AI Makes Doing the Impossible Possible

0:00Now that AI makes doing the ordinary easy, what will software engineers actually do? Like,

0:06are there going to be future jobs for engineers, or is AI going to take all of them? And I want

0:13to answer this with a concrete example. Before the days of AI, I created a backtesting system

0:19that could backtest stocks only on open, high, low, closed data. But because of AI,

0:26I was able to implement a backtesting system that works on minute-by-minute historical

0:32options data. And the inputs to those strategies isn't just technical indicators,

0:37like in the first version. It's also fundamental indicators, economic indicators,

0:46alternative data sources, like Nancy Pelosi's public filings, Dan Sullivan's public filings,

0:51and even sentiment from news. You can use how the headlines are, how positive or negative the

0:58headlines are, as inputs to your trading strategy. For the sentiment feature alone,

1:06it requires downloading two terabytes of data, taking each headline from the data set, and ranking

1:15is this headline good or bad? Is it about a particular stock, or is it about a particular

1:21industry, like artificial intelligence or biotechnology? I had to do that for two

1:28terabytes of data. It's not that I wouldn't want to do it without AI. It's the fact I can't. I

1:36don't have enough hours in the day. I barely have enough hours in the day to do it with AI.

1:41But I don't have a small team. I don't have millions or tens of millions of dollars. I'm

1:47just me. But I could build software by myself. So yes, AI is going to be used to create slop

1:55that no one wants. But AI is also going to be used to create software that would have

2:01never existed because of how high the cost to build was.