What if AI thought in many ways at once?

A thinking chamber, not a chatbot wrapper

Swiftest.ai fans questions through diverse cognitive lenses — code, math, symbols, diagrams, simulation, critique, product thinking, human context, memory, and research — then a final synthesizer model weighs and distills everything into a single answer.

Not every mind helps every time. That is part of the research. The goal is to measure when cognitive fan-out actually improves reasoning, creativity, accuracy, and problem-solving — and when it does not.

Swiftest.ai is being built as both a product and an open experiment: a way to explore what comes after single-model prompting.

About me

I'm Linh Ngo — a solo developer, AI builder, filmmaker, dancer, and longtime internet creator.

I have been fascinated by artificial intelligence since college, where I wrote a predator/prey artificial-life simulator and became obsessed with early AL systems like Tierra, Avida, and other digital evolution experiments in the spirit of Core War. Long before today's LLM boom, I was already captivated by the idea that intelligence could emerge from interacting agents, competition, feedback, memory, and environment.

That thread runs directly into Swiftest.ai. The "many minds" architecture is partly inspired by those older artificial-life ideas: intelligence as an ecosystem, not just a single monolithic brain.

I'm also the founder of Dance.net, once one of the web's major online dance communities, which grew out of my life as an avid hip hop dancer, instructor, and creative technologist. I have taught dance, performed, made films, written music, and built online communities around creativity and expression.

That mix — software, AI, music, movement, storytelling, and internet culture — is what drives my work now.

My current projects

Swiftest.ai is one part of a larger ecosystem.

The questions I'm trying to answer

I believe the next generation of AI products will not be defined only by model size. They will be defined by architecture. There is still enormous room for independent builders to experiment with how models are scaffolded, evaluated, cooperate, and fail.

• Can a coding model help solve a non-coding problem by reframing it as an algorithm?
• Can a symbolic model expose hidden constraints?
• Can a visual or simulation mind catch spatial reasoning errors?
• Can multiple cheap models, properly orchestrated, sometimes produce insights a single expensive model misses?
• Can we measure when the extra "cognitive noise" becomes useful signal?

Transparency is part of the product

Swiftest.ai, Uncen.ai, Multiplayers.ai, and FullAuto.ai are all being built by me as a solo developer. The systems are rough in places, fast-moving, experimental, and very personal. But they come from a unified vision: AI as a creative and cognitive amplifier for people who want to build things, tell stories, solve problems, and think better.

I'm not trying to make AI feel magical by hiding how it works. I want to expose the process — the minds, the traces, the disagreements, the failures, the recoveries, and the final synthesis. Swiftest.ai is especially interested in building systems that can show their work in a useful way: not by dumping hidden chain-of-thought, but by presenting structured cognitive artifacts that humans can inspect.

Where this could go

If the "many minds" approach works even some of the time, it could become useful for complex reasoning, software design, product strategy, creative development, legal and procedural thinking, research synthesis, visual and spatial problems, education, decision support, and AI evaluation itself.

Work with me

I'm building at the intersection of AI, software architecture, creative tools, artificial life, video generation, agentic coding, and human-computer interaction.

I would love for this work to reach people building the future of AI — including teams at Meta, OpenAI, Anthropic, Google DeepMind, xAI, and other frontier labs — not because Swiftest.ai has all the answers, but because it is exploring an important question:

If that question matters to you, I'd love to connect.