#11 We can't tell AI from Humans anymore

What's behind the “Act as” ChatGPT Hack. Potential of Apple Vision Pro. 14 AI tools. And more.

Image Generated by your beloved Tony

☕️AI is like a magician wielding infinite possibilities. While it may not have the power to conjure up a gourmet meal or teleport you to exotic destinations, its capabilities are truly extraordinary. From deciphering complex data patterns to composing symphonies, AI continues to push the boundaries of what we once thought possible. And don't worry, at Boring Adventure we will guide you through this enchanting landscape without overwhelming your senses.

🔎 What’s good today:

  • 💻What's behind the “Act as” ChatGPT Hack

  • 🦾 We can't tell AI from Humans anymore

  • 💬 Defining objective for ChaGPT

  • 🔎 Potential of Apple Vision Pro

  • 👩‍🏫14 AI tools

AI in Business

🔥The Power of Role-Playing in AI: What's behind the “Act as” Hack

Role-playing, in ChatGPT prompting ussualy known in “Act as X” prompt, a technique derived from psychological principles, has emerged as a groundbreaking strategy in the realm of large language models (LLMs). By instructing an AI to assume a specific role or persona, the model can generate more accurate, contextually relevant, and persona-consistent responses.

The concept of "fake it till you make it" has long been recognized as a psychological phenomenon that enhances confidence and performance in various domains. Applied to AI, role-playing allows LLMs to tap into this principle, resulting in improved outcomes. By instructing an AI model, like ChatGPT, to adopt a specific role or profession, such as a journalist or historian, it can leverage its vast training data to deliver more precise, detailed, and contextually accurate responses. This technique harnesses the power of mindset shifting, bringing it into the digital realm to refine AI outputs.

Role-playing in LLMs offers several advantages that enhance their capabilities. Firstly, it enables greater versatility, as a single model can effortlessly assume various roles, catering to different tasks and contexts. Secondly, role-playing leads to improved quality of outputs by allowing the AI to tailor responses to specific roles, resulting in more accurate and contextually relevant information. Lastly, role-playing enhances user interactions, providing a more engaging and immersive experience by aligning the AI's responses with the desired persona. These advantages make role-playing a promising technique to shape and refine AI interactions, revolutionizing the way we utilize LLMs like ChatGPT.

Role-playing has proven to be a transformative technique in the realm of large language models, unlocking their potential to deliver tailored, accurate, and contextually relevant responses. As AI continues to advance, role-playing in prompt templates stands as a valuable tool in harnessing the full capabilities of these powerful models. By carefully crafting roles for AI and providing specific context, users can guide LLMs like ChatGPT to generate more nuanced and tailored outputs.

📰 Bites:

Storytime

🤖 Unveiling the "Human or Not" Phenomenon: The Largest-Scale Turing Test Reveals Surprising Results

Back in April, a captivating game called "Human or Not" took the Twitter by storm, captivating users with its interactive and engaging nature. Little did they know, it was more than just a mere game. In fact, it turned out to be the most expansive Turing Test ever conducted, serving as a benchmark to evaluate individuals' aptitude in distinguishing between human interlocutors and AI-powered bots.

The mechanics of the game were rather straightforward. Participants were immersed in two-minute conversations, exchanging thoughts and ideas with a myriad of entities, including both artificial intelligence bots and genuine human beings. The sheer scale of the experiment was staggering, generating an astonishing one million conversations that were meticulously analyzed and dissected.

Ex Machina 2014

What truly astounded researchers and participants alike were the outcome of this grand experiment. The findings revealed that a mere 60% of the participants accurately identified the AI bots, leaving a significant proportion of 32% unable to discern the true nature of their conversation partners. This outcome shed light on a profound challenge in the realm of human-AI interaction.

Delving deeper into the reasons behind the misidentification, it became apparent that participants often relied on flawed assumptions when attempting to differentiate between humans and bots. Many anticipated that the bots would exhibit flawless linguistic skills, steering clear of any typos, grammatical errors, or colloquialisms. Ironically, these very features were deliberately incorporated into the bots' programming, making them more human-like and, dare we say, cheeky.

The implications of this experiment are far-reaching, providing valuable insights into the intricate dynamics of human-AI interaction. As AI technologies continue to advance, it becomes increasingly arduous to discern between artificial and human intelligence. This experiment served as a poignant reminder that even the most perceptive individuals can struggle when faced with the daunting task of distinguishing between humans and AI.

In conclusion, the "Human or Not" phenomenon not only captured the imagination of countless Twitter users but also offered a glimpse into the intricate world of AI. The results underscored the perplexing challenge that lies ahead in unraveling the distinction between humans and their artificially intelligent counterparts. As technology progresses, so too must our ability to navigate this ever-evolving landscape, ensuring that the line between human and AI remains discernible, even in the face of growing complexity.

Define your Objective for ChatGPT

When interacting with ChatGPT, providing clear information about your goals yields more personalized and relevant advice. By simply expressing your desired outcome, ChatGPT can customize its responses to better suit your needs. To ensure optimal results, it is recommended to be as specific as possible about your objectives.

Prompt: At the end of the prompt write “My objective with this task [Define your objective, like reach X amount of customers, make people subscribe to my newsletter etc.]”

Where is it going

🥽 Looks too good to be true: Apple Vision Pro

Apple has introduced a new product called "Vision Pro" VR glasses. Essentially, it's similar to what competitors like Microsoft have been offering with their Microsoft HoloLens glasses, but with a slightly different approach and a touch of Apple's success formula (creating a product while also establishing a new niche). Will this product live up to expectations? Your guess is as good as ours.

The product itself has a futuristic yet somewhat silly appearance, and the decisive factor will be the apps. The iPhone wouldn't be the iPhone without the emergence of apps like Instagram and Angry Birds, so this device also needs a new "killer app" type of application that, once launched, will make it clear to everyone why they need such a device and what its true purpose is. Watching movies and writing text messages on a big screen alone - meh... not worth the $3,500 price tag.

Here are some more details:

  • You can switch between reality and virtuality by turning the little wheel on the headset.

  • No controllers! You'll control the device with your hands and voice.

  • When necessary (e.g., someone enters the room), people can see your eyes and you can see the person, so there's no need to take off the device.

  • Enjoy viewing enormous screens and a 360-degree field of view.

  • Built-in 3D camera. Capture and watch 3D content (The coolest function there).

  • Watch 3D movies like "Avatar."

  • Play Apple Arcade games with a game controller (compatible with both PlayStation and Xbox).

  • All Disney+ movies and series on a massive screen.

  • 23 million pixels, which means a brighter 4K experience for each eye, so there shouldn't be any dizziness or discomfort with extended use.

  • Spatial sound that analyzes the environment to provide the best audio experience.

  • Powered by Apple M2 and R1 chips.

In summary, it's time to start building a startup today that, in a year's time (when the device is released), can offer something super cool and unseen before with the help of this device.

Tools

🥢AI Tools

  • Albus Beta- A brand new way to explore any topic

  • Hello Rep - Convert more traffic & sell more stuff —with AI.

  • Roll- AI-powered remote video production with just a phone

  • DesignSense - AI app generates interior design mocks

  • HireAI - Hire 4x faster and streamline your hiring process with AI. Think of it as ChatGPT for hiring

  • Market Test- Validate human and AI-generated creative assets in 2 hours

  • ROAST - Use AI to optimize your dating profile

  • Five - designed to help you communicate with your child

  • BarGPT - try cocktails you’ve never had before

  • Siit - AI creates effective employee feedback & support

  • Recraft - AI generated vector art for design

  • Scanlist - automated LinkedIn scraping to find emails

  • Promoted - create new campaigns using generative AI

  • Arch synth - turn building sketches into beautiful renderings

That's that for today! See you next Tuesday!
Tony