Abstract luminous fibers — Latent field: thought, taking shape.

Artificial intelligence. Human possibility.

On a journey to humanity’s next chapter.

Exploring the ideas, models, and human questions shaping our future with artificial intelligence.

Explore the journal

The journal

Thinking in the open.

Ideas worth exploring.
Models, systems, and the human side of AI.

Speech & audioTechnical note

Teaching a model how to speak

A closer look at conversational speech fine-tuning with CSM and LoRA. The real question is what happens after the loss curve.

13 Oct 20264 min read

Language modelsTechnical note

From fluent to useful: a note on DPO

Fluent language is a starting point. Preference optimization asks a more specific question: which response is actually better?

13 Oct 20264 min read

ConversationsConversation

The human side of artificial intelligence

In Vyvo Voices, a conversation about the context, personality, and privacy of an AI companion.

26 Sept 20252 min read

AI & systemsInterview

Privacy starts before the model

Data integrity, consent, and personal AI. An interview on connecting wearable signals to intelligent systems.

01 Aug 20253 min read

AI & systemsPerspective

An AI landscape that won’t stand still

A look back at a fast-moving AI market and the role of open models in the VAI OS strategy.

30 Jan 20252 min read

From ideas to models

Open lab

Selected models, shared in the open.

Visit Hugging Face

Conversational voice

CSM fine-tuned on English podcast speech.

Preference, learned

GLM-4.7-Flash adapted with DPO and TRL.

Beyond language

A Qwen3-VL-30B-A3B-Thinking fine-tune.

Hakan Kozaklı

I’m interested in the space between a capable model and a useful system. This journal explores language, voice, personalization, and data ownership—one note at a time.