AI round-up: Week of October 28, 2024

Happy Friday! Before we get too far, I wanted to let you know there won’t be an AI Round-up next week. I heard next week might be a little disruptive in general, and we all might be distracted. Something about an election…

We’ll pick back up the week of November 11, which puts us 17 days from Thanksgiving and 19 days from ChatGPT’s two-year anniversary.

Wow. We have come a long, long way since 2022.

the BIG five

1.    AI in the workplace: Answering three big questions.
This article should count for three of today’s big five. There is so much packed into this summary of Gallup’s most recent AI adoption report.

There are some key themes here to be mindful of:

• Companies are behind in terms of putting programs in place that enable employees to learn this technology.
• Employees are using it anyway.
• Employees are concerned their employers don’t have education measures in place or, at the very least, policies/guardrails around usage.

It’s quite a fascinating read that simply and smartly spells out just how important it is for organizations to start this journey before it’s too late.

2.    Gamechanger? Meta has a multiyear deal in place with Reuters.
ChatGPT goes back to October 2023.
Anthropic/Claude to April 2024.

By signing a deal with Reuters, Meta users are able to access answers in real-time. Now, if it could just tell a way to predict the future, as I really want to know who Juan Soto is going to sign with.  (Axios)

3.    Gartner says Agents will transform work in the latest hype cycle update
Great read from Venture Beat. Now, what’s interesting here is the juxtaposition of the trough (of disillusionment) where executives are saying the technology isn’t living up to the hype.

Think back to our first article … is it because they’re just not using it? Because if you fast forward in the article, Gartner gives three examples that show just how fast we’ll be moving with AI:

• By 2025, 30% of enterprises will have implemented an AI-augmented and testing strategy, up from 5% in 2021.
• By 2026, more than 100 million humans will engage with robo or synthetic virtual colleagues, and nearly 80% of prompting will be semi-automated. “Models are going to get increasingly better at parsing context,” said Chandrasekaran.
• By 2027, more than 50% of enterprises will have implemented a responsible AI governance program, and the number of companies using open-source AI will increase tenfold.

Something doesn’t add up for me.I will be very interested to see where things stand in their next report. 

4.    OpenAI says it has no plans to release Orion in 2024. (TechCrunch)

“No lyin. We’re not releasing Orion.”

-Sam Altman (ok, he didn’t say that, but it’s what I imagine he would say if he was a dad)

To borrow from college football icon Lee Corso…

5.    But OpenAI is going to Search ChatGPT.
In fact, it already has. (TechCrunch)

I’m asking/saying this in all seriousness:

Do you think at any point we will be faced with the decision of whether to use Search ChatGPT or Google in terms of environmental impact? Working within ChatGPT means I will be much more likely to do a simple search within that platform. (Rather than go to Google.)

However, as we’ve learned, the impact and output to use AI v. search is significant. Do I need to burn a few bottles of water to find out who the villain was in the first Friday the 13th? I mean, we’ve moved on from Styrofoam containers. Plastic straws. Fast fashion. What will change down the road for AI? Maybe we’ll pay a premium to use it?

Questions for another day, I guess.

For today, we need to take the warnings of AI’s impact on search seriously. (If you weren’t already.) Because how are you going to get people to your site if the person searching doesn’t need to leave the LLM they’re using?

Just a plug for the ‘Did you hear’ section below … make sure you read the first article as it pertains to why this matters so much in the short term.

Learn a little

What’s an RAG database, and why do you need to know about it? Shelly Palmer explains.

Did you hear…

…AI search could break the web. (MIT) Technically, this should’ve been a Big Five story. Wowzers, what a read.

…more than a quarter of new code at Google is generated by AI. (The Verge)

…Google’s detection tool is now available for everyone to try? (Digital Trends)

…Google’s Jarvis could be the biggest Chrome update in 15 years? But why? (Tom’s Guide)

…Miles Brundage, who is basically the guy in charge of AGI readiness at OpenAI, has left. Because no one is ready for it. Ok then! (The Verge)

Must read/must discuss:

Remember when social media debuted? And everyone was an expert simply because they had a channel and a Klout score? (RIP Klout)

If you’re waiting for that moment in AI, it may be here. There are already “super users” out there. Some of them are profiled in this piece from The Washington Post. Check it out if you want to see what constitutes a super user, what they think about this technology, what gets them excited, makes them nervous and how they are benefitting from it.

-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.