Listen in while I do the research with you

Hands-on notes on investment research with AI: the tools, the checks and the misses, all shown. Every post ends with one small thing you can take with you today.

Posts 361Today 1,176Total views 24,783

Original shorts

All videos →

A Little More Every Day: The Turn She Couldn't Land, Landed on Day 30

A 27-second original AI short: mio can't land a turn, practices a little more on day 1, day 12 and day 30, and the day it finally lands, someone calls her in to try out for lead.

  • On day 1 the turn won't hold, and she quietly tells herself "One more time."
  • On day 30 it lands and even she is surprised; then a title card: 1% better every day is 37.8x in a year.
  • Someone at the door calls "Mio, come try out for lead," and she turns: "Me?" No single day made the difference; a little more each day added up until others could see it.

Read the full article →

Latest posts

Building a Moving Espresso Machine in a Day: Drawing It Part by Part, and Simulating a Shot With Four Formulas
Research Notes

Building a Moving Espresso Machine in a Day: Drawing It Part by Part, and Simulating a Shot With Four Formulas

On October 14, 2026 I replaced the blog's espresso lab with a 110-part machine that moves with the sliders. Here is how it was drawn, how the code checks its own drawing for mistakes, and the four simplified formulas behind the extraction simulation, along with where they go wrong. A learning note, not advice.

Three Shots in One Clip, 24 fps Doubled to 48: An AI Short Film That Took Four Runs
Research Notes7 views

Three Shots in One Clip, 24 fps Doubled to 48: An AI Short Film That Took Four Runs

A small experiment log: making the 27-second AI short 'A Little More Every Day' with MiniMax H3 on a Google Colab A100. New this time: cutting three shots inside one clip, pinning storyboard stills to exact frames, and doubling 24 fps to 48 with open-source RIFE. Timings for each step, what broke in each of the four runs, the compute units spent, and whether the dialogue actually came out. The character is an AI creation.

Making AI Short Clips with MiniMax H3 on Colab: Is 2K Worth It?
Research Notes191 views

Making AI Short Clips with MiniMax H3 on Colab: Is 2K Worth It?

On October 12, 2026 I ran the MiniMax H3 video model through ComfyUI on a Google Colab Pro A100, generated 3-second clips at 1344×768, and upscaled them to 2K with Real-ESRGAN. Each clip took 206 seconds end to end with a GPU memory peak around 52GB. After comparing both versions, 768 was enough; the story is what decides whether a clip works. Includes the Colab gotchas I hit, the license terms worth reading, and three routes to 2K side by side. A personal learning note, not business advice.

Reshooting an AI Short Film: Testing Three Quality Settings for MiniMax H3 on a Colab A100
Research Notes27 views

Reshooting an AI Short Film: Testing Three Quality Settings for MiniMax H3 on a Colab A100

A lab note: nobody understood the first cut of our AI short film Daijoubu, so I rewrote the script around four problems, generated a fresh starting photo for every shot, and tested three quality settings on two shots on a Google Colab A100 before shooting all 13 clips. Timings and compute units for each setting, what the full shoot cost, a first run that produced nothing, and how the characters got to say their own lines this time. All characters are AI.

Podcast Notes Weekly: Oct 05–Oct 11, 2026
Investing & Markets5 views

Podcast Notes Weekly: Oct 05–Oct 11, 2026

The takeaways from 22 podcast notes I wrote this week (Asianometry, Macro Musings with David Beckworth, Bloomberg Odd Lots, The Circuit, Invest Like the Best, Tech Wave (科技浪) and more). Two to four sentences per episode, with a link to each full post.

Does the Bigger Bankroll Always Win? A Man Who Bought Nokia Every Month Waited 25 Years to Break Even
Mindset & Growth23 views

Does the Bigger Bankroll Always Win? A Man Who Bought Nokia Every Month Waited 25 Years to Break Even

A popular saying claims that with enough capital and patience you always win in the end. Roulette math and real dot-com stock prices show which half is true and which half hurts people. Educational and methodological, not stock advice.

Did Microsoft Make Its Spending Look Smaller? Burry Said So. The CFO Said It First, on July 29
Investing & Markets31 views

Did Microsoft Make Its Spending Look Smaller? Burry Said So. The CFO Said It First, on July 29

On September 24, 2026, Michael Burry wrote on Substack that Microsoft stretched the useful life of its data centers from 15 to 25 years so its capital spending would look smaller, and the story later spread as a Chinese-language news flash. This post goes back to the July 29, 2026 earnings call and Microsoft's fiscal 2026 annual report, checks six claims one by one, and then looks at the lease numbers that actually deserve attention. An educational note on reading accounting disclosures, not investment advice and not a recommendation to buy or sell any stock.

When the Answer Comes Too Fast, the Learning Doesn't Happen: Notes on “How to Keep Learning in the Age of LLMs”
Research Notes15 views

When the Answer Comes Too Fast, the Learning Doesn't Happen: Notes on “How to Keep Learning in the Age of LLMs”

In an October 4, 2026 post, engineer Oğuzhan Olguncu admits that instant AI answers drained his motivation to learn. He lays out four fixes: use AI as a tutor rather than an answer machine, struggle before you look, plan your stages ahead of time, and make progress visible. This post walks through his approach and adds my own take: the question worth asking is which part of the work we hand to AI. A personal learning write-up.

Days of EQ Tweaking, Then One Room Measurement Fixed the Bass: SweetVox V2.0 and an Ordinary Mic
Research Notes17 views

Days of EQ Tweaking, Then One Room Measurement Fixed the Bass: SweetVox V2.0 and an Ordinary Mic

On October 8, 2026, I used an AI assistant to rebuild SweetVox, my homemade audio mixer, into V2.0. I'd spent days nudging the equalizer within ±2 dB. What finally took the sound from good to great was measuring the room with an uncalibrated microphone: 63 Hz went from +9.0 dB to +0.1 dB. Here's the case for and against, eight steps you can follow, and the open-source code. A technical learning share.

Podcast Notes Weekly: Sep 28–Oct 04, 2026
Investing & Markets7 views

Podcast Notes Weekly: Sep 28–Oct 04, 2026

The takeaways from 26 podcast notes I wrote this week (Catalyst with Shayle Kann, Macro Musings with David Beckworth, Bloomberg Odd Lots, Serenity Weekly, Asianometry, The Circuit and more). Two to four sentences per episode, with a link to each full post.

We Lost the English Textbook CD, So I Built a Reading Machine from a Webcam and an AI
Research Notes26 views

We Lost the English Textbook CD, So I Built a Reading Machine from a Webcam and an AI

My kid needed the textbook audio to practice English, and the CD was nowhere to be found. I clipped a webcam onto an old laptop and wrote two tiny scripts with an AI: one takes a photo of the page, the other reads it aloud slowly in a natural offline voice. Here is how it works, how it compares with other options, what it can't do, and the full code at the end so you can use it. A personal write-up.

I Let AI Turn My Old Laptop Into a Home AI Server
Research Notes24 views

I Let AI Turn My Old Laptop Into a Home AI Server

An old HP laptop still running Windows 7 had been sitting in a drawer. I plugged in an install USB stick, typed a few short commands and entered my own passwords; AI did the rest over SSH from my other computer, installing Arch Linux and Omarchy, full-disk encryption, Tailscale and input methods, while keeping Windows 7 intact. What it did, where it got stuck, what it couldn't do, and the steps if you want to try it. Personal notes.

See more →

Is My AI Assistant's Data Safe on DeepSeek? I Read the Privacy Fine Print at OpenCode Go and OpenRouter
Research Notes4,900 views

Is My AI Assistant's Data Safe on DeepSeek? I Read the Privacy Fine Print at OpenCode Go and OpenRouter

In October 2026 I wanted to move my home AI assistant off DeepSeek's official API, mostly for data sensitivity. This is what I found, in order: OpenCode Go costs $10 a month and lists DeepSeek as '0-day retention', but that agreement only runs to October 31; on OpenRouter, 31 providers serve the same DeepSeek V4.1 Flash, 25 of them are on its zero-data-retention list, and DeepSeek's own endpoint is not. Then the math on when a subscription beats pay-as-you-go. Personal notes, not a recommendation for any service.

Making AI Short Clips with MiniMax H3 on Colab: Is 2K Worth It?
Research Notes191 views

Making AI Short Clips with MiniMax H3 on Colab: Is 2K Worth It?

On October 12, 2026 I ran the MiniMax H3 video model through ComfyUI on a Google Colab Pro A100, generated 3-second clips at 1344×768, and upscaled them to 2K with Real-ESRGAN. Each clip took 206 seconds end to end with a GPU memory peak around 52GB. After comparing both versions, 768 was enough; the story is what decides whether a clip works. Includes the Colab gotchas I hit, the license terms worth reading, and three routes to 2K side by side. A personal learning note, not business advice.