Thursday, April 23, 2026

Highlight reel from today…

  • YAML frontmatter
  • File over app!
  • List of people I would actually want to talk to for my fuzzy “Do You Want To Talk About AI?” project
  • Dreaming about a 10,000 year clock in Texas…

Blog

  • For some reason I decided to make a list of vectors I am observing. I’m sure there are more. The only reason is to find people who want to talk about AI.
    • I kinda want to post this to my blog today? This is my version of “working with the garage door up”
    • lol nerds pls don’t snipe me! Connect me?? I live in a garage, we have enough but we are poor! I automated like 75% of my job, so I have time to dive now. But I know the clock is ticking because AI is killing our click-through rates.
    • This “working with the garage door up” is the exact opposite of my experimental creative writing era, where the garage was very closed and what I wrote was weird, dark, and edgy. That was an internal project I did for myself.
    • This is not that. This is about PEOPLE! =)
    • It could be a fun project?
    • Only problem is I’m a ghost! I’m not on LinkedIn, Facebook, any platform because I haven’t needed to be. I’ve been mining SEO gold out of this niche for so long!
      • The world is changing and I am being forced to adapt. Fast. I have a family and the job market sucks. Things are expensive. I feel like people might relate?
    • I dunno, Humans Of New York (HONY) worked! This would be my humanist mama version of HST/Kerouac embedded in their own revolutions in the 60s/70s. “Stay hungry. Stay foolish.”
    • I’m also really interested in oral history. I want to help people tell their story to the world. Then it is not locked in a silo, but open-source!

Philosophy

  • Humanism is a philosophical stance that emphasizes the individual and social potential, and agency of human beings.

    • This ^ is me rn!
  • File over app

    • file > app
    • This is a digital ’new philosophy’. If you want to create digital artifacts that last, they must be files you control, in formats that are easy to retrieve and read.
    • Ethos: File over app is an appeal to tool-makers: accept that all software is ephemeral, and give people ownership over their data.

Code

  • I am re-evaluating the YAML frontmatter on the shorthand schema I wrote up yesterday, which was such an exciting ‘green check-mark’ moment as these learning logs grew.

    • Now I am thinking of this as…
      • A human/machine interface! Lite Markdown!
      • This is my current way of thinking about the world:
        1. My life is a firehose of information.
        2. AI agents can use that datastream, IF…
        3. …We have a system that works for both of us!
        • Thinking backward from this endpoint: I realize that in order for us to communicate and work well together, I need to refactor not WHAT I am saying but HOW I say it.
      • It’s like the Diet Pepsi of AI… where the diet version of Markdown may actually taste better than the original! At least to a human who is trying to touch-type all this out!!
    • It still needs work, like I’m not super excited about long XML tags like <verbatim>. I wonder if I could get away with something like %% instead? And fix the YML too.
    • Then of course when I actually try to use it with an agent, tons of flaws I never considered are sure to surface!
      • In fact maybe this whole idea will explode! I don’t know. It’s experimental. hashtag #mad-scientist-vibes #lol
  • Obsidian YFM (YAML Front Matter)

    • Metadata block at the top of markdown files used for organizing, filtering, and querying notes. Allows users to define variables. Essential for organizing knowledge, linking ideas, and managing projects within Obsidian.

Questions

  • What is YFM? (YAML Front Matter)

    • Optional section of valid YAML that goes at the top of a page to maintain metadata for the page and its contents.
  • What is Obsidian?

    • https://obsidian.md
    • Obsidian is a note-taking app for organizing personal notes, journaling, knowledge bases, and project management.
    • Users write in Markdown
    • Primarily for note-taking
    • Also used as a ‘personal wikipedia’, ‘digital garden’, or ‘interactive notebook’. Users organically grow a personal knowledge base in an organized way.
    • Obsidian advertises privacy, security, and personal ownership of data. Data is stored locally. Users do not need to share personal information if they do not want to.
    • Obsidian is run by a small team, low overhead. Engineering team only has 3 people. Total staff: 7 people.
    • Offers paid tiers, such as Publish: $8/site/month/billed annually, and Sync: $4/user/month/billed annually.
    • This is a niche market that seems to work for now!
    • In April 2026, this article was published: “Three Engineers, Zero Financing, No Meetings: Obsidian, a $350M-Valued ‘Small but Beautiful’ Company”
    • The founders’ core mantra: “Files over apps.”
    • I love Kepanos’s X post on 2025-8-9:
Obsidian is weird:
- 7 full-time employees
- ~1 million users per employee
- fully remote
- 1 in-person meetup per year
- no scheduled meetings
- no stand-ups
- deep focus is prioritized
- our manifesto guides our product

Ideas

  • Do You Want To Talk About AI?
    • This is floating around my head as a project. I don’t have much direction for it beyond… my overwhelming urge to talk to literally everyone about AI. Not because I want to talk, but because I want to listen.
    • This is like the ‘blind men and the elephant’ fable, where it seems like everyone has their hand on a different part. AI is a big animal. The deeper I fall into this dive, the bigger it feels. I’m using a microscope to look at bacteria, feeling overwhelmingly small in the universe.
      • So this morning I had some direction. Over the past 9 weeks, I have felt out some angles or vectors that seem like they might hold people who would be interesting to talk to.
  1. Genius mad scientists. This is the “cyber psychosis” thing in action. It is a real vibe in tech now.

    • [Others] in AI have also described “cyber psychosis.” Most will not talk to me but maybe some will?
      • Odds are 1/1,000,000 but someone might link me up?
  2. YouTuber who made $700,000 using AI to make extremely boring videos. He said he is a doomer whose niche is closing quick. I’m in the same boat, so we might relate?

    • I think most people in the world do not realize how much money there is to be made in small niches.
    • Like my job too, I know millions of dollars flowed through me! Yet when people ask me about my job and I try to explain it, 9/10 just go glassy-eyed. I explain it in simple terms, but it is just unusual for most people to think about money-making systems that do not fit their mental model.
    • Conclusion: AI is civilizational technology. Things are going to change. Change is hard! How will we adapt?
  3. Philosophers

    • I love what this person said about niches on his blog post: “A blog post is a very long and complex search query to find fascinating people and make them route interesting stuff to your inbox” (https://www.henrikkarlsson.xyz/p/search-query)
      • He had a great point about small niches becoming extremely useful and scalable when your spread is, like, billions of people on the internet. Specificity is key.
      • This idea is extremely intuitive to me, with an SEO background. In SEO we have a saying: “The riches are in the niches.” No idea who said that first, but the idea is that you don’t want to find ALL of the people! You want to find the RIGHT people. In my experience as a copywriter, finding ALL of the people is actually a bad thing because then you are drowning in calls from people, 99% of whom do not have a case that will actually turn into a lead.
      • There are philosophers who have posted on the small web and polymathic learning. There seem to be some patterns in thinking about non-linear thinking!
      • Which brings me to this gem:
        • “You Can’t Think about Thinking without Thinking about Thinking about Something.” Papert, Seymour. Contemporary Issues in Technology and Teacher Education, 2005.
        • [note: cognitive science in a nutshell?]
        • The best title of a research paper or the greatest title? #lol
        • This paper claims that the new idea of children using computers as tools for creativity was described in “Teaching Children Thinking” (1970).
        • https://eric.ed.gov/?id=EJ1080976
      • Someone also said:

[2024-12-17] “…just come walk around Emeryville and West Berkeley. It would look like that! All the tumult of Twitter would shrink into a single weird cafe…”

  • Is this cafe an actual physical place?! [?]
  • Not sure if any of these people would talk to me, probably not, and I might not know what to say, but maybe.
    • Also there is a quote on Matuschak’s notes:

“If you ever needed another reason to learn in public … This will get you invites to very cool exclusive events filled with high-achieving, interesting people.”

  • See that!!! LEARN IN PUBLIC. I feel seen. *_*
  • Not sure if I am actually ready to ‘work with the garage door up’ but it is calling to me.
  • I am normally a bit socially awkward and uncomfortable at parties that do not involve at least one dog or screaming child (usually mine), but I also have a very deep personal practice of travel and tourism. I am a tourist everywhere I go! Even after all this time in California I still feel like a tourist! From like an anthropological perspective, I am deeply fascinated by unseen systems.
  1. smolweb

    • This is the smolweb angle. This is a niche on the web that I have been fascinated to browse. It is so decentralized and the voices are so human and intimate. But discoverability is an issue (actually a feature, not a bug) and I am hoping that I can design an agent to browse it or Kagi Smallweb to find people and content to send to my inbox? This might be useful for my Hop-Protocol.md project.
    • I have no idea who I would talk to about this, but I am praying that the deep-research agent that I am designing (her name is Seek!) will help me find the right people.
    • Kagi Smallweb: https://kagi.com/smallweb
    • Gemini protocol: https://geminiprotocol.net/
  2. Boots on the ground

    • People like Blizzard dad, LinkedIn mom, etc. I want real-world stories of how this is affecting people. Tom completely flipped his business model in 2 months, going from a project-based marketing company to a company that supplies clients with curated AI material for a monthly fee.
  3. Deep learning models / fractals / physics / math / art

    • I am absolutely blown away how the deep learning/ML engines running this whole AI machine under the hood are connected to fundamental concepts in math and physics. Not sure if I’m too far over my head tho.
    • This is mesmerizingly beautiful!!! -> “Neural network training makes beautiful fractals” (Sohl-Dickstein, 2024)
  4. Fintech / RENTEC

    • If you want to know the future of AI, follow the money!
    • It also seems that the history of AI is rooted in some big money being poured into early ideas by people who had good instincts.
      • I find the structure of RENTEC absolutely fascinating. It’s a very secretive financial company that infamously does not hire finance people, but ‘creative problem-solvers.’
      • This also ties into my idea about finding “riches in the niches.” RENTEC was all about one very specific niche. The result? The most profitable hedge fund in history!
      • Was this the movie Moneyball? [?]
      • I think their financial models may have also played a role in the leadup to the AI revolution? [?]
    • This is interesting to me from a corporate structural standpoint, also the question: Why do some teams succeed?
      • Same angle as my Uncle who invented a thing that generated billions and is still generating billions. He wrote a book about corporate structure and how to hire people with “T”-shaped deep learning domains (basically a spread of skills at the top, which allows them to connect, and one very deep dive, which gives them great power).
    • I know I’m dreaming here but it would be cool to go on a trek through Texas to visit that 10,000 Year Clock, because ya know I love trekking after going to Nepal!! lol lol #lol
  5. Climate

    • AI data centers eat up a ton of resources and this is understandably hitting a nerve in San Francisco. I don’t know what side of the fence I’m on, but I like the way climate activists think in terms of 10,000 Year Clocks.
  6. Tower of Babel / Gokarna, India / LumiChats

    • This was a tiny thing I read about LumiChats, which found a niche by flipping the typical AI business model on its head… Instead of a $20 monthly fee, which is expensive and not useful for students in India crunching for exams during a specific time-period, offer the service as a DAY PASS.
    • They discovered an absolutely massive untapped market! INDIA! Now extrapolate this idea worldwide and… you get… $$$
      • I deeply believe that there is untapped human potential all around the world. What would happen if you gave every kid in the world access to the smartest brain and a room full of interns to help him/her build stuff? I don’t care about people making dumb puppy videos… I care about the kid in Dhaka or Bangalore or Burkina Faso who is a savant, a neurodivergent who just needs access to the right tools.
    • Another good idea from LumiChats is to combine multiple models on one platform! So you ask ChatGPT to brainstorm, Claude to write it, and Gemini for specific statistics. THIS IS A BIG IDEA… but hasn’t hit here in the U.S. yet.
    • This is about unlocking human potential on a global scale.
    • Gokarna, India…
      • I love India and I would love to talk to students about how they are using AI… maybe after their exams! ha!
      • I also would love to visit Om Beach in Gokarna after I catch my breath and finally slow down on this AI deep dive. I miss India so much… maybe if I get a mental refresh on India again, I will finish my travel book once and for all!
        • [note: my 1/2-finished unpublished gonzo travel memoir was my humanist rebuttal to Eat, Pray, Love (Gilbert)]
  7. And for the HOLY GRAIL of this extremely, um… unlikely Dreaming In AI grail quest…

    • I have an idea but…
    • This person seems so chill and sincere in her podcasts and lectures. This is a prime example of the quietest person in the room being the most important. I would love to talk.

[AI — Claude Sonnet 4.6]: AI is civilizational, the money trail matters, the human stakes matter more, and nobody is talking to the right people yet.