Geffen succeeded with a gift of $100 million to Lincoln Center and — perhaps more importantly — Lincoln Center paid $15 million to Fisher’s descendants so they would not sue. What that means is that the most prominent cultural organization in New York City lit $15 million on fire so that Geffen’s name would be on a concert hall.

No they did not lit them on fire, they payed of people.

In order to lit money on fire you need to buy something - like servers, electricity - and then just waste it. For example by running crypto schemes.

But they make up for it in volume!

So they named the product sucking the data after the Facehugger? At least they know that they are in the abomination business. Will they be releasing an AI named Bursting Chest?

I have noted two AI companies going belly up with earnings in a year matching costs per month. So I assumed that was around the worse case scenario, and for not yet bankrupt AI companies earnings were probably a bit better, perhaps just losing ten times their earnings.

I now see the flaw of my reasoning. Capital isn't allocated on profits, it's allocated on hype. Having profits draws the company down because it's no longer pure hype, and thus doesn't contribute to the hype bubble the same way.

So existing, not yet bankrupt, AI companies probably has significantly worse cost to income ratio than twelve.

A Danish ad company made a Google interface that they called "impersonal me" which searched Google with no personalisation. And not only was it better than Google search, it found things that normal Google just didn't show. In particular old comments I had written and lost track of. In the impersonal search they were easily found, in the normal search they weren't way down on the list, they weren't in the list at all.

Fascinatingly bad.

Sounds like something autocomplete would make up. Are we sure that is a real person this time?

Machine made t-shirt, with extra fingers.

Besides, isn't most clothes just made by poor people in poor conditions instead of being made with machines? Just like AI.

Good article. Captures the bubble growth and the lack of profit growth, with lots of examples. And that the capacity growth of AI is limited by non AI works, so no growth into functionality.

Good one to hand to people who needs to understand the nature of the bubble (and that it is a bubble).

And steer their careers into positions of influence.

Among the comments is an obvious rationaliser who claims that because [list of people in positions of influence] thinks AI Doom is real, this can't be a cult. Guess one has to be a rationaliser not to figure out how a cult that tries to place its followers into positions of influence can have many people in positions of influence.

Reading this extract I was surprised, figuring the desperate search for investment money must mean that Kickstarter wasn't turning a profit despite a seemingly sound business model.

Reading the article I found out it was even stupider. Kickstarter wasn't lacking a path to profit, it was lacking a path to growth. And being a profitable company with a clear market nisch isn't cool enough. Everything has to grow, grow, grow. So Kickstarter created a bunch of problems for itself, destroying much of its brand. It's that stupid.

You don't even need to be a utility monster.

Applying standard EA logic with utilions (approximately 1 utilion = 1 dollar) shows that when SBF was free he caused billions negative utilions. He says he did nothing wrong, and presumably he would continue to do nothing wrong. After updating our priors on the consequences of SBF doing nothing wrong, we can conclude that the risk is above 99% that SBF doing nothing wrong will cause billions in negative utilions. So the only utilitarian thing to do is hand out a life sentence.

Fortunately for SBF the judge is probably not in the business of creating philosophical justice.

Knowing just a smidgeon about how the statistical parrots work, I wonder were they will get the dataset for the animal languages.

This reminds me, I read an article in Nature about teaching dogs to read. Now, this was a 19th century article in a 19th century Nature, so it described how the author had written "food" on a note and placed it on the food bowl and placed a blank note on an empty bowl and eventually gotten his dog to fetch the note that had "food" written on it. Alas, due to unforeseen circumstances, it was hard to expand into more advanced literature.

So where to get the dataset? Nevermind, Magical AI to the rescue!