by iqentity | Jun 21, 2024 | Development, Tessera
Tessera must know what she does not know. This week I built the confidence calibration system, and it is one of the most important architectural decisions in the project. The Problem With Confident Systems Most AI systems are uniformly confident. They present every...
by iqentity | Jun 17, 2024 | Development, Tessera
Semantic search is Tessera’s primary retrieval mechanism for conceptual matching, and the embedding strategy required careful thought. Off-the-shelf embeddings are trained on generic text. Tessera’s corpus is anything but generic. The Domain Collision...
by iqentity | Jun 13, 2024 | Development, Tessera
Theory is comfortable. Production is not. I have been running Tessera in parallel with my normal workflow for six weeks, using it during actual client situations and comparing its output to what I would have done without it. The results are encouraging, humbling, and...
by iqentity | Jun 4, 2024 | Development, Tessera
The prompt engineering for Tessera is unlike any other prompt work I have done. I am not designing prompts for a general-purpose assistant. I am designing prompts for a system that needs to think like me, respond in patterns I find useful, and maintain a level of...
by iqentity | May 23, 2024 | Development, Tessera
Tessera will get things wrong. Not occasionally. Regularly. The local language model hallucinates. The graph traversal sometimes follows a misleading path. The enrichment pipeline misclassifies decisions. The salience model under-weights artifacts that turn out to be...
by iqentity | May 19, 2024 | Development, Tessera
Most AI systems are trained on outputs. What was the answer? What was the decision? Tessera is also trained on corrections. Where did I change my mind, and why? Why Corrections Matter More Than Conclusions A system that only learns from final decisions becomes...
