passing
> 11 days, one engineer
> hundreds of agents in parallel, two reviewers on every single file
> (not in prod yet, but the rewrite works)

an entire language port a team would have spent 6 months on.. done in 11 days by one person

iteration = parallelization

the agent doesn't just work harder.. it deploys an entire team of itself, plus a second team trying to break what the first one made

[ the math ]:

repo-wide security audit:
- old: 1 agent, 1 file at a time, half a day
- new: 100 agents fanned out, 11 minutes, adversarial verification included

A/B flag cleanup across 100s of flags:
- sequential: hours of waiting
- workflow: <10 minutes, all parallel

one run pays for the week of human work it replaced

RESEARCH PREVIEW, AVAILABLE TODAY, ULTRACODE TURNS IT ON

I tried it 50 mins ago on a dead-code audit across one of my repos

found 47, fixed 43, opened PRs for the other 4
14 minutes (yes it burns tokens.. Anthropic literally warned about that.. still pays for itself)

P.S. it's the first agent feature in a while where I caught myself thinking "the bottleneck just moved off my own brain"

the whole night gonna spend on tests of new features in Claude Code and Opus 4.8

if you love my observations and explanations, turn on notis and stay with me during this night

or you can skip this, it's your deal ❤️

[Ruben Hassid @rubenhassid]

https://t.co/FBQx2r61sQ

[HandofArsenal @HandofArsenal]

Greatest picture I have laid eyes on

[Codez @0xCodez]

https://t.co/cLtV9sv1od

[Khairallah AL-Awady @eng_khairallah1]

Boris Cherny, the creator of Claude Code at Anthropic, just explained why most people aren't getting real results from Claude

in this podcast he breaks down exactly how most people never actually set up Claude:

- the 14% you lose to CLAUDE.md before typing a word
- the features that change how Claude thinks before you type a word
- the settings 95% of users have never opened
- the workflows hiding behind one toggle

if you've been using Claude for more than a month and never left the chat window, you have at least 30 untouched features. probably 38

instead of another show tonight, watch this

make sure to bookmark it before it gets lost in your feed

my breakdown of all 40 features is below

[Bukayo Saka @BukayoSaka87]

You deserve more Arsenal fans 💔

[Polymarket @Polymarket]

JUST IN: Google unveils Gemini 3.5 Flash, claiming it can code at 4x the speed of comparable frontier models.

[Gym Fiesta @GymFiesta]

Runners… don’t ignore your plyos ⚡ These are the drills I personally use to build more speed, power, explosiveness &amp; stronger runs. Just running by itself wasn’t getting the job done for me.

[Aria Westcott @AriaWestcott]

Stop telling your AI, "fix this error."
Stop telling your AI, "do this."
Stop telling your AI, "write code."

You're actually treating a senior AI like a junior intern.

8 prompts you can copy and paste directly:

[Miles Deutscher @milesdeutscher]

https://t.co/042ZheWx7F