Back in 2009, I worked at a place called Justin.TV. It would give birth to two significant companies… socialcam (arguably a precursor to tiktok and vine) as well as Twitch.

When I was there, I was a video producer. I didn’t work on product or code, but instead helped ensure high quality broadcasts worked without effort.

One day, I jumped out of a shower completely naked and announced to my wife: “I’ve figured it all out!”

“Figured what out?”

“I’ve figured out how video should work. If it was invented today, it would be assembled just-in-time on the client.”

Even with this soap-inspired revelation, I didn’t have a clue how to pull off that idea. I tried to pitch the idea to my coworkers at Justin.TV… but at the time, Justin.TV was barely surviving on its own. We had a doomsday clock during our weekly company meeting, and adding client assembled video would not prolong our life.

For years, I tried to crack the nut. I worked on server-assembled video, podcasts, and other various projects to try to make it work. I did my own shows for years trying to figure out how I could make an amazing quality video in as few minutes as possible.

Eventually, I realized nobody was going to work on my idea unless I built it myself. I moved away from SF, started learning to code, and eventually birthed what would become Vidpresso.

History doesn’t repeat itself, but it rhymes.

This weekend I spent some time with GPT4. I was astonished by the speed and quality of the results. (which I’m consuming through poe.com)

I was able to explore and learn concepts I’d casually studied in little time, and even better, I was able to build code in unfamiliar languages (AppleScript!) without needing to understand any of its quirks.

GPT4 helped me build a text based todo and notes app in about 3 hours. It was an app I’d been thinking about for years, but since it was a personal project and central to my workflow, it would require a lot of attention to build it correctly. I hadn’t been able to make the time budget work.

But reduce the cost from days / weeks to hours? Who doesn’t love a good sale?

AI rhymes with 1992

Imagine what it was like to be in computing in 1992. Pre iphone, pre internet, pre windows 3.1, only a few years post Mac, and not that much time since the first spreadsheet.

Today feels like 1992. GPT4 is the spreadsheet + Mac + iPhone + internet, and will enable makers to build at a speed that seemed impossible 6 months ago. Now coders can create apps as fast as video creators can create TikToks.

Today’s shower thought: I can build anything without needing anyone else.

If I architect things correctly, I think I might be able to build a company as impactful as Facebook, but without the 100,000 people.

Since Trump was elected, deepfakes were invented, and the environment of general mistrust of media which continues to spread, I’ve been trying to solve a problem. How can we trust media without any intermediary? How can we place a level of trust in something from someone we’ve never met / seen before?

I’m still working through it, but I think with a combination of peer-to-peer technologies, merkle trees, and public cryptography, it’s possible to ascertain the origin of a piece of content, and figure out how trustworthy it is.

Trustless Trust

Imagine every social media account on the internet as a public / private keypair. Now, imagine each account can rate any other account with a trustworthyness score ranging from -5 to +5.

-5 is you’re impersonating me. (Fraud) -4 you’re a criminal gang / scary big entity / monster out to get me. (Fear) -3 you are a troll or an individual I’m scared of, and I think the people you trust might be sketchy. (Foe) -2 You’re someone who says stupid things regularly, and while not actively destructive to the world yourself, you cause bad effects. (Fail, faltering, faulty?) -1 You’re honestly just misinformed, and you make decisions that make me roll my eyes. (Fumbling) 0 I have no information about you whatsoever. +1 You are a real person who is making decisions I value, and I think you’re trying to make the world better (relative to my perception of better.) (follow) +2 You say things that I respect, and you tend to influence others to make the world around you better. (Fan) +3 You’re someone I closely trust, resembling family, a close circle of friends, or someone I would feel comfortable saying anything in a group chat to. (Friend) +4 You are a service or device which I trust to act on my behalf, and I’m pretty sure you’re going to do the right thing most of the time because you’re incentivized to do so. (Fiduciary ) +5 is you are literally me. You are my fingerprint, private key, retinal scan, passcode, secretest of information which allows people to authenticate me. (Fingerprint)

With the advent of GPT4, I can finally start working on this project. At a high level, my strategy is to continue to author content like you’re consuming now, as well as videos, podcasts, and even products, which I’ll treat just like content.

At some point, I’ll build a CMS to help me make this content more effectively and distribute it on other platforms. Later, I’ll encourage anyone who values stuff I’m making to come get it from the source: My website.

Someday way in the future, I can use this system to bootstrap the trustless trust system. People who subscribe to my content become the first participants in the system. Eventually, you could imagine a “proof of agreement” style system creating small groups of people exchanging value with each other, without having to go on public cryptocurrency networks.

I imagine you’d, for instance, get certain benefits for subscribing. You could also earn “karma” just like on Reddit, which could lead to free subscriptions. Karma could be used to purchase ad space, and if people contribute content or code to the site, we could pay in Karma which could turn into real dollars via Stellar etc.

Parts of this idea are years old. (Judging points, +/- 5, etc.) The most important parts of this idea are only hours old.

I think I can build it by myself now because of GPT4, and I’m pretty sure it’s possible to build a company as big as Facebook using cryptocurrency principles.

Things feel very different now.