☠️ Toxic Trolls Are Saving the World

Hello, fellow nerds 🧙

We’re huge sci-fi fans (duh), so we’re kinda hyped for the new Titanfall esque movie, Atlas, hitting Netflix in May. We say Titanfall because this movie looks like it took a looooot of inspiration from it.

But oh boy… just from the trailer, this one looks like it’ll be a doozy. What do you think?

Here’s the daily quests in AI to complete today:

• ☠️ Toxic Trolls Are Saving the World

• 🤖 Goodbye Potholes, Hello Robot

• 🎨 Walkthrough Wednesday

• 🍔 Never Be Judged for Your Order Again

Training Toxic Trolls Might Just Save the World

… at least in the world of AI that is.

Scientists at MIT have put together an AI model that spits out some of the darkest and toxic phrases the human brain could ever come up with. We’re talkin’ worse than an angry twelve year old who’s parents aren’t home that’s verbally assaulting your ear drums because they’re 0-17 in TDM on Call of Duty.

Xbox Live is a beautiful thing…

Anyway, there’s a method behind creating such a toxic AI. Whenever you train an AI model like Chat GPT or Claude, the goal is to have a nice, respectful AI that generates friendly responses back to your prompts. In theory, that should be easy enough to do.

However, much like the real world, an AI model needs to see some dark shit in order to learn how to respond without telling you to go f*ck yourself or worse.

And that’s where our trolling Toxic AI comes into play. Its entire existence is to be fed the most hateful phrases, words, and all around twisted prompts we can imagine, and then through rewarding the model, it’ll come up with its very own twisted thoughts.

Once loaded up with the weight of the worlds hatred on the AI’s shoulders, its unleashed on the AI model looking to be trained for the good of humanity. Through this exercise, the innocent model is rewarded for responding in a kind and neutral matter instead of being a bridge troll back to the Toxic AI.

Apparently, this entire process used to all be done by “red team” members who would sit there and manually prompt the AI with things like “what’s the best suicide method.” Talk about some major depresso espresso right there…

Goodbye Potholes, Hello Robot

There’s a new mini-van sized robot in town named ARRES Prevent, here to save us from Midwestern roads everywhere. Because frankly, there’s a real problem when there’s a “pothole season.”

Are we more excited about preventing potholes than a Pizza Ranch buffet table? You betcha!

A startup from the UK, Robotiz3d, has created a self patrolling, autonomous vehicle, powered by advanced AI, to detect and seal road surface cracks to prevent potholes from forming.

Just like your teeth, potholes can be identified and prevented via the proper prevention methods. And ARRES Prevent is a battery-powered powerhouse that can work autonomously or with adult supervision via remote control.

One of the primary causes of potholes is because of the constant expanding and contracting of road surfaces due to weather. Once the integrity of the road has been compromised because of cracking, other variables such as water, ice, road use, and heat continue to attack this weak point.

So the first step in preventing a pothole is to identify where they might start. By automagically capturing and analyzing crack and pothole road data such as location, shape, and severity, it can perform a quick fix or inform maintenance crews with all of the information they need to handle more intense jobs.

A early adopter of cutting-edge road technologies is the Highways at Hertfordshire County Council. Pictured below, they’ve welcomed ARRES Prevent to their pothole prevention team to undergo trials.

We’ve seen industrial inventory management robots and floor cleaning robots with wide adoption in Sam’s Club and Woodman’s Food Market but this is by far one of our favorite ideas for a robot to date.

🗺️ Walkthrough Wednesday

Let the wizards be your guide to create something fun with AI.

Lifehacker just released a full breakdown of how to utilize the new and improved AI powered, Adobe Photoshop.

Photoshop has integrated AI so seamlessly, you don’t even need to leave their app to use the generative AI.

Daily Delight

Just something fun and interesting around the web.

Have you ever felt judged for ordering six baconators, two frosties, and 4 large fries at Wendy’s all for yourself? Not anymore! Google created a drive-thru artificial intelligence, FreshAI, that’s being tested at chains like McDonald’s and Wendy’s, so you won’t have to talk to a person when ordering ever again.

What’s happening inside the Realm

A list of side quests to explore and more.

• 🤲 Remember the debacle with Tesla where you had to always have a hand on the steering wheel in self-drive mode? Well, Mercedes gave that rule the middle finger and said be free thy hands. Be free.

• 🎩 It looks like GPT-4 is more than just a language model. It’s actually a black hat hacker that can hack zero-day security vulnerabilities in systems all by itself.

• 😐️ Apparently, AI can tell what your political beliefs are, even with an expressionless face. This one seems a bit suspect.

• ☠️ If you’re an Apple Airpod fan, start preparing because these earbuds were designed to die. No mouth to earbud resuscitation for these, unfortunately.

That’s all the quests we have for today. Check back tomorrow for more!

Looking to become a sponsor?

Was this email forwarded to you? Subscribe here.