Live data from Hacker News

Mark Zuckerberg’s new goal is creating artificial general intelligence

theverge.com

241–250 of 362 posts

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#243
post #165

Earlier quoted context omitted.

What does irresponsibly open sourcing something look like?

Depends on the thing, detailed instructions of how to build nuclear weapons and enrich uranium is one example.

So in your opinion, how should we determine who should have this knowledge?

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#245

Earlier quoted context omitted.

In 2007 or 2008 or so, when I saw that they had implemented a feature that let people tag you in photos without permission, I assumed it was to gather data for identifying people with AI at some future date (albeit based on my rather different understanding of AI at the time). So, I deleted my account: FB didn't provide me any value, and had introduced an unknown future risk by taking away control of my identity. Eve…

> Ever since then, I've asked friends and family not to upload pictures of me to Facebook, with the natural amount of awkwardness and eye-rolling that request evoked. In the future, systems will probably infer what you look like based on connections. DNA relatives, etc.

How accurate do you expect that to me? If I tell you two people have children how well do you think you can predict what that child looks like?

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#246

Whatever happened to replacing the US dollar, or creating a metaverse, or curing disease in his children's lifetime? The guy has all the focus of a disco ball.

> Now, Meta CEO Mark Zuckerberg is entering the race. If this goes the same way as when Zuck entered the race to replace the US dollar, create a metaverse or cure disease, Ai will be dead in the next 2 years.

We can hope.

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#247

Can someone provide insight into why there's so much insistence from business that magic happens at scale with LLMs? We're a long way from AGI. The lack of meaningful details in these announcements makes me pessimistic.

> why there's so much insistence from business that magic happens at scale with LLMs It's already happening. See latest Google layoffs. They are automating a lot of things. Most people don't realize it, but the change is going to be dramatic. > We're a long way from AGI This is a big question, what is AGI? LLMs are quite generic and 'intelligent'. Not human-like, but. Next is going to be incremental evolution. Till w…

Boy, we must have a totally different understanding of whay AGI is.

Intelligence is not about parroting an answer you've seen before. It's about using your environment to gain an evolutionary advantage.

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#248

Whatever happened to replacing the US dollar, or creating a metaverse, or curing disease in his children's lifetime? The guy has all the focus of a disco ball.

Companies pivot, this isn’t a bad thing

Sure but you have to actually establish something first to pivot from. Declaring some half-baked idea and then not executing and then replacing that idea with some new half-baked idea is not actually pivoting. It's blathering. See:

"Facebooks Endless Pivots":

https://www.axios.com/2022/05/04/facebooks-endless-pivot-met...

Has Facebook as a business ever really pivoted from "social network advertising company?"

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#249

Whatever happened to replacing the US dollar, or creating a metaverse, or curing disease in his children's lifetime? The guy has all the focus of a disco ball.

To be entirely fair, the currency attempt was hampered by governments (predictable, maybe), and as for the metaverse... well he created something — technically it's impressive, commercially and in terms of product market fit, it's a flop.

I don't think they're done with the metaverse thing either, whether it will succeed, that's another question.

Re: Mark Zuckerberg’s new goal is creating artificial general intelligence

#250

Earlier quoted context omitted.

> Ever since then, I've asked friends and family not to upload pictures of me to Facebook, with the natural amount of awkwardness and eye-rolling that request evoked. In the future, systems will probably infer what you look like based on connections. DNA relatives, etc.

How accurate do you expect that to me? If I tell you two people have children how well do you think you can predict what that child looks like?

1. With the child's complete DNA or key markers, I bet we'll eventually have algorithms for DNA -> adult phenotype

2. With only parental or relative DNA, if we are able to gather an enormous amount of extra information -- social group, gym use, hobbies, dating circles (mate selection probably yields a huge amount of info!), profession, fashion purchase activity, etc. -- then I bet we can get pretty reasonable confidence intervals on a lot of physical characteristics. Height, eye color, eye distance, face shape, lip thickness, etc. etc. Perhaps even elements such as voice timbre, etc.

If we had their familial phenotypes, photos and phenotypes of all a person's dating history, then I bet that lets us cut the search space down tremendously. If we add music listening habits and hobbies, I bet we'd also be able to cut it down.

I'm hypothesizing that physical appearance might correlate from behaviors, social graph information, and limited DNA information. I have nothing to show for that, however. (I would love to read literature on this topic.)

I'd imagine that constructing such a predictive model would require an enormous amount of data. Facebook + 23andme could probably pull it off.

Post reply on HN