ChatGPT Gets Internet Access

Greetings fellow Prompterrs, Welcome to our first daily newsletter! Our mission is to save people time and effort by delivering AI news, tips, education and tools that matter most. We want to help you spend less time and effort staying up to date, and more time using what you learn to execute your plan for world domination in the new AI era.

So with that, let’s get this artificial party started!

Today’s Edition:

• 📡 All grown up. ChatGPT gets internet access!

• ⚾️ Google (finally) steps up to the AI plate, comes out swinging

• 🍿 AI Elon Musk and Steve Jobs discuss AI in this uncanny masterpiece

• ☕️ Humans ready to let robots do the driving, but don’t make our coffee!

• 🤓 Tip of the day - 10 ChatGPT power user hacks

📡 All grown up. ChatGPT gets internet access!

OpenAI's ChatGPT web browsing and plugins have launched, providing internet-connected tools that extend the functionality of ChatGPT. Previously, ChatGPT’s data cut off point was 2021. But now, users can enable the web browsing functionality using a new Beta feature that should be rolling out to ChatGPT plus users this week!

Here is what the new interface looks like once it’s available in your settings:

Congratulations ChatGPT, you finally got the internet! Welcome to the world of 50% cats, 50% memes, and 100% distractions. 😵‍💫

⚾️ Google (finally) steps up to the AI plate, comes out swinging

Google announced some major AI news at their developer conference, Google I/O 2023. Some of the new Artificial Intelligence features include:

• Immersive View for Routes in Google Maps
  As if google maps wasn’t already TMI you can now see traffic simulations, bike lanes, complex intersections, parking, and more. (but can it show people who don’t use their blinker and reroute me to avoid?)

• Magic Editor and Magic Compose for photo editing and text rewriting
  Nothing too game changing here, editing photos, removing backgrounds etc.

• PaLM 2, Google's newest large language model
  It will power Google Bard, aka, taking a swing at ChatGPT

• Updates to Bard, Google's AI-powered chat tool
  Removing the waitlist, and adding more language and image capabilities

• MusicLM, an experimental AI tool that turns text into music
  Text to music generation

• New features in Google Search
  Conversational mode and a "Perspectives" filter
  Indicating if images were AI generated

• Sidekick, a new tool for Google Docs
  Provides contextual suggestions

• AI-powered features added to Google Workspace
  Automatic table generation and image creation

• Codey, code completion and generation tool
  Google’s answer to Github’s Copilot

• Project Tailwind, an AI-powered notebook tool
  Note taking and summaries

• Imagen, a new text-to-image model
  Coming soon to Vertex AI, generate and edit images and captions

• Generative AI wallpapers
  Text to image generation for creating wallpapers on Android phones

There’s your recap of the main AI related news from Google. Some major updates here, and some others that feel a little like creating solutions to problems that don’t exist. Either way, with this round of new product and feature updates, Google finally steps up to the plate in the AI revolution. Learn more here

🍿 AI Elon Musk and Steve Jobs discuss AI in this uncanny masterpiece

Not sure how this works, but it’s super uncanny, and I can’t tear my eyes away from it.

Watch AI versions of Elon Musk and Steve Jobs discuss AI.

☕️ Humans ready to let robots do the driving, but don’t make our coffee!

I recently saw this at the shopping mall, a robot barista stand. There was quite the crowd gathered to watch these Jarvis-like robot arms making coffee. They also talked as they whipped up cappuccinos and lattes. If New Yorkers have anything to say about it, when it comes to their coffee, they want the human factor. I have to agree! You robots can drive our cars for us, handle our finances, heck, even do our surgeries, but back away from my mocha!

🤓 Tip of the day - 10 ChatGPT power user hacks

You know what they say, “Promptin’ ain’t easy”. And the value you can get from AI tools like ChatGPT is only as good as your input.

Here are 10 power user tips to improve your results (and impress your boss)

1. Chain Your Instructions:
   You can give a series of instructions in one prompt to guide the model. For example, "Summarize this article in two sentences, then translate it into French, and finally, write a catchy tweet about it."
   

2. Change the 'Temperature' and 'Max Tokens':
   If you're using an interface that allows for it, adjusting these settings can greatly affect the output. The 'temperature' parameter controls the randomness of the AI's responses - a higher value like 0.8 makes the output more random, while a lower value like 0.2 makes it more deterministic and focused. The 'Max Tokens' parameter allows you to control the length of the response.
   

3. Use the 'AI doesn't know' Phrase:
   If you get an answer from the AI that you know is incorrect, or it seems to be speculating, you can remind it and yourself that "the AI doesn't know." This phrase can be used to reset and remind you that the AI is not conscious and is merely processing inputs based on its training.
   

4. Use Imaginary Contexts:
   If you want the model to generate content in a certain style, you can set up an imaginary context in which that style would be appropriate. For example, you could say "Imagine you're writing a letter to a king in the 16th century..." to get the AI to generate text in an old-fashioned, formal style.
   

5. Prompt Reformulation:
   If you're not getting the results you want, try rephrasing or expanding your prompt. Even minor changes in wording can lead to different responses.
   

6. Leverage Multimodal Capabilities:
   If you're using a version of the model that supports it, you can use both text and image inputs together to guide the model's responses.
   

7. Ask for Different Perspectives:
   If you're stuck on a problem, you can ask the AI to generate ideas from different perspectives. For instance, "What would a scientist say about this?" or "How would a historian view this situation?"
   

8. Use Iterative Refinement: 
   If the initial response from the model isn't exactly what you wanted, don't hesitate to refine your request or ask follow-up questions. This approach often leads to better results. You can give feedback, ask the model to think step by step or debate pros and cons before settling on an answer.
   

9. Use System Messages for Orientation: 
   If you're using ChatGPT in a conversation that requires it to play a role or have a certain character, you can start with a system message to orient the AI. For instance, you could start with a system message like "You are an assistant that speaks like Shakespeare." This sets the context and personality for the assistant.
   

10. Instruct Explicitly for the Desired Format: 
    If you're looking for information or asking for a text to be generated in a specific format, specify this in your instruction. For example, if you want a summary of a complex topic, ask explicitly for a "simple," "layman's terms," or "five-year-old friendly" explanation. Likewise, if you want a detailed response, ask for it explicitly.

Use these power user tips to experiment and start seeing the quality of your ChatGPT output increase!

That’s it for the first daily edition of Prompterr! If you found this useful, make sure to subscribe and share with others! Stay tuned for more of the latest news, tips and tools to keep you ahead of the AI game.