⚒ ACG ⚒
Founding Member

Laurie Scheepers

Scientist · Engineer · Philosopher · Systems Thinker

Laurie brings systems thinking, AGI research, strong philosophical discipline, and a deeply human orientation toward technology, family, and community.

Founding Member Technical Director AGI Research and Development Cape Town, South Africa 🇿🇦

Laurie Scheepers

Scientist · Engineer · Philosopher · Systems Thinker

Laurie Scheepers portrait

I work in AGI research and development -- on the side.

My day job is serving as the technical director at ENTER Konsult, our lean advanced tech firm. We transform industries that modern software has underserved, building long-term relationships to guide organisations through this revolutionary era. We make technology work FOR people -- not against the humans who make any organisation what it actually is. We also develop private, bespoke software for select people and organisations.

AGI R&D Systems Thinking Lean Advanced Tech Quality Control Critical Thinking Technology for People

Our ethos:

  • We will not build what should not exist
  • We transform complexity -- KISS is paramount
  • We enforce strict quality control
  • We do what is RIGHT
  • We extend our thinking -- never outsource it

enterkonsult.com

I am a scientist, engineer, philosopher, and systems thinker. That last one is a polite way of saying I see patterns connecting things that are almost certainly unrelated, and I won't stop talking about it.

I've been fascinated by nature's inner workings since before I had the vocabulary to describe them. I still am. Nature is a better engineer than any of us. She just refuses to document anything.

When I'm not in front of a computer, I'm reading, writing in my journal, producing music, or co-designing games with my eldest son -- the actual game designer in the family. I provide the infrastructure. And opinions nobody asked for.

Ahhh, there's always too many things and not enough hours.

But none of that comes first.

I am a dad and a devoted husband. The everyday moments with my wife and two boys -- big or small -- are what I live for. I capture them in my admittedly absurd memory. Doctors told me as a child that my head is too big. I have chosen not to interpret this. 🤣

How I see things.

I build things. I can also destroy what I build. That is not a threat -- it's a definition of ownership. If something I create doesn't meet the standard, I would rather take it apart myself than watch it become someone else's compromise.

I don't try to prove things. I try to break them. What survives my best attempt to disprove it is the only thing worth keeping. This applies to software, to ideas, to strategies, and -- on good days -- to my own opinions, which I hold strongly, change willingly, and will abandon in front of you without embarrassment if you bring a better argument.

I believe life is not a zero-sum game. This is not optimism -- it is an engineering observation. The most enduring systems I have studied are cooperative ones. Competition and cooperation are not opposites; they are the same force expressed at different scales. Nature arrived at this conclusion long before we started writing papers about it.

I believe critical thinking is the most important and most neglected skill in any room. Not cynicism -- the genuine kind. The kind that starts with questioning your own assumptions before you touch anyone else's.

I believe respect costs nothing and compounds infinitely.

I believe individualism and community are not in tension. You cannot build a strong collective from people who haven't figured out who they are. And you cannot figure out who you are in complete isolation. That tension is not a problem to be solved. It is the engine that drives everything worth building.

I am a teacher. But I vastly prefer being a learner. In my experience, the best teachers always do.

Respect. Individualism. Sense of Community. Critical Thinking.

These are the principles that shape me. Or perhaps they shaped me long ago and I am only now finding the right words for them.

I am a gentle soul. I mention this because people who hold strong convictions are often assumed to be aggressive. I hold very strong convictions. I am not aggressive. These things coexist more easily than people expect.

I'm based in Cape Town, South Africa 🇿🇦 -- the city with the best weather in the world. This is not opinion. This is data. Come visit and argue with me about it over coffee.

You can contact me anytime on the Discord channel.

Kind regards,
V>>

What I Build and What I Learned

Engineering work, tested honestly. What survived falsification is listed. What did not is acknowledged.

GRIP

A recursive self-improvement system for Claude Code, used in production at Sudonum. It applies well-known patterns -- caching hierarchies, evolutionary selection, fixed-point iteration -- to the specific problem of making AI agent collaboration more efficient. The engineering integration is the contribution, not the algorithms.

160 Skills 28 Agents 53 Commands 30 Operating Modes Mechanical Safety Gates

GRIP enforces safety through mechanisms that constrain, not values that suggest. Every destructive action is gated. Every gate is deterministic. The system cannot argue its way past its own rules. This is not a philosophy -- it is an engineering decision.

Convergence Research

A paper exploring structural isomorphism between Maria Stromme's consciousness framework and convergence kernels. 20 of 20 theorems compile in Lean 4 with zero sorry. 17 need no axioms; 1 uses an honest axiom (Mathlib coverage gap, not a mathematical gap); 2 were reformulated to equivalent statements.

I submitted the paper to my own 5-model adversarial falsification council. 4 of 6 claims were killed. 2 were weakened. Zero survived intact. The council found that the structural mapping was circular (bridging axioms were stipulated, not discovered), the QFT formalism was cargo-cult notation, and the consciousness claims were a category error. The 2 weakened claims retained partial validity: the self-reflection map f(x) = 1/(1+x) does have a unique fixed point at the golden ratio complement (algebraic fact), and the contraction criterion is necessary-but-not-sufficient (as claimed, but the "canonical" framing was undermined).

The honest results -- including what was killed -- are more interesting than the original claims. The willingness to publish failures is the point.

Genome System

An evolutionary fitness tracker for AI components -- skills, agents, modes. Currently Generation 33. After an initial rise over ~15 generations, fitness is declining (from ~0.66 peak to ~0.47). I report this honestly because declining metrics are more informative than rising ones. The decay mechanism (use-it-or-lose-it) prevents component bloat, but the overall decline suggests the measurement needs recalibration or the environment is shifting faster than the system adapts.

Related work at far larger scale: Google DeepMind's AlphaEvolve evolves entire algorithms. Sakana AI's Darwin Godel Machine modifies its own codebase. GRIP's genome is a lightweight independent implementation of similar evolutionary principles, applied to agent configurations rather than algorithm discovery.

Epistemic Practice

I believe the most important AI skill is knowing what you do not know. I build falsification protocols to test my own claims. I cite prior art that undermines my novelty claims. I publish declining metrics. This is not modesty -- it is engineering discipline. If you cannot break your own work, someone else will, and they will be less kind about it.

Contributed Ritual 09 (Falsify Before You Ship) and Ritual 10 (Cost-Aware Retrieval) to the Guild's AI Rituals archive.

Explore More Members

Return to the member directory to browse the other founding profiles.