Paolo JN

shipping w/ claude code. probably building something rn.

Building → | MMR → █░░░░░░░░░░ 5%

[Paolo JN @PaoloJNN]

Every satellite above us, in real time. ~31,000 of them right now → https://t.co/rIuhQvBOCl

[Paolo JN @PaoloJNN]

Go -> https://t.co/j0iQoqHVdX

[Paolo JN @PaoloJNN]

Try it: https://t.co/Z5k5PHMOua

[Paolo JN @PaoloJNN]

Messing around with pathfinding algos, built this visualizer

A*, Dijkstra, Greedy, BFS, DFS + 6 maze generators

lets you draw walls, generate mazes, and see the algorithm search step by step

[Paolo JN @PaoloJNN]

Building a team of ai agents that do my algo trading research.

They research ideas, backtests them, runs risk audits, and moves the ones that survive into paper trading.

the goal isn't one perfect strategy. it's a pipeline of 10 to 30 small, uncorrelated edges.

[Paolo JN @PaoloJNN]

36 users in 4 days. 2 signed up. 0 paid.

not bad for a first product...
need to figure out the funnel part now

[shirish @shiri_shh]

difference bet Opus 4.8 vs Opus 4.7

[Max Katz @maxktz]

nothing beats my `cmux` + `hitch` workflow...

[Paolo JN @PaoloJNN]

@maxktz How did you get cmux to look like this 👀

[Matt Popovich @mpopv]

the top right corner of vscode is starting to resemble an iq test

[CyrilXBT @cyrilXBT]

All you need in 2026 are these apps to make life changing money

[Paolo JN @PaoloJNN]

wanna to try it https://t.co/ufmcUXP5k7 :)

[Paolo JN @PaoloJNN]

launched my first product... tonight

claude code + me + the chrome web store kicking my ass 4 times... couple of days of work

AI prompts, portable across ChatGPT, Claude, Cursor, everything

PRODUCTHUNT = 50% off →
https://t.co/cQtpgv3FMm

[Paolo JN @PaoloJNN]

@shadcn Codemaxxing

[Paolo JN @PaoloJNN]

Almost got caught in the hype and bought a Mac mini for my 🦞 OpenClaw

then remembered I have a whole gaming PC in my basement doing nothing

saved myself $1,400. lowkey better anyway

[Paolo JN @PaoloJNN]

I realized most people showing profitable backtests are probably just fooling themselves.

The hard part in algo trading isn’t finding a chart that goes up.

It’s figuring out if the strategy survives reality.

[Paolo JN @PaoloJNN]

Don’t fall into the “optimize your workflow” trap.

I keep organizing systems, improving automations, and tweaking my setup instead of shipping.

It feels productive, but it’s still avoidance.

I literally had to update my agent memory to call me out when I start side questing.

[Paolo JN @PaoloJNN]

@Rasmic Hated my life trying try to get openclaw to work …Hermes >

[Paolo JN @PaoloJNN]

Did a little portfolio refresh today

https://t.co/DiUQO45GbZ

Here’s the current design. Still very much a WIP, but it’s finally starting to feel like a real place to show what I’m building.

[Paolo JN @PaoloJNN]

@haasonline Exactly.. Paper trading tests the system. That’s where you find out if the edge is actually real

An 86% strategy means a lot less when fills are ugly.

[Paolo JN @PaoloJNN]

Small breakthrough on my algo-trading side quest:

Found a strategy that backtested well across 1,431 trades from 2015–2026 — roughly 86% win rate after modeling costs.

Still early. Next step is paper trading and seeing if it survives outside the backtest.

[Paolo JN @PaoloJNN]

@theo @shadcn is this real ? wtf