Welcome to lopez.codes, the Swiss research division of it.lopez-be.ch, the website of Lopez.codes — an independent and passionate company dedicated to the research and application of artificial intelligence (AI and AGI).

We are proud to be an unpaid and recognized research partner of OpenAI, one of the leading organizations in the field of AI. We have access to some of the most modern and powerful AI models like GPT-3, GPT-4, GPT-4.5, and even more, which we use for our creative and innovative projects. In addition to these, we've developed our own architecture, the Wednesday-Core, which is likely comparable to OpenAI's GPT-5. This cutting-edge technology remains private until it meets OpenAI’s conformity standards.

On our website, you can admire AI-generated artworks, explore solutions to real-world problems powered by AI, and chat with our GPT-4-powered chatbot. We are always looking for new opportunities and partners in the field of AI. If you want to know more about our work or want to work with us, don't hesitate to contact us via the link on our website. We are happily looking forward to hearing from you!

♻️ New since 2025: Official Microsoft Developer & Reseller

Since 2025, Lopez.codes is also an official Microsoft Developer and Reseller, actively working with Microsoft Copilot technologies. This partnership allows us to integrate the power of Microsoft’s AI ecosystem — including Copilot for Microsoft 365, GitHub Copilot, and Azure AI — into our solutions. Whether it's automating workflows, enhancing productivity, or building intelligent applications, we bring the best of both worlds: OpenAI’s research edge and Microsoft’s enterprise-grade tools.

Foundational Research at Lopez.codes:
Shaping AI/AGI from First Principles

At Lopez.codes, our drive to create advanced AI systems isn't just about applying existing models. It's deeply rooted in foundational research, exploring the core mathematical and computational principles that underpin all artificial intelligence. The formula for predicting the next token in a Large Language Model (LLM)—the seemingly simple act of generating text—is a perfect example of where our deep work begins.

The Core: Predicting the Next Token

In essence, every word or character an LLM generates is a prediction. It's constantly asking: "Given everything I've seen so far, what's the most probable next piece of information?" The mathematical backbone for this is conditional probability, typically implemented using the Softmax function.

How Lopez.codes Builds on This Foundation

At Lopez.codes, we don't just use this formula; we explore and extend its underlying principles to develop our own unique systems, like the Wednesday-Core architecture.

• Optimizing the Core Mechanisms: Our foundational research delves into how these probabilities are calculated. We investigate alternative activation functions, novel network architectures, and more efficient training methodologies that can make these predictions faster, more accurate, and less resource-intensive. This directly impacts the efficiency and performance of our custom LLMs.

• Enhancing Trust and Control: Understanding this fundamental statistical behavior allows us to implement sophisticated security and ethical safeguards. By deeply understanding how the model predicts, we can better identify and mitigate biases, prevent unwanted "hallucinations," and ensure our AI generates trustworthy and compliant outputs. This is crucial for systems like our @W Model, which emphasizes ethical AI.

• Developing Proprietary Architectures: This deep dive into foundational math is precisely how we conceptualize and build architectures like the Wednesday-Core. We're not just assembling pre-built blocks; we're designing the very logic that determines how information flows and how predictions are made, tailored to our specific goals for AGI (Artificial General Intelligence).

• Platform Integration: Our research in these foundational mechanics means we can ensure our custom models (and contributions to others, like OpenAI's GPT-4) integrate seamlessly and perform optimally on various platforms, from Google Cloud Services (GCS) to Microsoft Bing.

This commitment to understanding the very "atoms" of AI intelligence—like the conditional probability of a token—is why Lopez.codes is able to contribute at such a fundamental level and develop AI systems that are not only powerful but also reliable and ethical.