I feel bad because I feel that I haven’t progressed much in Augmented Awareness development. This is a problem. I mean, this feeling is a problem.

Productivity has evolved into a cultural obsession, mirroring the structure of religious devotion. Like traditional faiths, it demands adherence to rituals, fosters guilt when unmet, and promises salvation through measurable outcomes.

Corporate culture fuels this dogma. Companies like Amazon and Tesla glorify "hustle culture," where burnout is rebranded as "grit." Employees internalize this, working overtime to avoid being labeled "lazy." The cult of productivity even has its prophets: Tim Ferriss’s The 4-Hour Workweek and Cal Newport’s Deep Work are treated as sacred texts, their methods seen as moral imperatives.

I don't want to fall for it. Essentially, the obsession with productivity feeds on fear: fear of being idle, fear of becoming irrelevant. A personal project should never be a victim of these feelings.

Another issue is that I don't like the current code. It's messy, and I haven't paid much attention to git etiquette because it's mostly throwaway code for testing ideas. Improving it isn't my main focus right now. What's important is deciding the direction I want the project to take. Having messy code on my GitHub stresses me out.

Where is Augmented Awareness (Aww) going?

Let's first remember what Aww is supposed to be. It's designed as a robot or agent that interacts with an environment by observing it, understanding it, deciding what to do to achieve specific goals, and acting on those decisions. The environment includes my life: both my mental and physical states, as well as the physical and media environments I live in. The goals are my own. The actions are the information that Aww shares with me through dashboards, notifications, and nudges.

We have many systems in the world with similar designs, but they are controlled by third parties, giving them the power to influence our lives. I'm referring to all the notifications apps try to push on us, the promotions we receive through emails, the advertising we see, and all the information fed to us.

I have currently implemented reading my journal, schedule and tasks from my Obsidian vault, and obtaining raw data from activitywatch. Both data sources can be converted into Arrow datasets.

The plan is to use my Obsidian journal as a source of labels and activitywatch as a source of features to automatically label my activities using semi-supervised machine learning.

I have also tried using LLMs to read my journal and provide insights, which has been helpful for my weekly retrospectives.

I am starting to experiment with the Model Context Protocol (MCP) to see if Aww should be implemented using it instead of relying on its own tightly-integrated code and control loop.