🚗 Cars are a privacy nightmare ⋆ 🤖 AI of the week ⋆ 🏛️ Museum of Passion projects

Starting with a shocking revelation of the data your car collects and shares about you...

Continuing with AI updates - how much AI improves the performance of people using it.

Finishing with appreciation for all the great stuff we see around us.

Thing 1 - Cars are a privacy nightmare

Cars are the worst product category we have ever reviewed for privacy.

- Mozilla (yes, the same one that makes Firefox)

Your car might know (and share) more about you than your best friend.

The Data Dive: All 25 car brands they reviewed collect more personal details than you'd expect. Think medical histories, genetic data, sexual information, and more!
What's worse, they might be selling this data or using it for marketing.

At least my data's safe, right? Think again. None of these brands met basic security standards, and 17 faced breaches in the last three years. Plus, they might share your info with authorities without a formal request.

Feeling uneasy? You're not alone. Mozilla is rallying against these invasive practices. You can join by adding your name to their petition.

Thing 2 - AI of the week

I've been telling everyone (who doesn't run away) about GPT-4's emerging features.
Here's the source/inspiration (AI updates follow after):

There is some intelligence in this system… Beware of the trillion-dimensional space. It's something which is very, very hard for us as human beings to grasp. There is a lot that you can do with a trillion parameters…

It could absolutely build an internal representation of the world, and act on it as the processing progresses through the layers and through the sentence temporally…

We shouldn't think about those neural networks as learning a simple concept like "Paris is the capital of France." It's doing much more, like learning operators, it's learning algorithms. It's not just retrieving information, not at all.

It has built internal representations that allows it to reproduce the data that it has seen succinctly… Yes, it was trained just to predict the next word. But what emerged out of this is a lot more than just a statistical pattern-matching object.

- Dr Sébastien Bubeck, Microsoft Research

• StabilityAI released Stable Audio.
  Think StableDiffusion for audio
• Meta released the DINOv2 training code and model weights.
• HeyGen's lip-syncing using the speaker's original voice. See demo.
• Open Robot Parkour - think Boston Dynamics performance, but open-source.
  I wanna get deeper into this...

A minute-and-a-half demo, a mix of terrifying and cute

• Osmo gives computers a sense of smell. Robo-chefs, anyone?
  Think smell-to-text conversion.
• Anthropic and BCG form a new alliance to deliver enterprise AI to clients.
• AI benefits are real to people who use it - research.

Consultants using AI finished 12.2% more tasks on average, completed tasks 25.1% more quickly, and produced 40% higher quality results than those without.

- Ethan Mollick

Distribution of output quality across all the tasks. From the article.

Thing 3 - The world is a museum of passion projects

As you become an adult, you realize that things around you weren't just always there; people made them happen. But only recently have I started to internalize how much tenacity everything requires. That hotel, that park, that railway. The world is a museum of passion projects.

- John Collison, Stripe co-founder

Anyone who's ever tried building anything beyond what the world handed them felt this.

Cheers, Zvonimir