weekend ai reads for 2025-03-14

AI search engines send 96% less referral traffic to news sites and blogs than traditional Google search, per a new report by content licensing platform TollBit, shared exclusively with Forbes. Meanwhile, AI developers’ scraping of websites has more than doubled in recent months, the report found.

Similarly to how self-driving car companies rely on remote human operators to provide guidance in exceptional situations, perhaps the knowledge worker of tomorrow will be dropped in to nudge an AI agent in the right direction.

The resulting ten roles were the highest-paying, lowest-AI-risk jobs with much faster than average growth — and they were all in health care.

BREAKING🚨 So, I tested this new LLM-based system. It generated this 200-page report I didn't read and then this 150-page book I didn't read either, and then a 20-page travel plan I didn't verify.

All I can say: it's very, very impressive! 🔥🚀

First, the number of pages it generated is impressive 👀

⚽ But not just the number of pages: The formatting is so nice! I have never seen such nicely formatted 200 pages in my life.✨⚡

⚠️🌐 A game changer! ⚠️🌐

The company has taken down more than 75,000 AI-generated copies of songs by its artists, with popular performers like Harry Styles, Queen, and Beyoncé being the primary targets, according to a source familiar with Sony's efforts.

That is where artificial intelligence comes in and, ideally, does the calculations to lower energy consumption by, for example, providing information on when to charge the car or when to sell the excess energy the solar panels produced on the roof.

First, be aware of what you get from your AI assistant.

Learning from text data means systems such as ChatGPT are language models, not knowledge models. While it is truly amazing how much knowledge gets encoded via the mathematical training process, these models are not always reliable when asked knowledge questions.

New research shows that ChatGPT, Claude, and other AI systems leave distinctive “fingerprints” in their writing.

Here’s how you can use this knowledge to identify AI content and improve your AI-assisted output.

94% of Americans have heard of AI

Al, it's bigger than Doritos, TikTok, and Viagra

The critical first element to powering an effective AI strategy is a unified customer data foundation. The tricky part is that accurately unifying customer data is hard due to its scale and complexity — most consumers have at least two email addresses, have moved over eleven times in their lifetimes and use an average of five channels (or if they are millennials or Gen Z, it’s actually twelve channels).

Successful AI navigation requires a change in the relationship between strategy and execution and between the C-suite and the rest of the organization. A firm’s ability to adapt along these lines will make or break its AI transformation.

Similarly, at another university, a social sciences professor’s digital twin answered an astonishing 2,000 student questions per week, leading to an average grade increase from a C to a B.

The reality is sobering: It’s very easy for students to use AI to do the lion’s share of the thinking while still submitting work that looks like our own. We can manually edit AI responses to be more “bursty,” employ one of the myriad programs that “humanize” text or blend AI-generated ideas with our own prose.

The problem isn’t technological — even perfect detection software couldn’t prevent intellectual plagiarism when students can harvest ideas from AI and put them in their own words. Ultimately, the problem is behavioral.

It turns out that if there is anything more implausible than the idea that they might need to write as part of their jobs, it is the idea that they might have to write, or want to write, in some part of their lives other than their jobs. Or, more generally, the idea that education might be valuable not because it gets you a bigger paycheque but because, in a fundamental way, it gives you access to a more rewarding life.

Interfacing with rarely used, complex interfaces is a perfect LLM use case.

In this report, we provide a structured taxonomy of these risks by identifying three key failure modes (miscoordination, conflict, and collusion) based on agents’ incentives, as well as seven key risk factors (information asymmetries, network effects, selection pressures, destabilising dynamics, commitment problems, emergent agency, and multi-agent security) that can underpin them.

Use Gemini 2.0 Flash to tell a story and it will illustrate it with pictures, keeping the characters and settings consistent throughout. Give it feedback and the model will retell the story or change the style of its drawings.

We believe interviews, and in particular technical interviews that are solely focused on algorithmic tricks or memorization are ruining the job market by missing out on good candidates like yourself!