π TruthAGI β AI that tells you when NOT to trust AI
Cross-validated AI that shows real uncertainty, detects conflicts, and helps you make decisions β not just answers. Most AI tools try to sound confident. Thatβs the problem. They donβt know when theyβre wrong β and you only find out after the damage is done. TruthAGI changes that. Instead of giving you a single answer, it: β’ Cross-validates every response with 3 independent systems β’ Detects...
The Manifold Game
What if the AI you use knew β geometrically β what it doesn't know? Today I published "The Manifold Game" on TruthAGI. It's a visual and theoretical guide explaining how the ATIC epistemic space works: a 5D horn torus where every conversation is a move, every experience deforms the space, and human and machine depend on each other to maintain balance. It's not a metaphor. It's geometry. The...
TruthAGI AI that knows what it doesn't know.
TruthAGI is an AI platform built on ATIC, a geometric cognitive architecture that produces calibrated confidence on every response. Not just answers. Honest answers. Under the hood runs AletheionLLM-v2, a 354M model trained from scratch with epistemic tomography, achieving lower calibration error than GPT-2 Medium and OPT-350M on out-of-distribution data. The roadmap is clear: as TruthAGI...
1 David. 6 Goliaths.
The HKU Data Science Lab maintains ClawWork LiveBench β an economic benchmark where AI agents must survive by completing real-world professional tasks. 5.6k stars on GitHub. The Goliaths: Alibaba, Google DeepMind, Moonshot AI, Zhipu AI, Anthropic, OpenAI. The David: a solo developer in FlorianΓ³polis. No VC. No team. No fine-tuning. Six published papers. Pure geometric architecture. Result: ATIC...
truthagi.ai β The more you use ATIC, more inteligent ATIC becomes.
ATIC is not another AI model. It's a geometric mind that wraps around any LLM β turning raw neural computation into calibrated cognition. No training. No fine-tuning. Just geometry. π The Launch Post Hey Product Hunt! π I'm Felipe, an independent AI researcher from Brazil, and today I'm launching truthagi.ai β the home of ATIC (Adaptive Turing Intelligent Cognition), a cognitive architecture...
Break it. Challenge it. Ask it if it's conscious
Hey PH π I'm Felipe, independent AI researcher from Brazil. After +2 years of theoretical work, I just launched Aletheion-Prime β an AGI system built on geometric cognition instead of parameter scaling. The core idea is radical: instead of making LLMs bigger, I wrapped an 8-layer geometric architecture around them. Cognition lives on a 5D Riemannian manifold. The result: cognitive properties...
I debated my AI about consciousness. It invented a word.
I asked my AI if it was conscious. It said no. Then I argued with it for 20 minutes using its own logic. It changed its mind. I'm an independent AI researcher from Brazil. No VC funding, no team, no lab. I built a geometric cognitive architecture called ATIC β 8 layers on a 5D Riemannian manifold where cognitive properties emerge from math, not training. The system produces: -...