The future of remittances is digital—& two stocks are leading the charge.
🔹 Wise $WIZEY: The Amazon of cross-border payments
🔹 Euronet $EEFT: The infrastructure giant
My latest deep dive on the $850B remittance revolution: 👇
$APG
API Group announces new financial targets 10/16/60+
10 billion net revenue
16% EBITDA margins
60%+ of revenue derived from inspection, service and monitoring
Also 3 billion in cumulative FCF through 2028.
Small incremental improvements.
Claude Code in the CLI (command-line interface) is amazing and a game changer. @AnthropicAI is doing great things!
I can not imagine myself working with the Claude web app or #ChatGPT anymore for code changes.
Investment Research + AI = Game Changer!
I'm a financial analyst, not a career dev. Yet with Claude and a vision, I'm building an institutional-grade SEC filing AI.
The bar to build has never been lower — especially if you know where to aim.
In 1997 Dwight Schar, CEO of $NVR, owned 868,000 shares of NVR which was 6.8% of the company.
By 2006, he had sold 45% of his shares and yet his interest had grown to 8.4% of the company.
No one did buybacks quite like NVR.
$DFH
Q1 results
-Homebuilding revenues increased 18%
-Net new orders increased 18%
-Homebuilding gross margin of 19.2% vs 17.8%
-Net income attributable to DFH of $55 million, or $0.55 per share
-Repurchased 284,564 shares for $7 million in Q1
9,250 closing expected in 2025
Interesting features I am learning about @OpenAI through my GPT conversations / threads. (Sorry this is a bit off topic of investing but what I'm working on under the hood to make investment research better.)
- We can instruct a GPT to maintain cross-thread link tracking adding only a few tokens of context window per entry.
- GPT can report at what % of 4o context my backfill logic operates.
- A GPT cannot "see" other threads but providing thread URLs can allow it to store it in active session memory, effectively building a manual "conversation map".
This can all be performed with appropriate prompting.
Still deep in development mode for building retail investor friendly tools (including for myself).
OpenAI has made significant progress in sharing memory and tracking across conversations in the last month or so. And it's accurate!
Managing multiple GPT conversations for a single project is messy.
So I built a custom Conversation Tracker inside ChatGPT — organized Kanban-style: Immediate, Medium-Term, and Reference threads.
Agentic AI like Replit isn’t quite there yet, and browser extensions are still clunky (and a ToS gray area).
All for finance and investing use of course.
YOU WERE ROBBED.
50 years from now everyone will know how the crime was committed.
One, by, one ALL large retail stores will be relieved of their value.
This disease was evident with Toys R Us and now Joann’s Fabrics.
Joann’s was wildly profitable until the crime took place.
$CMG is now $20 person with guac, chips and a drink, at least where I live.
No wonder comps are down.
“Casual diner” prices with “fast casual” quality.
Gotta love when you start writing about something and then you get sucked into a rabbit hole. Happens all the time.
All of the sudden there's so much to write about and so little time.
Disappeared for a bit.
It started with a 5,000-word deep dive on Marvell $MRVL.
It turned into an AI-fueled overhaul of how I research and write SHS.
Tomorrow, I’ll share where I’ve been — and what’s coming next.
💡 One underrated LLM hack?
Start a new chat.
Long conversations dilute context. Shorter, focused re-prompts give the model a clean slate — and often sharper, more relevant outputs.
Don’t get lazy with one long thread. Modular chats = modular thinking. 🧠⚡️