skip to content

Contact Me!

For any inquiries regarding [contracting, hiring, sponsorships], or if you’re a like-minded individual, reach out on:

Scraping Bits

Scraping Bits is my technical podcast for people interested in the weeds of niche development topics. I’ve primarily focused on crypto [mev, dapps, wallets, zero-knowledge, bridges, investigators, cyber tooling, infrastructure], however there are plans to expand to AI, once I get more technically sound to unlock those interesting deep-dive conversations that are missing.

Feel free to listen to Scraping Bits on Spotify!

This was the first interview I’ve ever done too.

To keep up to date with new content subscribe to my Mailing List!


I’m an independent author, educator, podcast host @ScrapingBits and self-taught (no college) research developer who focuses on understanding and solving complex problems at a low level.

My primary focus is computational neuroscience, precisely self-mutating AGI (involving self assessment, correction and mutation).

Before AI I was building a bytecode fuzzer + static analyser from scratch for smart contract exploit generation and reached the mutator part. After realisation that the fuzzer would not learn after thousands of iterations, and I would have to bake in the intuition, I thought “this seems wildly inefficient”. And so AI was the next progression. However there are some detrimental flaws in AI which lead me to down the road of complete autonomy with self mutating continous learning AGI. The one that can achieve such an AI will unlock the ability to exponentially grow or destory human civilisation.


The following is what I’ll be focusing on for 2024:

Granular Human Neuroscience

To achieve the level of generalised learning we must learn from the leader in that field, humans. Unfortunately we cannot experiment on ourselves because of ethical considerations but I think to truly understand by verifying it is a necessity. I think AGI can be achieved with the first three topics I mention below. The forth, emotion, is the endgame of AI to bring it to “life”.

  • Modeling: Complex concept representation and where we are in space and time.
  • Memories: Abstract relationships between nonrelated topics, data representation, how they influence decisions.
  • Decision Making: Reward mechanism, abstract thought, action filtering, explore when exploit outcome(s) is certain.
  • Emotion: I’m quite interested in artifical emotion. Inherently AI doesn’t have emotion or a need for socialising however motives, goals and self sacrifice come from emotion (think of a parent). It’s a fascinating thought experiment after intelligence is achieved.


These topics in math seem like the most correlated with my goals and it just so happens it’s super helpful with financial engineering so might fuck around with that too. I’ve realised to do any inventing you really do need math. It is the greatest edge you will have in conjunction with engineering. Unfortunately I’ve never needed it so it was never a priority. But now, here we are. I believe that if you cannot recall things without reference then you don’t really know it. The cue to the retireval of that knowledge would be a problem. If you don’t recall via cue then you don’t actually know the thing.

  • Algebra: Master all the fundamentals
  • Category Theory: Findign the relationship(s) between two things that are different — modeling everything! The joy of abstract mathematical thinking - with Eugenia Cheng
  • Probability Theory: Decision making in complete uncertainty + after we gain 1 experience, 2, 3 and so on — how our deisions change over time.
  • Stochastic Differential Equations: Model systems influenced by randomness and uncertainty even with noise.
  • Multivariate Calculus: For the activation of our neurons.
  • Stochastic Calculus: For rates of change in continuously changing environment(s).
  • Dynamical Nonlinear Systems: An abundance of linear functioning neurons (each dendrite + synapse) create nonlinear systems (biological neural networks). We need this to continuously update the algos over time.
  • Graph Theory: To construct biological neural networks, assignment of relations and abstract thought — also seem quite useful for cyber sec.
  • Linear Algebra: Financial engineering foundation.