Old news but pertinent based on recent events. Unless “AI” becomes more than a probabilistic model, it’s not going to be anymore useful than any person’s gut instincts trained on millions of years of evolution. Gut instincts, like models which generalize poorly to edge cases, are famously wrong all the time.

And by national data, he means personal data on its citizens. There are use cases where a centralized data warehouse for government would be helpful, but this is not that.

Yeah except they are falsifying national data and no longer collecting a lot of it. Garbage-in...

Sure, because he wants to convince world leaders this is great idea and then sell them a massive Oracle Cloud contract to analyze all this data with AI.

If he really wants to crack down on fraud, perhaps the government should unify all corporate financial transactions and chat records and feed it to AI.

Imagine what interesting things the tax authorities, the regulators, the enforcers of securities law could find that way.

And if they find a bunch of false positives, well, you can't make an omelette without breaking a few eggs, right?

The more I hear of Ellison’s rise, the more I am reminded of Theil’s easily mocked warnings.

Whether we like it or not, China will igore our IP and data laws, which will give them a clear technological advantage when it comes to AI.

some of these comments get disappeared.

So the AI bubble is still gonna be here for the next year and a half I guess, it'd be better if it popped sooner but situations surpass men so we gotta make do

Post from February.

Some discussion then: https://news.ycombinator.com/item?id=43038190