The Rabbit Hole Was Worth It: SHS Enters the AI Era

What started as a Marvell deep dive turned into a whole new AI-powered direction for SHS.

📣 Where I’ve Been (Hint: It Involves Python, AI, and a Lot of Coffee)

The world used to move fast. Now it just blinks past you. It was once said that software ate the world. Well, now AI is eating software: raw, no salt, straight from the GitHub repo.

For the past few months, preparing research for SHS (and myself personally) has been a hands-on, manual grind — filings, earnings calls, spreadsheets, rinse, repeat. But something shifted recently. A month ago, I dove deep (really deep) into Marvell Technology (MRVL), an AI infrastructure company hiding in plain sight. Five thousand words later, I realized two things:

1. The investment case for MRVL is compelling.

2. The way I’ve been doing research? Outdated.

It wasn’t just the nature of Marvell’s business — custom silicon, hyperscaler partnerships, compute offload — that lit the fuse. It was the realization that I was spending hours on tasks that could now be partially or fully automated. If AI is going to eat software, I might as well feed it SEC filings for breakfast.

So, I pressed pause on writing and jumped into building.

🛠️ SHS, Meet the Machine: How I’m Bringing AI to Investment Research

Over the past month, I’ve been waist-deep in Python, VS Code, and various flavors of large language models (LLMs). I’ve been wiring up the SEC’s EDGAR database — the raw filings themselves — into a set of AI-powered tools designed to supercharge how I (and soon you) consume and analyze company data.

Here’s the mission:
Build a research assistant that actually helps. No fluff, no gimmicks. Just real signals, faster workflows, and fewer hours lost reading boilerplate risk factors.

✅ I've already built a pipeline that pulls 8-K filings directly from EDGAR.
✅ It summarizes key events (like earnings releases or M&A) using GPT.
✅ It layers in real-time market reaction and links to the original exhibit.
✅ It stores the summaries in clean Markdown format (ready for SHS posts).

In short, I’ve started to automate the very research muscle I rely on to write SHS, and I’m just getting started.

👥 What This Means for SHS Subscribers

This isn't just a backend upgrade for my own research process. I plan to open up these tools to SHS subscribers via a front-end app — think:

• Real-time filing alerts and summaries

• Event-driven insights from earnings, guidance, and M&A

• A searchable archive of SEC summaries, filtered by company or event type

All powered by open data and custom AI workflows, not expensive third-party apps or terminals.

I’m also toying with launching a new companion Substack thread, a sort of “building in public” mini-newsletter, where I document the AI and software side of SHS in real time. Think of it as the research lab behind the polished deep dives.

🧠 What’s Next: Marvell, Taxes, and Publishing Again

So yes, I went dark for a bit, but not without reason. Between personal taxes (K-1s love to party late) and this AI rabbit hole, I chose to prioritize building over publishing. But I see daylight again, and the next wave of SHS content is lining up — starting with that long-awaited Marvell deep dive.

Thanks for sticking with me through the quiet. What’s coming next is sharper, faster, and powered by more than caffeine.

Talk soon,
—Kris

---

Thanks for being part of the Safe Harbor community! Follow me for more insights: LinkedIn | X (formerly Twitter) | Instagram | YouTube

Disclosure: This information is provided for informational purposes only and should not be considered a solicitation or recommendation to buy or sell any securities. The author or entity providing this information may hold positions in the securities discussed. This is not investment advice.