The one who is truly special is Ilya Sutskever. Not Sam.

Ilya is the major researcher here. Sam is a Ycombinator guy that knows about startups. But the AI guy is Ilya. He's the real brains behind OpenAI.

Pick Sam over Ilya is like picking a bean counter over the original inventor. Sam isn't the "Steve Jobs" in this story. He's John Sculley.

Picking Sam over Ilya would be a grave grave error.

The real issue is whether you want to keep Sam who's a money guy, a startup guy, and someone with no clue of how to make this technology advance. Or if you want to keep Ilya Sutskever who is the BRAINS behind OpenAI.

Ilya has been an author on some of the most important papers of the last 10 years including the ground breaking AlexNet. And continues to be a major contributor to the advancement of Large Language Models.

Sam can be replaced. Ilya IS OpenAI.

So watch what you wish for.

So a classic way of solving this is to systematically throw convolutions or cross-correlations at the data until you see something visible. Basically to search the space of convolution/cross-correlation functions until you find one that fixes the data.

Well, guess what? That's exactly what a convolutional neural network does.

Yeah, headline should be: MBAs scratch their heads and don't know what to do.

MBAs realize they don't really know much about products or how to ramp up production in a new competitive world. Expecting to make money of the coattails of innovators that died 100 years ago, they can't explain how they can't make money doing new things. In response, they suggest that we go back to the old way of doing things that made them money for the last 100 years.

Henry Ford was so inspiring, lets stick to his tech. We know that'll work. Building batteries would require us to actually have to take risks and work for our bonuses.
EVs are broken.

So we have companies that are far behind like GM and Toyota saying "EV doesn't work".

Meanwhile Tesla raking in millions and growing sales each year.

Yeah, they're just trying to pivot to their advantage.

Max Tegmark most recent paper is about:

https://www.researchgate.net/s...

So I guess his raising the panic level makes him more relevant.

The main "danger" LLMs present is the dumbing down of public discourse with all this end of the world talk.

The very people making these "panicked" calls are the very people developing this tech.

I think they are fear mongering so that the government puts so many restrictions that the competition can't catch up.

It's very cynical.

"Learning" in the sense of Machine Learning is transformative, not derivative.

In fact, as others have stated, GPT's main value is it's ability to "understand" things, rather than its ability to recall facts or quotes from articles or books.

I thought such behavior is forbidden by the GPL.

"The problem in general with ad targeting is that "target by gender" and "target by age" are not valid. There is no reason to target this way, "

Methinks you don't know much about marketing.

In Marketing you literally always start by targeting people by gender, by age, by income and other factors.
Why?

Well, imagine you're selling tampons. Do you want to pay for all the times the ad is shown to men?
Or to menopausal women?

No. You want to target the ad to get to the people most likely to be interested in the product and buy it.

Insurance companies may have ad campaigns to attract people of certain age or gender as well. Certainly targeting tools could probably be used in a discriminatory way, but that's on the company advertising it shouldn't be on Facebook or Google.

If basic marketing is going to be rules "discriminatory" then we're in for some big changes. And we're going to start seeing a lot of irrelevant ads all over the place. And the cost of advertising will go up. FB will actually make MORE money from it, but it won't benefit us or the companies doing the advertising.

How much was the rent if 149M SAVED them money?

All this push for premature regulation isn't because anyone serious is scared about the "singularity".. what it really is, is a push to make the regulations so tight that small companies can't reproduce what they've done.

This sounds like a GDPR violation. As an EU member shouldn't they been fined?

If the climate isn't hot you could heat treat it by letting it cool off in pipes or in some uncovered pool and when it gets cold enough you dump it back into the river.

WaffleMonster is clearly too young to have lived through the waterfall era, or wouldn't say such silly things.

WF was terrible. And most people doing it would admit, with some guilt, that they weren't doing it according to "serious" WF principles.

Scrum is not the problem. The problem is bad managers that don't understand software development. Maybe going back to an oldie such as "Debugging the Software Development Process" by Steve MacGuire can give some insights on how to do it right. And then you do it right, and along the way when people ask you what you're doing, just say "Scrum" and all will be fine.

But not understanding how to do software development is not compensated by any methodology. What people need to understand (and the book by Steve McConnell, "Rapid Application Development' illustrates well) is that a Software Engineering manager has a set of tools in his or her toolbox. And can choose what tools to use. These include, daily scrums, unit tests, UI designs, code reviews, etc.. etc.. The correct mix is achieved according to what team you have. It has to be dictated by the capacity of the team, and the conditions of the project.

Also, businesses typically want estimates. So read McConnell's "Software Estimation: Demystifying the Black Art" so you can learn how to handle that business need as well.

Running a software project requires more know how, that a single Scrum or Agile course. But Scrum and Agile tools can be very useful in real world projects.