Week 24 — Wild and surprising AI news
Week 24 was the week when Apple truly entered the AI race, but it was far from the only wild thing that happened during the week.
Apple: AI for the rest of us
The big event this week was Apple’s Keynote at their World Wide Developer Conference on Monday. The initial reaction from those (us) deeply involved in AI development was disappointment, but after some reflection, the mood has changed.
Most people can see the logic in Apple going their own way and focusing on consumers and on such deep integration into everyday tasks that we won’t think of it as Artificial Intelligence (but as Apple Intelligence).
The market has also responded well to the news, sending Apple’s stock so high that Apple was again the world’s most valuable company this week.
There has been a lot written about it. Here is a link to the sharpest analysis I’ve read, which I think is thought-provoking: https://www.sectionschool.com/blog/apple-intelligence-and-consumer-ai
Here you can read an overview of all the new features that Apple Intelligence is expected to deliver: https://www.apple.com/apple-intelligence/
AI Steve runs for the UK parliamentary election
There is an election in England. And this time, there is an AI on the ballot in Brighton.
It’s Steve Endacott — an enterprising entrepreneur from Sussex who came up with the idea.
The idea is that the chatbot AI Steve will engage in dialogue with all the citizens in the region to develop a policy that aligns with what the civilians want, and then human-Steve will show up physically to represent AI Steve’s policy.
It may sound completely crazy, but in the first 3 days, AI Steve had 2,500 dialogues with citizens — a feat no human candidate could have managed, and now an AI algorithm is condensing it into policy proposals.
AI Steve was created by the company Neural Voice, which specializes in conversational AI and where Endacott happens to be the chairman. So, there are several agendas at play here.
But it’s exciting innovation in terms of democracy development. And there is a basis for an interesting dialogue on whether this is a good or bad development.
You can read more here: https://shorturl.at/kPVG2
Wired also has an interesting article on the case: https://www.wired.com/story/ai-candidate-running-for-parliament-uk/
And check out AI Steve’s website for more specific details here: https://www.ai-steve.co.uk/
Lumalabs overtakes Sora and Kling and launches accessible text-to-video for everyone
This week I finally managed to make my first real AI video. And I could do it because Lumalabs this week not only introduced but actually gave access to their new text-to-video algorithm: Dreams Machine.
Together with my daughter, we agreed we wanted to see: “A pink Elephant wearing a tutu singing Christmas carols on the street” and sure enough, we got 5 seconds of convincing animation out of it.
It took all night, but it was partly because there is extreme pressure on their servers after the launch because we all want to try it.
The last few weeks have been wild on the AI video front with Kling from China last week and now this. Exciting to see how long OpenAI can wait before releasing Sora.
Lumalabs is also known for using AI to give you the ability to create complete 3D models directly from your mobile phone camera — another wild technology you can see here: https://lumalabs.ai/interactive-scenes
But do check out Dreammachine and try to make your own first AI video: https://lumalabs.ai/dream-machine
AI reveals: Elephants call each other by name
Speaking of elephants, a wild story came out this week: By collecting 469 elephant calls — and responses over a period of 36 years, researchers with AI have been able to document that elephants call each other with “name sounds” and respond when called but ignore calls not directed at them.
We have long known that elephants are incredibly intelligent animals that live in families like us and have a rich social life. But this revelation could easily lead to a lot of new knowledge. Beautiful story.
You can read an abstract of the original Scientific Paper here: https://shorturl.at/fvCJi
The Guardian also has a story about the discovery: https://www.theguardian.com/science/article/2024/jun/10/elephant-names-study-ai
Sal Khan revolutionizes the education system
One of the areas most under pressure from Gen AI is our entire education system. What should we actually learn in the future?
How should we be taught? What kind of teaching materials should we develop? What is the teacher’s role in the future? How do we examine in a time characterized by AI?
Just some of the very interesting and complicated questions we must ask.
Back in January, DTU realized that this development could not be stopped, so they opened up for the full use of Gen AI across teaching preparation, assignments, and eventually examination. Read the story here: https://via.ritzau.dk/pressemeddelelse/13776043/dtu-abner-for-brug-af-kunstig-intelligens-i-undervisning?publisherId=13561337&lang=da
One of those who have thought the most about this transformation and has long been working to rethink the school system is Sal Khan. He is the founder of Khan Academy, which operates based on a “flip the classroom” concept. His instructional videos have been viewed more than 2 billion times!
Now he has launched an AI tutor: Khanmigo, which is publicly available. And he has written a new book about how we must expect the entire school system to be rethought but that it is actually the best news in the world. I will definitely read his book during the summer vacation!
If you also want to add it to your reading list, you can find the book here: https://www.penguinrandomhouse.com/books/740806/brave-new-words-by-salman-khan/
His biggest investor has been The Gates Foundation, which over the years has supported Khan Academy and ensured it has been free worldwide.
You can hear him talk to Bill Gates about the revolution in the education system here: https://www.gatesnotes.com/Brave-New-Words
He also has a really interesting Video Podcast where he talks to all the sharpest AI experts: https://www.youtube.com/watch?v=z0S1ORbbQUU&list=PLSQl0a2vh4HBnzzKJ5JQuWoLV87nCeHs4
Some extra reading:
One of the experts we closely follow is Wharton professor Ethan Mollick. He recently released his mid-year report on how we are currently best using Gen AI.
It is worth diving into, and you can find it here: https://open.substack.com/pub/oneusefulthing/p/doing-stuff-with-ai-opinionated-midyear?r=58aey&utm_medium=ios
