assets The Brainstorm AI conference has just wrapped up in San Francisco, and I’m still processing the litany of intriguing tools and concepts discussed by some of the leading minds in the field of AI, as well as the many cautionary tales from those those who do understand the power of this technology better than anyone else.
Generative AI was the star of the show, continually popping up in discussions on stage, informal conversations between participants in the hallways, and after-hours dinner conversations.
Since OpenAI’s new ChatGPT had only been released days before, it seemed like every guest who came to the conference had a story to share about their experience playing around with the tool or thoughts on how pivotal this moment is for artificial intelligence tell.
“It still feels surreal to be born into this period of history and to be in the middle of this technology,” said Fei-Fei Li, co-director of the Stanford Institute for Human-Centered AI, of recent innovations like ChatGPT.
Microsoft Chief Technology Officer Kevin Scott said that ChatGPT is showing the world how far AI has come and all its possibilities. “One of the things it’s doing is really opening the door to who can use AI now,” Scott said during a discussion on stage Tuesday.
According to Scott, big language models like the one used by ChatGPT provide foundations on which new products and services that harness the power of AI can be built, without anyone having to have a computer science degree.
“We’re now able to invest in building these really big things, and then thousands, tens of thousands, hundreds of thousands of people can use it to build all sorts of incredible things,” he said.
Of course, generative AI is not just about generating text. Image generation is progressing at an equally dizzying pace, and several examples including OpenAI’s DALL-E 2 and Stability AI’s Stable Diffusion were presented at the conference.
Emad Mostaque, CEO of Stability AI, gave the audience a first public look at the latest version of the company’s image generator, Stable Diffusion 2.1. The image generator understands 2 billion different concepts, which it can turn into images at ever-increasing speeds — by next week, Mostaque said, it will generate 30 frames per second, compared to an image in 5.6 seconds for the first version of the product published in August.
And while Stable Diffusion is open-sourcing its technology, Mostaque’s ambitions for the business potential of its product are sky-high.
“Every single creative industry in the world will have to react to this. As of next week, that’s 30 frames per second, which is basically video,” Mostaque said.
Mostaque showed the tool’s on-the-fly interface by asking it to create an image of Eye on AI’s Jeremy Kahn as an astronaut (he noted that the tool was pre-fed with 4 photos of Jeremy to complete the task enable). In fact, Jeremy’s smiling face soon appeared in a realistic-looking space suit, with Earth floating in the background.
When Mostaque next asked it to turn spaceman Jeremy into a Van Gogh painting, the tool demonstrated some of its remarkable capabilities and limitations. It quickly spat out the astronaut’s face, rendered in the broad, vibrant brushstrokes associated with the Dutch painter – unfortunately the face in the picture was no longer Jeremy’s, or at least not our Jeremy Kahn’s.
Read on for more news and highlights from the Accenture-sponsored Brainstorm AI conference.
Alexey Oreskovic
@lexnfx
[email protected]
INSIGHTS FROM THE CONFERENCE
On the potential dangers of AI:
“We shouldn’t totally delude ourselves about the negative impact and the human aspect. So it’s a mix of excitement and responsibility.”
— Fei-Fei Li, Co-Director, Stanford Institute for Human-Centered AI
“I think open source will have good uses, but I think this is one of the areas where we have to be careful when choosing. Because open source is used by people. And some people create great things like education tutors and other things and others create spam bots or other things that are harmful. So you’re trying to shake it off for good, not away from it.”
– Reid Hoffman, Partner, Greylock
For industry regulation and accountability:
“This may surprise people, but I think you’ll find that many of us embrace regulation because we have to think, ‘when is it appropriate to use these technologies?’ How do we put them out into the world so we don’t throttle them, but make sure we use them in the most useful way, the most appropriate way, with enough oversight and thoughtfulness.”
– James Manyika, Senior Vice President, Technology and Society, Google
“All of us in the industry need to figure out how to work on it, not just say it ‘Oh lookme regulator that does not understand the techhnology that works somehow magical make This happens.‘”
– Reid Hoffman, Partner, Greylock
On the rapid progress of AI:
“One thing we see over and over in technology it is what was Yesterday’s mainframe is today’s smartwatch. and it was like this Interesting to See how once sOhine shows that it is possibleble, Tons of others are starting to replicate it.”
— Andrew Ng, Founder and CEO, LandingAI
“The next few years are going to be one of those transformative moments in technology. Think iPhone, think Mosaic browser.”
– Reid Hoffman, Partner, Greylock
What the AI industry is doing wrong:
“Where I see the AI most stuck in the long tail of this distribution… As soon as I look outside the internet for consumer software, I see a lot of $1-$5 million projects that nobody is working on.
— Andrew Ng, Founder and CEO, LandingAI
ASSETS ON AI
“We become their visual storytelling partners”: How artificial intelligence is changing the film and entertainment industry, by Alena Botros
Google SVP says AI breakthroughs are accelerating but search giant needs to reflect on ‘extraordinary range of risks and challenges’, by Kylie Robison
Sasan Goodarzi, CEO of Intuit, explains how AI transforms TurboTax’s parent company into more than a “transactional” platform: it “actually does the work for you”, by Paolo Confino
BRAIN FOOD
Meet your new co-pilot. Another hot topic at the conference was the notion of co-pilots – AI tools designed to complement rather than replace human jobs.
Microsoft’s Kevin Scott talked about the GitHub Co-Pilot, a programming assistant for software developers. A developer can describe the functionality needed, and the tool will churn out computer code to accomplish the task.
“Just about anything you can think of where you’re doing cognitive work, you’re going to be able to see things like the co-pilot helping with that work,” he said.
In fact, Linkedin founder and venture capitalist Reid Hoffman appears to be focusing on something very similar with InflectionAI, a stealth startup he recently co-founded with DeepMind co-founder Mustafa Suleyman.
Hoffman was hesitant when asked what exactly InflectionAI is working on, but he did point to “human enhancers” and the concept of AI co-pilots. “Too often, The discourse on AI is about human replacement,” said Hoffman. While there are some tasks that are better delegated to machines, Hoffman says there is a huge opportunity for AI to also act as a human amplifier.
“Let people realize that whatever you do for a living, you’re going to have a co-pilot. Right, so there’s this Microsoft product for engineers, whether you’re a journalist or not, whether you’re an investor, an academic, a lawyer or a doctor, maybe within five years at most there will be a co-pilot two” , he said.
