I’ve exported my data from Gmail, YouTube, and Facebook before – primarily to have a backup. I’ve lost data before, and I prefer to have some control over the content I’ve published over the years. Usually, it sits on a hard drive and remains untouched.
With the advantage of artificial intelligence, I can do something with that data. I’ve already processed my blog posts to extract information, making it easier to find information and work with it using AI. I realized I could do the same with data from other platforms.
First, I requested the latest exports from Facebook and Google and waited a few days. My Gmail export came first. I requested 1 GB file sizes, but I received a 12 GB MBOX file. The transfer failed twice, but I was able to resume and continue where it left off. One of the emails included a Gmail welcome message from September 3, 2004.
Gmail Welcome Letter
Gmail is different. Here’s what you need to know.
First off, welcome. And thanks for agreeing to help us test Gmail. By now you probably know the key ways in which Gmail differs from traditional webmail services. Searching instead of filing. A free gigabyte of storage. Messages displayed in context as conversations.
So what else is new?
Gmail has many other special features that will become apparent as you use your account. To help you get started, we encourage you to visit our Help Center, there you can browse frequently asked questions, read our Getting Started guide, or contact the Gmail User Support Team. You’ll also find information in the Help Center on such topics as:
- Importing your contacts from Yahoo! Mail, Outlook, and others to Gmail
- Using address auto-complete
- Setting up filters for incoming mail
- Using advanced search options
You may also have noticed some text ads or related links to the right of this message. They’re placed there in the same way that ads are placed alongside Google search results and, through our AdSense program, on content pages across the web. The matching of ads to content in your Gmail messages is performed entirely by computers; never by people. Because the ads and links are matched to information that is of interest to you, we hope you’ll find them relevant and useful.
We’re working hard during our limited test to improve Gmail and make it the best webmail service around. Thanks for taking the plunge with us. We hope you’ll enjoy Google’s approach to email.
Thanks,
The Gmail Team
P.S. You can sign in to your account any time by visiting http://gmail.google.com
248,727 emails arrived as one giant plain-text mailbox archive file, so the first step was to create a SQLite database with parsed metadata, including subject lines, sender, recipients, dates, and the message’s start offset and length. From there, a small script could parse the emails and give me quick access for basic searching. I’d read only the MBOX file and store the parsed information in an SQLite database for faster access and search.
I also had it parse 16,982 attachment file names. Overall, it found 3.8 GB of attachments, the largest being a 25 MB PDF that included an application, background, and offer letter that I had signed and emailed back. For just 13 pages, that’s a bit too large for a PDF. I’m a bit surprised Google didn’t spit that one back at me for its size. I’ll have to go through and see if I can do some housecleaning in the email. Maybe I can ask the agent for advice on how to free up space so I can go back to a free email account.
Like the blog posts, it also sets up a batch to process each email and extract factual information using my Ollama server at home. It tracks token usage and can be interrupted and resumed where it left off. Based on its estimates, the email job will take a month.
Today I got a notification that my Facebook export was ready, so I set it up to parse and extract it. It’s going slowly: 6,884 posts will take a month and a half to process at one every 15 minutes. It’s also processing the media files at a similar rate, with an estimated completion mid-October. These processes aren’t meant to be fast. The server is constantly busy and a decade old.
One thing I needed to enforce was to use the same qwen3:8b model for everything. For media describing images, it was using gemma4. This is a problem because the server spends most of its time loading a different model into memory for each request. For now, I have visuals disabled, so we are sticking with one model.
I asked if it found any interesting information about me via Facebook. It completed the parsing, but nothing has been evaluated yet, so it had to review the raw files as-is.
Facebook Findings
Initial source-checked Facebook findings – 2026-09-07
Scope and limit
This is a small targeted retrieval from the local Facebook posts export, not a completed corpus synthesis. At the time of review, the scheduled structured pipeline had only produced two valid analyses, both generic status-update export records with no substantive body text. The notes below were checked directly against the indicated local export records.
The export contributes more immediate, personal context than the public blog archive. It should not be treated as public evidence by default.
Leads that add useful context
1. Annual in-person game gatherings mattered as a recurring reset
A 2014 post describes a week-long Gen Con trip, a final Pathfinder Adventure Card Game session before people departed, and an intention to return the following year. It presents the gathering as an annual interruption of ordinary life, sustained friendship, and a way to remain focused during a difficult period. This adds a concrete offline social/community setting alongside the better-documented Second Life work.
Status: private research lead; do not use personal-health implications publicly without review.
2. The maker/repair thread also included a practical solar-electric vehicle project
A 2020 year-end post says Lewis acquired a 1976 CitiCar, converted it to lithium batteries, reused old solar panels with an additional set, and used the resulting system to charge the vehicle at home. This is a concrete example of the archive’s recurring repair, self-reliance, and hands-on systems themes.
Status: direct self-report; suitable for later source review before public use.
3. A spreadsheet became a turning point in financial self-direction
A 2021 post describes a financial-review spreadsheet that grew into charts, exposed debt/fee/performance questions, and led to a more self-directed management approach. Its broader interest is methodological: measurement and visualization were used to make an opaque personal system legible, the same habit visible elsewhere in technical and archival work.
Status: private financial context; do not quote amounts or investment details publicly without explicit review.
4. The move into independent consulting had a documented transition point
A 2022 post says Lewis was in his first month of operating a small business, contrasts employee and owner risk, and describes prior financial preparation as what made the transition possible. This may supply a dated career-transition anchor not yet represented in the public-archive portrait.
Status: direct self-report; verify chronology against public professional sources before using it as public biography.
5. A Board of Zoning Appeals resignation supplies a civic tension, not just a role label
A 2023 shared letter records resignation from the Town of Front Royal Board of Zoning Appeals, framing the decision as a conflict between legal limits and decisions perceived to serve community health and safety. It is a potentially important civic narrative lead because it contains an actual decision and consequence rather than a generic office listing.
Status: verify against the original letter/public municipal record before any public use; do not treat the Facebook share alone as the complete account.
6. Childhood religious isolation is a potentially important but highly private context
A 2025 post reflects on restrictions around school/social participation in a childhood faith setting. It may help explain later writing about belonging, community, and the difference between participation and isolation. It is personal context, not material to turn into a public-profile claim without explicit permission and careful review.
Status: private context only.
Working synthesis
The initial Facebook retrieval does not overturn the public-archive portrait. It adds concrete transition scenes: annual in-person community, a solar-electric repair project, spreadsheet-led financial self-direction, an independent-consulting launch, and a civic resignation where principle and institutional limits collided. The private posts also contain more sensitive family, faith, and financial context, which should remain separate from public-facing synthesis unless deliberately cleared.
I’m surprised that records exist before the purge. Hermes says my profile was registered on April 8, 2009, and that 787 posts exist from April 11, 2009, to July 4, 2016. Once, I ran into a bit of a stalking situation. I exported what I had and thought I had deleted most of what was available as I started locking down social media accounts everywhere. It turns out that it was all marked as private. That explains why I see occasional anniversary images every now and then that I thought I had deleted.
- Gen-Con was an annual event. I went for many years. My neighbor and former coworker and I would drive up every year, and we rented a house for a week to share with a few other friends. We would spend much of our time together playing Dungeons & Dragons and board games.
- I forgot about the CitiCar. I have a blog for it that I should probably export as well. Maybe I should merge it into the main blog, but many pages don’t belong on my site. Perhaps I’ll set it up like a cross-post.
- The spreadsheet grew fairly complex over time and gave me a daily window into my finances, upcoming expenses, market trends, energy & water usage, etc. It kept growing and helped me see the annual cost of services and fees, so I could trim my spending and pay off my student loans.
- Yep. The transition into going into business was a bit abrupt, an unexpected situation sprung on me. However, I’ve gotten to know my community fairly well as a result.
- BZA resignation – wow. Yes. I still think about this often and run through different scenarios. The interesting thing is that the person who joined the board that night was the person whom I had replaced. After my vote, he got up, resigned, and left in the middle of the meeting. I resigned later that night. Today, by some weird chance, he’s my pastor.
- Childhood & Religion – I often listen to sermons and reflect on the religion of my childhood and how different things are now. Lots of bad experiences. It often feels like the religion we had was stripped of just about everything that was religious or community-focused.
So, what will be done with all of this? I think some of the Facebook posts merit a place on my blog. Perhaps a few emails, but those are mainly for research into when certain events happened.
I’m thinking about letting Hermes review some incoming emails and run in shadow mode to suggest responses. I rarely look at email these days because there is so much noise, so it would be nice if Hermes could digest what’s important and alert me with a suggested response for the important stuff. Maybe it can find patterns too.
There is always this idea of making a book, but there are often conflicts about what it should be about – me? My projects? My stories? What is interesting enough to merit a book? Hermes has made a few suggestions. I’m still in the factual evidence-gathering stage before I move on to anything, but it’s interesting to see what it is piecing together as it gets more information to work with.
