'Godfather of AI' quits Google

Geoffrey Hinton, father of deep learning, says that the new generation of large language models—especially GPT-4, has made him realize that machines are on track to be a lot smarter than he thought they’d be. He’s scared about how that might play out, so he has quit to do 'philosophical work'. ??? Hinton is talking about “few-shot learning,” in which pretrained neural networks, such as large language models, can be trained to do something new given just a few examples. This means that these language models can string a series of logical statements together into an argument even though they were never trained to do so directly. Will our edge as humans truly vanish? #jobs #ai #deeplearning

It's not often someone changes their mind about something they've worked on for more than four decades. But Geoffrey Hinton - pioneer of modern AI, Turing Award winner and (as of Monday) former Googler - has done just that. He spoke candidly with me about the recent revelation he's had thanks to the latest crop of large language models like GPT-4. They are so good, he says, that he now believes machines are - or very soon will be - smarter than humans. In effect, Hinton thinks there are now two forms of intelligence in the world: human and digital. And the digital kind might just be better. “These things are totally different from us,” he told me last week when I visited his house in London. “Sometimes I think it’s as if aliens had landed and people haven’t realized because they speak very good English.” What's the upshot? Hinton says he is now scared of the technology he helped usher in. Hinton's new ideas have been brewing for a few months. But he waited until leaving Google to speak openly about them. Not because he has dirt on Google; but to avoid self-censorship. He simply wants to speak his mind. And, wow, it was remarkable to hear him out. I don't agree with everything he says at all. But when someone like Hinton has something to say, it's worth listening. You can read my in-depth interview below. And follow MIT Technology Review for all things AI. #ai #google #technology #interview

An interesting yet worrisome perspective on how AI might change our lives in the upcoming years, delivered by one of the people that gave birth to it. Geoffrey Hinton, a computer scientist who's been active for over 50 years in the artificial intelligence field, just left his job at Google after 10 years of working on the development of deep learning techniques and new AI applications and the reason why he did is concerning: with the widespread adoption of LLMs and the arrival of the first AI Agents (AI applications that can autonomously execute tasks and acquire knowledge without continuous human-feedback; popular examples include AutoGPT or BabyAGI), he realised that machines will be smarter, more efficient and more powerful than humans much earlier than he expected. This doesn't just apply to learning but to communicating as well. "If you or I learn something and want to transfer that knowledge to someone else, we can’t just send them a copy,” he says. “But I can have 10,000 neural networks, each having their own experiences, and any of them can share what they learn instantly. That’s a huge difference. It’s as if there were 10,000 of us, and as soon as one person learns something, all of us know it", he mentioned, hinting that we are subtly standing in front of a completely new form of intelligence. Given his experience and competence, Hinton's pessimistic take on the subject should certainly be taken into account. AI might be the technology that will turn us into super-humans or the one that haunts us down, we don't know yet. What we do know already, however, is that it will impact our lives in one way or another, so we better be prepared for when that happens! #ai #technology #intelligence

Here it is. An exclusive talk Will Douglas Heaven had with Geoffrey Hinton last week ahead of the news that he's leaving Google—and focus on raising awareness of the dangers of super intelligent AI. Great read that gives so much more context and background than the NYT piece you might have seen. Also has some hair-raising passages like: Hinton now thinks there are two types of intelligence in the world: animal brains and neural networks. “It’s a completely different form of intelligence,” he says. “A new and better form of intelligence.” Read it all here—and don't miss Will's live interview at EmTech Digital this week —tickets still available. https://lnkd.in/eHAVrM2i

"I don’t think [researchers] should scale this up more until they have understood whether they can control it.” Those are ominous words when they come from Dr. Geoffrey Hinton, a pioneer in the world of artificial intelligence who just resigned from Google so he can "speak freely" about potential risks posed by AI. Hinton, who helped create some of the fundamental technology behind today's generative AI systems, fears that the tech industry's drive to develop AI products could result in dangerous consequences—from misinformation to job loss or even a threat to humanity. Hinton joined Google in 2013 after Google acquired Hinton's company,DNNresearch. His departure a decade later marks a notable moment for the tech industry as it simultaneously hypes and forewarns about the potential impact of increasingly sophisticated automation systems. His long-term worry is that future AI systems could threaten humanity as they learn unexpected behavior from vast amounts of data. "The idea that this stuff could actually get smarter than people—a few people believed that," he recently told The New York Times. "But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that." Some critics have cast a skeptical eye on Hinton's resignation and regrets. In response to The New York Times piece, Dr. Sasha Luccioni of Hugging Face tweeted, "People are referring to this to mean: look, AI is becoming so dangerous, even its pioneers are quitting. I see it as: The people who have caused the problem are now jumping ship." Read more via Ars Technica. https://trib.al/vXKp3D0

Yes, I read this article, and yes, I agree with everything in it. Releasing generative AI could be dangerous. It could make it difficult for the average person to know what's real and what's not real. It could replace rote, monotonous jobs and unemploy millions. I feel like I should apologize for what I'm about to say but I won't: (1) so what? (2) really? and (3) GREAT! Wow. That was insensitive. Let's continue. 1) I was in high school in 1996 when they gave teenage me and millions of other people access to something called the WWW. I probably used this powerful 'information superhighway' to download naughty pictures. I'm certainly the only person who did this, but it was inappropriate and dangerous. I also read a lot of stuff my parents didn't know about. I used a chat tool to talk to girls. And strangers. Crazy stuff. Almost no one in my grandparents' generation knew this thing even existed, and it changed just about everything about my personal and professional life and probably yours too. You're reading this on that! 2) I definitely agree that it will make it hard to tell truth from fiction, but did you know that until very recently the most-watched media personality was (is?) Tucker Carlson? Do you know how many millions of people believe totally false things? About half the US believes the opposite of whatever you believe about everything. We are way past the point about worrying about how #AI is going to make #2 less of a #2-show. I'm not sure why anyone cares what anyone thinks is true anymore. 3) This is super mean, but society has always had jobs that were replaced by improvements in technology. People used to manually connect telephone calls. They also used to manually answer them when you called somewhere! Typing pools. Assembly lines. Bank Tellers. New jobs are always created and they are usually better. But also, humans might have a lot less routine, boring work to do. #3 is probably going to be amazing. It also might allow YOU to do things you can't easily do today. Medical research is super hard and seeing doctors is expensive. You have to pay a lawyer to correctly draft legal documents. You might not know how to write any kind of computer code. But what if you could do all those things "yourself" with AI enhancement?

Wow. Geoff Hinton, one of the most important founders of AI, leaves his job so he can speak out about the potential for harm. "I console myself with the normal excuse: If I hadn’t done it, somebody else would have."

Thought provoking interview with Prof. Geoffrey Hinton who is commonly referred to as the “godfather of AI” has now left Google in the recent days. In his interview with CBS News highlighted two concerning points. 1. AI is more prevalent than ever, with OpenAI, Microsoft and Google all offering easily available AI tools. I don't think the technology we're building today naturally leads to artificial general intelligence, but asked to exercise caution than ever as it may likely be a reality in the near future. 2. The bigger issue, he said, the technology could give a handful of companies or governments an incredible amount of power. Good points to reflect on. #aitrends