AI round-up: Week of March 31, 2025
We’re going heavy on Ethan Mollick to start this week, and for good reason – he’s an absolute must follow anyway, but he has dropped some truly mind-bending intel in the past few days.
Let’s get to it.
The Heavy Stuff
Story 1: Your AI coworker will save you 12-16% of time. (One Useful Thing)
Ethan Mollick worked with a team to conduct a study of 776 professionals at Procter & Gamble to determine just how much AI impacts our work.
The results show one big thing: there is an impact. To the tune of double digits.
Still think the future of work won’t include AI?
Story 2: AI interviewed itself. (Ethan Mollick, LinkedIn)
Remember that study we just talked about? Ethan Mollick uploaded it to HeyGen, which then proceeded to generate an AI host interviewing an AI version of Mollick about said paper.
AI Ethan being interviewed by AI host.
We live in incredible times.
Story 3: One More Mollick: AGI is coming. Hello, is this thing on? (Ethan Mollick, LinkedIn)
I just finished rewatching Avengers Infinity War with my youngest daughter. The opening sequence, concluding with Bruce trying to warn people that Thanos is coming reminded me of where we stand with AGI. Which is essentially what Ethan Mollick is saying in this post regarding the fact that Google Deep Mind put out a course on AGI Safety and it has less than 1000 views (as of a day ago).
Story 4: Google Deep Mind put out a course on AGI safety. (Google DeepMind, YouTube)
You saw that coming, right?
Story 5: Meta is stealing books to train its AI. (Lit Hub)
Looking to rob us of more than our soul, the Facebook parent company is literally pirating books from LibGen to train its AI. Against the advice of lawyers.
One author you might know ‘caught’ Meta doing this. Here’s Ann Handley’s account, who broke the news on LinkedIn that Meta used all three of her books to train without ‘consent, compensation, copyright concern, credit.’
Oh, and no class.
Story 6: Stuff I don’t understand, but it’s bad for web traffic.
Speaking of training, AI crawlers are absolutely crippling websites, sometimes forcing blocks on ‘entire countries’. (Ars Technica) It might cost us Wikipedia in the near future as the traffic to the site has spiked by 50% since January 2024 thanks to more AI agents being trained. (TechCrunch)
We’ve talked about the large percentage of web traffic that is AI (crawling/scraping) and what we can do about it…not much. But it’s worth knowing for two reasons:
- Stay in constant contact with your IT team. This is not a siloed approach – communications/marketing and IT have to be working together, sharing knowledge, etc.
- The impact on search. So, more bots are crawling…which means we need to make sure they are finding your brand or else…we get left behind.
Story 7: Generative Engine Optimization (GEO) – what it is and why it matters (NoGood)
Or, what I like to refer to as ‘How to not let your brand get left behind’ when it comes to search in the age of AI.
Breaking news - Story 8: Thanks to this going out late, you can be one of the first to read about the big announcements from Microsoft regarding the latest Copilot news! (Courtesy of Conor Grennan)
The not-so-heavy stuff
Story 1: Don’t use AI to make food images. (PetaPixel)
It’s making people uncomfortable!!
Story 2: Amazon’s new AI agent will shop third-party sites for you. (TechCrunch)
Question is: Will an AI agent think AI food looks appealing?
Story 3: Runway claims to have finally achieved consistency in AI videos. (Ars Technica)
But can it make videos of food?
Story 4: Apple and Mulberry Jam (PYMNTS)
Apple announced Project Mulberry which is an initiative to launch an AI doctor on its next OS update as early as the summer of 2025.
Also, I couldn’t resist the urge to ask ChatGPT for an image of apples and mulberries.
Story 5: ChatGPT users have generated 700M images since last week. (TechCrunch)
Including that fruit bowl.
Story 6: So it seems like a good time for Midjourney to release a new image gen tool, eh? (TechCrunch)
Its first in over a year!
Story 7, AKA Shameless plug: Mike Lawrence wrote a great blog. (Akhia Insights)
Given all of these image gen tools…it seemed like a good time to ask a writer to talk about image gen. (You will enjoy this, trust me.)
A few that don’t fit in either category
Story 1: CEO of software company Replit says ‘don’t bother learning to code’. (Financial Express)
Instead, he recommends doing this…wait for it…learning to COMMUNICATE, like humans.
"Learn how to communicate clearly, as you would with humans, but also with machines."
Sounds like AI literacy to me.
Story 2: Speaking of AI literacy…OpenAI and Anthropic are fighting over your kids. (The Verge)
Which will become the default AI for this and future generations of school kids?
Listen, I just watched Adolescence and the last thing these kids need is more time on screens.
Oh, and then there's this ⬇️
Story 3: Teachers warn AI is impacting students’ critical thinking. (Axios)
Uh, it’s not just students…but yes, let’s start somewhere.
Story 4: How to write your first AI agent. (Even if you don’t know what that means yet.) (Shelly Palmer)
Shelly Palmer digs into what we should know about building our first AI agent. This might be a save and come back to article but definitely wanted to share.
I hope to have more to share on this, personally, in the coming weeks.
Final note: Bill Gates does it again
So when Bill Gates drops in on a late-night show, to discuss technology, we should listen. Here he is talking to David Letterman in 1995 about the Internet. It’s eerie.
Here is talking to Jimmy Fallon in 2025 about AI, which he claims will replace doctors, teachers and more, making us unnecessary for ‘most things’.
Happy Friday!
-Ben
As a reminder, this is a round-up of the biggest stories, often hitting multiple newsletters I receive/review. The sources are many … which I’m happy to read on your behalf. Let me know if there’s one you’d like me to track or have questions about a topic you’re not seeing here.