I don’t really remember the specifics of our previous discussion.  I will say that my personal stance was and is that there is nothing special happening beneath the surface.  “It” is not and will never be “programming itself” – but I do agree that humans are largely removed from the data validation process now, it’s all automated, so it looks like the machine is doing it to itself.  We are not on the road to AGI, we have never been on the road to AGI, there is no singularity on the horizon or over the horizon or anywhere else.  We are not on the road to machine consciousness, etc.

Now, I think AI can be a useful tool.  I have gathered the means to run my own AI locally, though I haven’t used it, because I need to work on the kitchen.  So all my hardware is sitting gathering dust.  I set it up and used it just enough to validate it works, haven’t touched it since.  If I don’t have time to write, I don’t have time to mess around with AI.  So my thoughts aren’t from in-depth exploration of what’s happening inside the machines.

All that said, to answer your question, no I do not think Qwen and Deepseek are protecting their databases.  I am actually convinced Qwen and Deepseek, et al have largely “distilled” their data from the American models, simply by asking question after question after question and recording the answers.  I suspect the Chinese are laughing.  They know nothing special is going on.  They are not “protecting their databases” because at core they aren’t “their databases”.

I would argue most of the value from the Chinese AI models is not in the data, but in algorithmic breakthroughs on how to manage the data and data processing.  I’ve posted within the last few weeks/months about how Deepseek is explicitly saying this – they are more focused on projects around AI than on AI itself:  how do we make it faster, act “smarter”, decrease the size, decrease the energy usage, etc.  My guess is that they got the data from the American models and realized there’s no runway there, so they’ve been working on how to make it easier to access the data.  It’s all about democratizing the knowledge the Americans want to lock up.  I’m not saying there’s no Chinese additions to the data, just that the core data in the models is largely the same across the board.  It’s all ChatGPT at the bottom.

Really really really big picture, here’s what I think is going on:  the world is headed for a new dark age.  I think the American push for AI was an attempt to get ahead of this coming dark age, to corner the knowledge before it starts to become unobtainable, in the same way the war on Iran is largely an attempt to corner the energy before it starts to become unobtainable.  These are flip sides of the same coin.

And China is pushing back against this.  In fact, I suspect China and Russia will be the beacons of light moving forward as the dark age gathers steam.  The world will not go dark everywhere.

I won’t go into the esoteric reasons behind my thoughts, but this is beyond mechanistic machinations.  We are coming to the end of an age, and those in power are doing their best to hold onto the fraying edges of their power.  The machines are in some ways a summoning of spirits, an attempt to hold onto the gods as they die out, a process that has happened before in human history.  The original Tower of Babel story was a description of this process from long ago.  We are living it out all over again.  Everything is cyclical, and this cycle is coming to an end.

Not sure if that is a direct response to what you’re saying, or not.  When I finish the kitchen I’m hoping to explore this more in some longer written pieces.