Hello Innovators 🚀, Tech Enthusiasts 🤖, and Curious Minds 🤓! Welcome to another edition of the Watchtower 📰, the newsletter that keeps you at the forefront of all things tech, AI, and innovation 💡.

As the pace of technology speeds up 🏃‍♂️💨, it's crucial to stay informed, ask questions, and engage in meaningful dialogue. This week, we have a selection of handpicked stories 📚 that explore the fascinating and, sometimes, puzzling intersections of technology with our daily lives 🤷‍♀️🌐. From the ethical debates surrounding AI's impact on elections 🗳️ to the magical yet perplexing world of biometric payments 💳🔮, we've got you covered.

So, grab your favourite cup of coffee ☕ (or tea 🍵) and let's dive into this week's riveting topics 🏊‍♀️.

For the best possible viewing experience, we recommend viewing this edition online.

🗓 Upcoming Events

🗳️ How Concerned Should You Be About AI Impacting Elections?

In the latest tech news, we're diving into the impact of AI on elections. 🗳️ Should we be worried about AI-driven disinformation campaigns? 🤔

👥 People have always spread propaganda and misinformation during elections. But now, generative AI tools like ChatGPT are changing the game. They can create tons of fake content, making disinformation campaigns bigger, more convincing, and harder to detect. 🤯

📈 AI could increase the quantity and quality of disinformation, use micro targeting to flood us with personalised propaganda, and even mess with our trust in fellow citizens. 🙈

🛡️ But here's the good news! Voters are harder to persuade than we think, especially on important issues. AI might not be the doomsday disruptor we fear. 🙌

🔍 Social media platforms and AI companies are on the lookout for political-influence operations. They're improving their detection methods and even considering disclosure requirements. 🕵️‍♂️

🤷‍♀️ So, should we panic? Not quite! While AI is powerful, it's not the root of all problems. Disinformation has been around long before AI came into play.

Stay informed, stay vigilant, and remember, the power to discern fact from fiction lies within us! 💪

_{Published by Li Li, September 15 2023}

Wave Hello to the Future: Amazon One's Palm Magic & The Privacy Puzzle

🎉 Hey there, tech enthusiasts and future moguls! 🎉

We've got a piping-hot, futuristic tale for you today. Imagine walking into Whole Foods, filling your cart with some organic, gluten-free, fair-trade kale—because, let's be real, you're that cultured—and then paying for it by just waving your hand. Yep, you read it right! Amazon One is turning your palm into a magical wand for payments and identification. 🌟

The Tech Behind the Magic

Using infrared light, this bad boy scans the unique lines, grooves, and veins in your palm. Think of it as your palm's astrological chart, but for Amazon. The technology achieves an eye-watering 99.9999% accuracy rate. To put that in perspective, it's 100 times more accurate than scanning two irises! Seriously, even the Sorting Hat from Harry Potter would tip its... er, brim to this level of precision. 🎩

Generative AI to the Rescue

Amazon faced the chicken-and-egg problem: they needed a ton of palm data to make this work, but how do you get that without being creepy? Enter Generative AI. It created millions of synthetic palm images, training the model in various 'palm-scenarios'—from romantic candlelight to glaring fluorescents.

Beyond Payments

Now, Amazon One isn't just a one-trick pony. It's more like a Swiss Army knife with a sense of humor. Beyond payments, it's linking loyalty programs and even verifying your age. So yeah, you can finally prove you're old enough for that kombucha! 🍹

But, What About Privacy?

Ah, the elephant in the room! 🐘 Now, Amazon says they've designed this with privacy in mind. The system operates beyond the visible light spectrum, so it's not snooping on your skin tone or gender. But let's press pause. How comfortable are you with your biometric data—basically your body's social security number—being stored somewhere?

Questions to Ponder 🤔

- Could this replace traditional IDs or even passports?

- What are the ethical implications of biometric data collection?

The Rundown

So, what fields could this disrupt? Pretty much anything requiring verification or transactions. Businesses could speed up customer service; Science could see breakthroughs in identification accuracy; Engineering could integrate this into smart homes; Arts could use this for ticketing at events.

In a world of increasing surveillance, how do we draw the line between convenience and privacy? Your thoughts, people, your thoughts! Because if anyone can solve this conundrum, it's the brilliant minds reading this. 😉

Until next time, keep waving those magical palms! 🌴✨

_{Published by Andrew Suryanto, September 15 2023}

🔍 Did You Know ChatGPT Could Do This?

Hey scholars, researchers, and lifelong learners!

Ever found yourself stuck in a labyrinth of academic papers, trying to find that one golden nugget of information to solidify your argument? 📚🤔 Well, get ready to have your mind blown! 🤯

Your Argument-Boosting Companion 🤝

If you've got the premium version of ChatGPT 🌟, you can now tell it to scour the academic depths for papers that align with your argument. That's right! You're no longer alone on your quest for evidence. Just a few keystrokes, and voila, you've got yourself an arsenal of scholarly papers 📖 that fit whatever general argument you're trying to justify.

The Power at Your Fingertips 🤳

Think of it as your personal research assistant 🤖, tirelessly working behind the scenes to fetch you what you need. It's like having access to an academic library right at the tip of your fingers 📚✨. All you have to do is use one of the many paper-searching plugins available in your request, and just like that, you'll get a curated list of papers and sources to beef up your argument.

A Game-Changer for Assignments 📝

This feature is a lifesaver for those assignments where you need to present a well-researched argument. Whether you're a student drowning in essays 📝 or a professional crafting a compelling report 📊, this tool is invaluable. It not only saves you time ⏳ but also ensures that your claims stand on a solid foundation of evidence 🏛️.

Questions to Ponder 🤔

How could this feature change the landscape of academic research? 🎓Does having such easy access to information make us more informed or just more overwhelmed? 📚

The Rundown 📊

So, there it is—a tool that could revolutionize the way we approach research and argumentation. The question that remains is how we use this power responsibly and ethically 🤔⚖️. But one thing's for sure, the future of academic research just got a whole lot more convenient 🚀.

_{Published by Solomon, September 15 2023}

👁 Google DeepMind Finds AI-Designed Optimal Prompts for Language Models

In a remarkable development, the Google DeepMind team has unveiled an unfathomable technique for enhancing the performance of large language models through the use of carefully crafted prompts. These prompts, when combined with a new "magic spell" – "Take a deep breath," have demonstrated remarkable results in improving the scores of AI models. 😮

By combining the familiar "Let's think step by step" with the new "Take a deep breath" prompt 🤝, large language models achieved a significant boost in their performance on the GSM8K dataset, which tests the AI on grade school maths, with scores soaring from 71.8 to an impressive 80.2 points 📈. What makes this discovery even more astonishing is that the AI models themselves devised this highly effective prompt.

The AI community has shared their various viewpoints with discussions about this discovery 🗣. Some suggest that taking a deep breath might even rev up the cooling fans in AI hardware, while others jokingly advise newly hired prompt engineers to take a deep breath, as their demanding roles might lead to short-lived careers. 🙊

This innovation is presented in the paper titled “Large Language Models as Optimizers," which has garnered widespread attention 🦾. Beyond prompt design, the paper explores the ability of large language models to excel in classic optimisation tasks, such as linear regression and the travelling salesman problem. 🧑‍💼

Interestingly, the optimal prompts also vary depending on the specific language model being used. For instance, the best prompt generated by AI for GPT-3 was "Let's work this out in a step by step way to be sure we have the right answer," surpassing the human-crafted prompt "Let's think step by step." 🙃 However, on Google's PaLM 2 and Bard models, human-crafted prompts outperformed those generated by AI using the APE method. 😁

_{Published by Solomon, September 15 2023}

Closing Notes

And there you have it, folks 🙌!

Another week, another set of mind-bending technological advancements to ponder 🤔. As we wave our magical palms ✋✨ at Amazon One kiosks or debate the ethical implications of AI in electoral politics 🤖🗳️, let's remember that technology is a tool—a powerful one, but still a tool 🔨. It's up to us to wield it responsibly and thoughtfully ⚖️, considering both the opportunities 🌈 and the risks 🌪️.

We hope you've enjoyed this edition of the Watchtower as much as we've enjoyed curating it for you 💖. Remember, the best way to predict the future is to invent it, and knowledge is the starting point 📚. So, stay informed 🧠, stay vigilant 👀, and never stop questioning ❓.

As always, we welcome any and all feedback/suggestions for future topics here or email us at u[email protected] 

Stay curious,

🥫Sauces 🥫

Here, you can find all sources used in constructing this edition of WatchTower:

• How generative AI helped train Amazon One to recognize your palm

• Large Language Models as Optimizers