filmov
tv
Two AIs time travel to envision the future of news
Показать описание
On stage in Sydney’s stunning Hunter Valley wine region recently, an audience of global travel writers asked me if and when AI would replace humans as creators.
Rather than thrust a finger in the air and offer one of those blah blah answers that placate audiences, I decided to do the research.
In the car on the way home, I used the ChatGPT-enabled Bing Copilot search app on my iPhone to go back 2.3 million years.
My goal was to research how human communication evolved, and what could be learned from it.
If you haven’t used the Bing app this way yet, it enables a two-way chat with GPT to deep dive into topics.
This is a new and better way to search, which is more satisfying than Google links, and conversational and human-like.
As we chatted, I discovered that humanity took an entirely different path to develop language than the methods being used by geeks to build AI today.
Early humans used gestures to communicate first, then took millions of years to evolve to images, then language, and finally to words, just 3,000 years ago.
The internet has taken an opposite evolutionary path, starting with words, then images, audio, and video, and only to conversation in the past year and a bit.
It shows the new AI web still has a long way to go to achieve the ultimate goal of human-level communication, which is what this post is about.
My conclusion is that if human communication is the goal, then the web is a massive failure for billions, and has a long way to go.
Which means it’s a giant opportunity for communication professionals and innovative publishers willing to try to solve it.
This is one of my favourite posts so far…
But, before we dive in, let’s welcome a flurry of new arrivals to our fast-growing almost 4,000-strong community.
His research shows that 93 per cent of human communication is not voice at all but relies on a highly evolved process of gestures and expressions.
That means AI for all its trillions in valuations and techno-whiz-bang wizardry, is still 93 per cent off replacing us.
That’s the answer I wish I had for the members of the Australian Society of Travel Writers who I was keynoting. 🤦♂️
But AI is developing exponentially faster than we humans did. It took us 2.5 million years to get from a gesture to a grunt to a GPU.
Now the world is in an arms race to get to whatever AI enables the web to be in the future.
And my belief is that the ultimate winner will be whoever can scale digital information delivery into the most human-like experience first.
Because, if all of human history is a signal, then it has to be inevitable that words and pictures will not be the communication format of the future.
And that begs the question why 99.9 per cent of news media resources are focused on something that’s dying. Or maybe that’s why 99.9 per cent of publishers are dying?
But it supports my 20-year belief that the future of news telling must evolve to be audio and visual, and as human-like as tech can be, to remain relevant.
To test this, I decided to conduct an experiment.
On the three-hour drive from my speaking engagement back home to Sydney, I had a long conversation with Microsoft’s Copilot.
I wanted Copilot’s AI to teach me about the history of human language, and then to merge my knowledge with the archives of the Wall Street Journal to join some dots.
Because I believe that the way we deliver news has to change.
And I decided to do the entire thing from my car, using only voice, and while driving, to fulfil another hypothesis I’ve held for years: That news media fails to engage its audiences while driving.
* 1.4 billion people commute in a car alone every day.
* The average commute is slightly more than 30 minutes each way.
* That’s 5,276 years of addressable media time in just the top 10 cities, and
* 121,348 years globally
* Every day.
With one ad every eight minutes, and a podcast costing near nothing to make, that’s 7.9 billion $50 CPM ads in addressable revenue. Just sayin’
So, here’s what I did.
Rather than thrust a finger in the air and offer one of those blah blah answers that placate audiences, I decided to do the research.
In the car on the way home, I used the ChatGPT-enabled Bing Copilot search app on my iPhone to go back 2.3 million years.
My goal was to research how human communication evolved, and what could be learned from it.
If you haven’t used the Bing app this way yet, it enables a two-way chat with GPT to deep dive into topics.
This is a new and better way to search, which is more satisfying than Google links, and conversational and human-like.
As we chatted, I discovered that humanity took an entirely different path to develop language than the methods being used by geeks to build AI today.
Early humans used gestures to communicate first, then took millions of years to evolve to images, then language, and finally to words, just 3,000 years ago.
The internet has taken an opposite evolutionary path, starting with words, then images, audio, and video, and only to conversation in the past year and a bit.
It shows the new AI web still has a long way to go to achieve the ultimate goal of human-level communication, which is what this post is about.
My conclusion is that if human communication is the goal, then the web is a massive failure for billions, and has a long way to go.
Which means it’s a giant opportunity for communication professionals and innovative publishers willing to try to solve it.
This is one of my favourite posts so far…
But, before we dive in, let’s welcome a flurry of new arrivals to our fast-growing almost 4,000-strong community.
His research shows that 93 per cent of human communication is not voice at all but relies on a highly evolved process of gestures and expressions.
That means AI for all its trillions in valuations and techno-whiz-bang wizardry, is still 93 per cent off replacing us.
That’s the answer I wish I had for the members of the Australian Society of Travel Writers who I was keynoting. 🤦♂️
But AI is developing exponentially faster than we humans did. It took us 2.5 million years to get from a gesture to a grunt to a GPU.
Now the world is in an arms race to get to whatever AI enables the web to be in the future.
And my belief is that the ultimate winner will be whoever can scale digital information delivery into the most human-like experience first.
Because, if all of human history is a signal, then it has to be inevitable that words and pictures will not be the communication format of the future.
And that begs the question why 99.9 per cent of news media resources are focused on something that’s dying. Or maybe that’s why 99.9 per cent of publishers are dying?
But it supports my 20-year belief that the future of news telling must evolve to be audio and visual, and as human-like as tech can be, to remain relevant.
To test this, I decided to conduct an experiment.
On the three-hour drive from my speaking engagement back home to Sydney, I had a long conversation with Microsoft’s Copilot.
I wanted Copilot’s AI to teach me about the history of human language, and then to merge my knowledge with the archives of the Wall Street Journal to join some dots.
Because I believe that the way we deliver news has to change.
And I decided to do the entire thing from my car, using only voice, and while driving, to fulfil another hypothesis I’ve held for years: That news media fails to engage its audiences while driving.
* 1.4 billion people commute in a car alone every day.
* The average commute is slightly more than 30 minutes each way.
* That’s 5,276 years of addressable media time in just the top 10 cities, and
* 121,348 years globally
* Every day.
With one ad every eight minutes, and a podcast costing near nothing to make, that’s 7.9 billion $50 CPM ads in addressable revenue. Just sayin’
So, here’s what I did.
0:01:24
0:05:58
0:08:54
0:00:27
0:02:18
0:16:43
0:06:05
0:25:08
0:19:29
0:00:57
0:01:00
0:14:33
0:11:15
0:02:33
0:02:37
0:04:17
0:22:18
0:43:02
0:03:18
0:02:21
0:08:05
0:02:48
0:02:45
0:02:42