AI round-up: Week of August 12, 2024
Breaking news – it’s getting wild out there, guys.
What the heck is going on here:
It’s days like this that I’m glad the round-up goes out later in the week. I was going to start today’s post by writing about how annoyed I’m becoming with the billionaires playing with these powerful AI tools and bragging about it on Twitter (X). If you missed it, a mysterious account popped up called @iruletheworld. The account profile is a pic of Joaquin Phoenix’s character from the movie “Her,” with three strawberry emojis. This is a reference to some inside jokes/references (too deep to go into here) between Elon Musk and Sam Altman.
So, for days, people have been trying to figure out what this is all about. And then, over the past day, we saw the release of Grok-2. So maybe that has something to do with it? But then @iruletheworld tweeted this to start the day:
Then, literally moments ago, everyone started tweeting about ChatGPT being down … and all of those tweets were retweeted by, you guessed it, @iruletheworld, with comments like “here we go” and “they’re starting to realize” and “hello world.” So, is a new version of ChatGPT coming today? Is it 5? Is it AGI? No one knows … but if you want to follow along at home, here are some references:
1. Keep tabs on @iruletheworld
2. Listen to this week’s AI Show, where Paul does a fantastic job breaking it down.
3. Read the backstory on the source of Project Strawberry/OpenAI.
4. Check out the Gemini summary Dillan pulled from the Reuters article:
OpenAI's Strawberry Project: Five Key Points
1. New Reasoning Technology: OpenAI is developing a new AI technology codenamed "Strawberry" aimed at significantly improving AI reasoning capabilities.
2. Deep Research Focus: Strawberry is designed to enable AI models to autonomously navigate the internet and conduct in-depth research, a feat currently beyond AI capabilities.
3. Post-Training Innovation: The project involves a specialized post-training process to enhance AI models after initial training, potentially leading to breakthroughs in reasoning.
4. Long-Horizon Tasks: Strawberry targets complex tasks requiring planning and execution over extended periods, such as scientific research and software engineering.
5. Industry-Wide Challenge: Improving AI reasoning is a critical goal for many tech companies, with OpenAI aiming to be a leader in this field.
The Big 5
Are you still with me? We will get through this together.
Obviously, the news above is our first big thing.
But there are other news items to note.
2. Are we blind to how far we’ve come?
Do you remember that summer between your 8th and 9th-grade year? You left school, hung out with your friends and then came back three months later to teachers and friends saying, ‘Wow, you look different!’ or you thinking, ‘Wow, they look different!’
I remember thinking, ‘Why didn’t my parents tell me I look so different?’ – well, of course, they didn’t know. Because they saw me every day. Kind of like we see AI every day. But a weird thing has happened to AI going into its 9th-grade year … it’s changed. Like, a lot. It has some facial hair. A new haircut. And discovered contacts.
Ethan Mollick does a great job showing us just how much things have changed in almost two years.
3. AI + the ‘A’ players – this is the future.
Whoa. Someone finally wrote it. The article that talks about how high performers will thrive in a world with AI … and C players … won’t.
You’ve heard it for months: “AI won’t take your job. But someone who uses AI will.” Well, this article in Chief Executive has the guts to actually tackle what that might mean.
4. Intelligence Decoupled from Humanity, by Shelly Palmer
Not the read we want but the read we need … if you’re like me, wondering why more people aren’t talking about all….this.
5. MAICON – it’s getting closer!
Learn a little
The ‘learn a little’ section has gone from truly learning a little (hey, try out this prompt!) to learning a lot, especially as we start to think about how we advance AI and focus on learning AND applying.
That’s what today’s article is all about, from HR Dive:
4 ways recruiting leaders are investing in AI
Which is good, because 50% of job hunters are using AI to find those jobs.
Did you hear about…
…Flux? It’s the new leader in the clubhouse for AI image generation. (Tom’s Guide)
…Only two of the 11 original OpenAI leaders are left? Here’s where they’ve gone. (Financial Times)
…Elon Musk agreeing to stop training Grok on European X users’ posts? For now, anyway. (Politico)
…Microsoft’s Copilot AI vulnerable to hackers? (Wired)
…Trump is claiming Harris is using AI to make her rally crowds look bigger? (Rolling Stone)
NOTE: This is not a political opinion of any kind. It’s a reminder to fact-check anything any candidate puts out there!
…92% of IT jobs will be transformed by AI? (CIO)
…Google’s Gemini Voice chat? (The Verge)
Must read/must discuss:
Paul Roetzer is making his second appearance in this round-up … and is in this slot for the second straight week.
Totally being transparent here … I thought long and hard about including this. Not because I don’t think you should see it … but because of the weight of such a topic. At his new company, SmarterX, Paul has tackled a topic I feel is critical to any organization’s AI path forward: the impact AI has on a job.
As Paul will tell you on this week’s podcast, this is his interpretation of how AI impact should be/could be measured. The tool, which is still in testing, will take your job description and break down the exposure to AI and the tasks you can expect AI to handle.
You can experiment with it yourself here. You do need ChatGPT as the tool is a GPT. And remember, as we mentioned at the start, ChatGPT is currently down…
Let me know if you try it. I did. I’m still processing what I think. But it’s kind of like looking into the future. (If you made it to season 3 of “Westworld,” seeing the future didn’t go so well for everyone.)
Thanks for reading!
-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.