What you see here is the last week's worth of links and quips I have shared on LinkedIn, from Monday through Sunday.

For now I'll post the notes as they appeared on LinkedIn, including hashtags and sentence fragments. Over time I might expand on these thoughts as they land here on my blog.

2023/12/05: The easy part is over

(Image credit: Idiocracy)

Looking at the exploits being waged against LLM-based chatbots and … I'm reminded of something I've said before: these systems are a terrible mix of "naive" and "powerful."

The latest example is the research group that convinced ChatGPT to share sensitive details from its training data:

"ChatGPT Can Reveal Personal Information From Real People, Google Researchers Show" (Vice)

(For the moment we'll ignore that these are researchers from Google, which has its own LLM products …)

Slowly but surely, we're learning that "build the robot" was the easy part.

Creating safeguards around it? That'll be the long slog.

2023/12/06: A distributed cyborg operation

I've written a few notes here about mixing people and AI in order to accomplish a task. This WSJ video isn't about AI, per se, but the lesson still holds:

"Why Target’s 2,000 Stores Are Its Biggest Asset Against Amazon and Walmart | WSJ Shipping Wars" (The Wall Street Journal)

Part of Target's logistics chain includes "sortation centers" that act as routing hubs: package comes in, package gets sorted into the right spot for its exit, package goes out.

Machines and conveyor belts move packages around. People are also present to fine-tune the operation. That means each label must be readable by both people and machines.

For the machines, there are the barcodes we all know and love. For the people, Target has included some extra details:

• "TLMD" indicator, for "Target Last-Mile Delivery"
• A large number (here, "1") to note partition number
• Route number and cart, which ultimately lead to an outbound delivery vehicle

All of this is printed on a single label to help people and machines handle the same objects in the same location.

(Several comments on the video point out that Target is far behind Amazon as far as its use of AI. That may be the case, but that's a separate discussion. What I'm highlighting here is the importance of helping people and machines work together to tackle a problem. Now that Target has sorted this out, they can identify the best parts to automate.)

What lessons can you take from this? What steps can you apply to your business, to help get the best of AI and humans working together?

2023/12/08: Asking the hard questions

For the data leaders, stakeholders, and product owners out there: how well do you know your company's privacy policy? And what do you know about the data you collect and use in your products?

Several auto manufacturers are thinking about that this week, as US Senator Edward Markey has called them out on their data handling practices. Like smartphones, modern cars have become devices that hold a lot of data about us … and it's not always clear who gets access to that data.

"Automakers’ data privacy practices “are unacceptable,” says US senator" (Ars Technica)

This is when you might say: "This is interesting and all, but my company isn't in the connected-vehicle space. Why should I care?"

That's a fair question! You want to care for two main reasons:

1/ You're not in the connected-vehicle space today … but perhaps next year, your company will provide services to those auto manufacturers and/or try to plug into their platforms. You'll want to know about data privacy and regulation matters in that field before you make too many plans.

2/ Today, Senator Markey is looking into auto companies. What field is next? Is it yours? Are you ready to respond to hard questions about your data practices from lawmakers and other officials?