Does AI Know Everything?
Now... I'd like to move on to another topic that I don't hear discussed very often.
Whenever people talk about modern AI models, they usually say the same thing. They've been trained on an enormous amount of information. And that's absolutely true. The amount of knowledge available to these models is incredible. It's probably more information than any human could read during an entire lifetime.
But I think there's another side to this discussion that people often forget.
Just because AI has been trained on a huge amount of data doesn't mean it has been trained on all human knowledge.
Those are two very different statements.
The Internet Is Only One Library
Think about the Internet for a second. It feels enormous, and compared to twenty or thirty years ago, it really is. Every day millions of new pages, videos, papers and discussions appear online.
But even with all of that, the Internet is still only one part of human knowledge.
There are thousands of books that have never been digitised. There are old technical documents sitting in archives. There are notebooks, drawings and research papers that exist only in physical form. Some companies still have documentation that's twenty or thirty years old and has never been scanned or uploaded anywhere.
So when people say AI has learned "everything from the Internet," I think we should remember that the Internet itself isn't everything.
It's a huge library.
But it's still only one library.
Knowledge Hidden Inside Companies
Now let's go one step further.
Even if every public website on Earth suddenly became available for AI training, there would still be an enormous amount of knowledge missing.
Think about private companies.
Large corporations have millions of pages of internal documentation. They have engineering reports, production procedures, customer data, design decisions, incident reports and technical discussions that never leave the company.
Software companies have private repositories containing millions of lines of code that nobody outside the organisation has ever seen. Manufacturing companies have production processes they've spent decades developing. Pharmaceutical companies perform research that remains confidential for years.
None of that normally appears on the public Internet.
And honestly, that's exactly how it should be. Companies invest years and sometimes billions into creating that knowledge. It becomes part of their competitive advantage.
So even though AI has learned an incredible amount, there's still another enormous world of information that's completely outside its reach.
Knowledge That Lives in People's Heads
But here's the part I find even more interesting.
Some knowledge isn't stored in documents at all.
It's stored in people.
Imagine an engineer who's been designing aircraft engines for thirty years. During that time they've seen projects succeed and fail. They've made mistakes, solved unexpected problems and learned hundreds of small lessons that were never written down anywhere.
The same is true for doctors. Mechanics. Architects. Electricians. Scientists. Lawyers. Craftsmen. Every profession has experts who've spent decades building experience that exists mostly in their own minds.
Sometimes they can't even explain why they immediately recognise a problem.
They simply look at something and say, "This doesn't feel right."
That judgement isn't based on reading one document.
It's based on thousands of situations they've already experienced throughout their careers.
And I think that's a very important difference.
Knowledge and experience aren't the same thing.
You can read every book about riding a bicycle.
You can watch hundreds of videos.
You can study the physics behind balance.
But until you actually get on the bicycle and fall over a few times, you haven't really learned how to ride.
I think many professions work exactly the same way.
That's one of the reasons I'm not convinced AI can simply replace experienced professionals overnight. It has access to an incredible amount of knowledge, but that doesn't automatically mean it has the same understanding that comes from years of doing the work.
What Happens Next?
Now... does that mean AI will never reach that point?
Honestly, I don't know.
Maybe in twenty years we'll have robots working in factories, hospitals and construction sites every single day. Maybe they'll continuously collect new information through cameras, microphones and other sensors. Maybe they'll build their own experience instead of relying only on human-created data.
I think that's entirely possible.
But I also think we're not there yet.
Today we're still in a world where AI learns mostly from information that humans have already created. It's incredibly powerful, but it's still only part of the picture.
That's why I believe experienced professionals will remain incredibly important for many years to come. AI will become a better assistant every year. It will probably become a better engineer every year too. But there will still be situations where someone needs to say, "I've seen something like this before, and I think we're heading in the wrong direction."
For me, that's one of the biggest differences between having information and having experience.
In the next part, I'd like to talk about something related to this idea. If AI becomes better every year, what happens to the knowledge we're creating today? Are we reaching a point where models begin learning mostly from content generated by other models? And if that happens, what challenges might it create for the future?