Collective observation

I’ve started thinking of a computer as a collective-looking machine.

There is too much happening for any one person to see the whole thing. So we build a machine and say, “Look at more of it for me. Find the patterns. Help me understand what’s happening and work out my next move.”

I think a lot of computing comes back to that. Collective observation. We gather information, recognise patterns, make predictions and try to improve what we do next. As individuals, as businesses, and eventually as a species.

Now look at a large language model. We’ve built this mathematical behemoth which can have billions of little adjustable knobs. Those knobs are its weights: numbers shaped by learning patterns in data. So much of that data comes from things humans have written, made and discovered.

All of us together have helped midwife this thing. Generations of people creating knowledge, building machines and leaving something for the next person to build on.

And now I can use it.

Even this journal entry is an example. I have a thought, probably a bit unorganised, and I give it to the machine. “Sharpen this. Help me explain what I’m seeing.”

It feels like taking an idea to a vast collection of human thinking and asking it to help me make that idea clearer. I still have to examine what comes back. The machine can produce a convincing mistake too. But the ability to have that conversation whenever I want feels like magic.

That’s the leverage I’m excited about. One curious person can draw on something much larger than themselves to learn, build and express an idea.

Then there are the bigger questions. Could a machine like this become sentient? What goals should we give it? We still have to decide what we want it to help us do.

I don’t have the answers. But I look at what we’ve collectively created, and I think this could be one of the best things that has happened to humankind.