ENIAC

In the 1940s, McCullough and Pitts introduced the first mathematical neural network model in 1943.

Claude Shannon's 1948 information theory paper gave probabilistic language tools, and early computers like ENIAC provided the hardware that made AI research possible.

Those early ideas were just the beginning.

Want to hear how they evolved into the AI we know today?

They were always faster than the alternative.

Uh, incidentally, that computer in 1945 that was built, called ENIAC, was built to do artillery trajectory calculation, and it was built to do it faster than the alternative.

The alternative were humans who actually did the calculations.

Now, the alternative typically is some other computer.

Please, let's not buy into AI is going to kill us.

As a software developer myself, as someone who was in the software development from its, I don't want to say its inception because that started with ENIAC and all those vacuum tube, that's way before my time.

But at the time I came at the, when CPU started to take off, central processing units, microprocessors used to take off.

So I programmed from there.

You see over here on the right, Harvard Mark I. There's Grace Hopper and other members of her team.

This is the ENIAC on the bottom left.

That's the first computer at University of Pennsylvania.

Eckhart Mochley, Jean Jennings, six women were the first digital programmers in the United States over there.

Right.

ENIAC or whatever, whatever those computers were.

And so just the expectation of the interaction with the benefits that a participant has has just been totally raised.

So just the – it used to be that we would have to struggle with our clients to get them to do something more than just – uh, kind of Xerox, you know, a word processed summary plan description to, to give to, um, to their plan participants.

If you mean a modern data center campus, large purpose-built warehouse-like buildings packed with rows of servers and surrounded by cooling equipment, substations, generators, fences, and security, the cleanest starting point is the late 1990s into the early 2000s.

We can trace data centers back to 1945, when ENIAC occupied an enormous dedicated room at the University of Pennsylvania.

By the 1950s and '60s, governments, universities, banks, and large corporations had specialized mainframe rooms with raised floors, heavy cooling, backup power, and restricted access.

Someone from our time could walk into one and recognize the ancestor of a data center.

The recognizable campus appears around 2006.

The giant AI campus, the one measured in gigawatts and discussed like its own power province, is the much newer beast currently kicking down the door.

Google identifies The Dalles facility as established in 2006, while IBM's history traces the underlying data center lineage back to ENIAC and the 1940s.

I don't wanna get ahead of ourselves.