We write to become less ignorant, and to build competence, sanity, and taste around things that feel early, strange, or important.

Why

We are ignorant.

We know that we are ignorant.

And when we tried the “jump into the water to learn swimming” theory, we realized that we do not even know how deep our ignorance is.

Some founders say they would not have started their companies if they had known how hard things would become along the way. So maybe starting immediately, once you see an opportunity, is a good idea. But today, things move extremely fast. If you do not see the opportunity far enough in advance, experienced engineering teams with more momentum will beat you almost every time.

Their victory is not directly against us. Still, we lose directly against them.

By “them,” we mean companies like Google, Cloudflare, Anthropic, OpenAI, and others. Even if an idea is brilliant and achievable for a small team like us, these companies can dominate the narrow gap we see in the industry within days, especially with the help of agentic agility. Even when something is only slightly outside their roadmap, they can move into it quickly. So we need a way to deal with the frustration of being beaten by some of the greatest institutions of all time.

We are not a legendary visionary team at the time being. Finding a gap, seeing a problem, convincing brilliant and hardworking people, and starting immediately is not enough. Not anymore, or not yet. And we need to beat our incompetence and ignorance before our taste can reliably turn into products.

What we do not know is almost everything. So trying to map our inabilities, or the full shape of our ignorance, does not feel time-worthy. It is frustrating, endless, and rarely gives something useful in return. This is the part we agree with those founders on: knowing too much about the difficulty ahead can kill the attempt before it starts. We have ideas that we believe may become important in the future. Betting on our taste is probably the best thing we can do.

A couple of months ago, we would have agreed that the right thing was to work on the idea we were most ambitious about and create something tangible around it. Today, because of the reasons above, that method often creates demotivation and breaks our agility. Also, we have a shit ton of ideas.

So instead of trying to build every idea immediately, we want to identify the basic topics that make those ideas important to us, and do practical research around them. We want to understand what might be just beyond the edge of the current industry timeline. And to counter the demotivation, we want to share what we find along the way, elegantly and honestly.

Looking back and seeing that we worked on things before they became barely real feels deeply satisfying.

That is why we want to share things. They may be called blogs, essays, notes, or something else. The label does not matter much. We want to do this to increase our competence, slightly beat our ignorance, debunk hype, and, most importantly, protect our mental health.

These pieces will be about anything we find interesting at the time. A topic may be hyped, controversial, technically strange, or simply beautiful to us: personal twins, agentic memory problems, IPA transcriptions, LeCun’s world-model bet, Gaussian splats, and whatever else keeps bothering us.

We hope they will be useful in the end.

Taha, Ömer, Barathan