It has been a while since I wrote on this blog.

Recently, everything around AI has been incredibly interesting. Claude, ChatGPT, Codex, Claude Code, Cursor. When I use them almost every day, coding AI no longer feels like a special tool. It feels like part of the foundation for work and creation.

There are three things I want to write about this time.

Why I think SpaceXAI and xAI will become very strong. Why we are entering an era where countries limit AI. And what we ourselves are going to build.

The Strength of SpaceXAI

Recent coding AI tools are truly amazing. I use Codex, Claude Code, and Cursor every day.

I think three things matter most for AI: electricity, a great design team, and chips.

In that sense, the world around SpaceXAI, xAI, X, and Tesla looks extremely strong.

Rockets, space infrastructure, data centers, Grok, real-time information from X, and robots. They do not feel like separate things. They feel like they could connect somewhere.

On Earth, data centers are facing bigger and bigger problems with electricity and cooling. Trying to solve that from the angle of space and rockets is already impressive to me.

Even if not everything happens immediately, I think it is amazing that they are seriously putting out that kind of idea.

Companies With Data

The model itself matters, but I also think the information a model can learn from, and how real-time that information is, matter a lot.

That is why X feels so strong. It has real-time information, a long history of posts, and human reactions. I feel like it may be one of the strongest information sources after Google.

OpenAI and Anthropic are amazing companies, but they do not own a huge information source by themselves in the same way. That means they can be affected by restrictions, partners, and external relationships.

Google, on the other hand, should be incredibly strong. Search, YouTube, Gmail, Calendar, Android, apps, services, researchers. It has everything.

And yet, when I use Gemini, I sometimes feel that it is wasting that position. Once, I asked a lightweight model, "Who was Oda Nobunaga?" and it gave an answer that made me seriously question what was happening.

Data alone is not enough. Integration, design, and the product experience all have to work together. If Google becomes more serious, it could be amazing, but GPT, Claude, and Grok will keep moving forward too.

This race is interesting even just to watch.

When Countries Stop AI

Another big thing is that we are entering an era where countries limit AI.

When I saw the news around the temporary suspension of Claude Fable 5 and Mythos 5, I was very surprised.

But honestly, I had felt since around last winter that something like this might happen. As large AI agents grew, I thought governments would eventually step in.

Now that kind of movement is actually happening.

I cannot simply say that government limits are good or bad. But this moment made me feel that we need to discuss much faster how AI should grow, and how we should think about model capabilities and training scope.

If AI keeps growing without enough thought, it could eventually move in a dangerous direction. That is why safety research and rulemaking need to grow alongside agent capabilities.

Our Turn

While the AI world is moving quickly, we are also building projects called Copernicus Sust and Two Brain, one step at a time every day.

I really believe these projects can become something wonderful.

I want to improve how AI is used and make people happier. In a bigger sense, I want to move closer to helping humanity.

I will share more soon. This month, I also want to create a bigger master plan.

I still think this is the year of AI agents. There is talk about Grok 4.5 on July 9, 2026, movement around the GPT-5.6 series, and Claude may bring something new too.

Recently, checking app updates and X after waking up has become part of my morning routine.

I think that will probably continue for a long time.

And someday, even when I am not looking, AI will report what matters to me. That is the future we are building.

Happy Two Brain.