Notes from Day1 of the Mindvalley AI Summit 2023

AI summary for Day 1

The Mindvalley AI Summit 2023 Day 1 focused on the potential and implications of Artificial Intelligence (AI). Key points include:

1. AI can be used for good or bad, depending on our consciousness. It can extend human capabilities and has the potential to change a billion lives.
2. AI technologies like ChatGPT are being adopted rapidly. Top users of AI can stay ahead of the curve and potentially increase their earnings.
3. AI is projected to reduce workforce requirements in some companies by up to 50% within 24 months.
4. AI can be used in various fields such as text generation, copywriting, customer support, code generation, image editing, video editing, sound editing, and hiring.
5. AI can be integrated into existing workflows and scaled up. It requires standardization, documentation,

Keynote

• AI can be used in good way or bad way based on our consciousness. It can extend us.
• Seeking to create services & products that can change a billion lives
• ChatGPT fasted adopting techology
• If you are in the top 1% of the users that learned that, you’ll stay ahead of that by far
• with AI enabled your salary goes up
• Mindvally will reduce about 50% of their workforce with 24 months
• projects 3.5 days work week
• AI more
• ChatGPT - be polite
• Ai is incredible to create ideas
• Prompt … Let me know how you understood the assignment
• Commit to going deep on AI!!!
• tool 11labs – voice digitization and reproduction
• Take a meditation - script it - ask ChatGPT for changing - ask 11labs to add audio … done
• Usages
  • Text generation
  • Copywriting (already now with 80-90% less invest)
• Give me 7 points that make Social Media articles go viral
  • apply that to your PodCast, Video
• Generate interview scripts
  • Describe company
  • Describe values
  • Generate interviewing questions
• Airtable
  • Spreadsheet combined with Database … NoCoder
• McKinsey - 20% of time of a worker is spent searching for information
• You can run AI on the data, when you externalized

How Entrepreneurs are using AI

• Text generation and Copywriting
• Customer Support and Experience
• Code generation and debugging
• Image generation and Editing
• Video generation and editing
• Sound editing and music generation
• Social media video analysis
• Hiring and interviews

Idea

• Story to Midjourney picture
• Generated Voice

How to write books in your voice?

• Don’t need to worry about a sucky writer.
• Speed up book writing time
• The Code of an extraordinary mind

Step 1 Analyze text

The Tool ChatGPT Chunker  divides a large input text into smaller chunks based on the specified token amount, and prompts ChatGPT accordingly.

Create a reusabe prompt

Output: Instructions: Write a conversational and engaging article about topic. Use a personal voice, share anecdotes and experiences to make the content relatable and easy to understand. Add a sense of humor, using light-hearted language and references. Maintain a friendly and encouraging tone of voice, aiming to guide the reader towards self-improvement and personal growth. Use clear and concise language, employing analogies and metaphors to better explain concepts. Combine expository and narrative literary styles, weaving personal stories and experiences throughout the text.

• 💡 Use my essay and analyse that regarding my writing style

• You have to know prompt engineering!
• You need to be in the top 1 percent of the AI people
• Mindvalley - the most transformative platform on the planet
• If you don’t keep up to this game, you will be super fast outperformed
• AI is gonna be part of our humanity (like books became part of it)
• get your kids on AI
• how fast can you enter the AI game
  • 10 minutes … to already make a huge mark on your business
• It is how you solve problems. It is a mindset shift

Great facilitation prompt: Just take 60’’ to write down at least one takeaway from this session

Using linked data

Who can code? Who can write logic in Excel sheets?

2 requirements

• Build linking data (with Airtable)
• add to your identity that you can program

enables complete automations with NoCode applications.

Push it to one database and establish a single source of truth. Everything gets connected. The inner creator gets unleashed.

Why linked data for AI?

• Learn first ChatGPT and MidJourney.
• Integrate AI into existing workflows and transform them
• Implement AI at scale

Steps

Standardize

• Will I use that process again? Yes .. standardize (at least document it)
  • define the data needed
• Define what AI tools you will use

Document

• the output you get from Airtable or elsewhere (direction of Second Brain)

Iterate and Improve

Airtable to AI linking to e.g. create copy variations. Just via API connection between Airtable and ChatGPT. Connected with AdSale to automatically create ads.

Airtable integrates AI directly with the database (… need to learn more about Notion and AI).

a few things to master

• Build Database Structure (Tables & Views)
  • Idea: need to get our pathways in database descriptions

Talk from Jasmine Taillon

Potential Legal Issues

• Copyright Infringement
• Trademark Infringement
• Defamation
• Privacy and Data Protection
• Ethical Implications

Copyright issues

• How much human contribution is needed in order to own the copyright?
• Advice: document what your human elements are

Example Zarya of the Dawn - How much human contribution is needed

• Copyright registration was granted using Midjourney.
• Copyright Office registered both the entire work and individual images contained.
• Copyright Office has the authority to cancel a registration after giving 30 days to defend a registration.
• Applicant was unable to show sufficient human authorship and the protection of the images was revoked.

US Copyright positioning

AI and Privacy

Privacy issues:

• GDPR, state private laws, industry specific laws
• what are you inputting and what are you outputting

INPUT CONCERNS

• What are you inputting?
• Is there consent or another justification?
• Can the personal information be used in a way consistent with privacy rights?

OUTPUT CONCERNS

• Is AI being used in a way that raises privacy concerns?
• Is it being used to make decisions that impact privacy?
• Do people understand how the technology is used and how it might impact them?

Ethical implications

We have to account for the biases. Information might be biased by:

• human bias
• data bias
• algorithm bias

Tips and Solutions

• Consider confidentiality and privacy
• Consider protectability of the outputs (human authorship still required)
• Beware that output may infringe
• Check accuracy of outputs
• Make outputs your own
• Always have at least one substantial human inventor
• Realize there is no current ruling on mixed inventorship. Document the
• human inventor side!

Presentation Iman Oubou

Nice teaser question: How are you feeling about this AI stuff

Mindset for working with AI: Be curious, be creative, ask the right questions.

Estimated AI value creation

For instance, PwC had projected that AI could contribute up to $15.7 trillion to the global economy by 2030, which is more than the current output of China and India combined. Of this, $6.6 trillion was expected to come from increased productivity and $9.1 trillion is likely to come from consumption-side effects.

Additionally, according to the McKinsey Global Institute, AI could potentially deliver additional global economic activity of around $13 trillion by 2030, which would be about 16 percent higher cumulative GDP compared to today.

Artificial Narrow Intelligence and General Intelligence - ANI vs AGI

• ANI (artificial narrow intelligence) - smart speakers, chatbots, Image recognition, web search
• AGI (artificial general intelligence) - do anything a human can do. AGI still a long way off (decades).

ANI

• focused on a specific, singular or limited task
• Examples Include image recognition, hyper- personalization, chatbots, predictive text
• Trained on specific tales by data scientist
• no self-awareness, consciousness, ability to think

AGI

• not fully realized, with some developers questioning if it will be possible
• machines that can handle a range of cognitive tasks with little oversight
• The ability to learn, generalize, apply knowledge and plan for the future
• Must consistently pass the Turing Test
• General intelligence that possesses common sense and creativity and expresses emotions

The AI Doomer

Short Term Concerns

Need to expect a tremendous shift in the job market.

Prepare yourself

• What kind of new skills do we need?
• How should we be organized?
• How do we start defining new jobs?
• How can we bring our people along with us?

Understanding Generative AI

Key Characteristics of the Foundation

• Pretrained with vast volumes of data
• Can perform a wide variety of tasks
• Extremely large (BERT was trained with 3 billion words, the equivalent of reading the entire lHarry Potter series 350,000 times)
• Highly adaptable via natural language prompts
• Self-supervised (Data is unlabeled, the modal learns by finding patters in raw data)

Overview on the Foundational Models

• Large Language Model (LLM) … Text
• Multimodal Foundation Model … Text, audio, Images
• Visual Foundation Model (VFM) … Text, Images
• Speech Recognition Model … Audio
• Language Processing Model … Text

Consider different main tools

Text mainly

• ChatGPT
• Google Bard
• Anthropic

Images

• DALL-E2
• Stable diffusion open source

Use cases for Business

https://www.insidr.ai/ - to find the best AI tools

5 Steps

1 - Define your goals and challenges

Clearly define your goals and identify the challenges you face in your internal business operations

Example Goal: Improve customer satisfaction/client support Challenge: Manual processing of a high volume of customer inquiries, leading to delays and inefficiencies

2 - Assess your data capabilities

Understand if you have the necessary data to drive AI solutions. Find the gap to address.

Access what you have data on (eg):

• Customer Purchase History
• Inventory Levels
• Customer Feedback

Access what you lack data on:

• Real-time customer sentiment

3 - Identify AI use cases

Identify specific use cases where AI can bring value to your business operations

For example:

• Using natural language processing
• Automate customer support Inquiries by developing a chat bot
• Reduce response times and free up your customer support team

4 - Find AI tools or develop AI solutions

Once you have identified the use cases, it may even be time to develop your own AI solutions.

For example:

• Partner With an AI vendor or build an in- house team to develop the chatbot
• Train the chatbot on historical customer interactions

5 - Implement and Monitor

Implement the AI solutions and closely monitor their performance.

Prompt engineering - How to structure prompts

• Specify the scenario
• Assign a role in the instruction
• Provide the model with additional information
• Use specific formulations
• Formulate the instructions clearly

Understand that the previous conversation forms context:

• either start new chat or
• prompt: ignore all previous instructions

https://www.promptpal.net AIPRM chrome extension for ChatGPT Google plugin

Understand APIs

• Commercial API-based distribution: OpenAI, Anthropic, Microsoft, Google
• Open-source distribution: Databricks, Stability Al, Meta (formerly Facebook)
• fast response
• leverages more sophisticated models
• low barrier to entry

Naveen Jain

Naveen Jain VIOME

AI to fundamentally transform our lifes.

• Why do you want to do that? Why do you care?
• How many other ways can go to reach the same goal?
• AI is the collection of all the knowledge that we have … new knowledge is created via AI
• Ancient intelligence is ours … that trains AI … through our storytelling
  • so we need to give it something
• massive amount of fear
  • AI is beginning of destruction of humanity
  • problem with misinformation and conflicts created through that
• AI can make us do any things …
• AGI still years to come
• LLM -
  • Underlying architecture of ChatGPT is around for 15 years
  • processing power increase a lot
  • a lot of money on the table
  • huge amount of data
  • mainly text, audio, video
• each educational stock went down by 50%
• ChatGPT can modify to teach different ways – through gaming, logic, … so specific teaching becomes super possible
  • their assets become their liabilities
  • teachers & teachers unions - instant liabilities
• Disruption in Healthcare
  • sensors - amount of data we can collect becomes cheaper and faster
  • amount of sensor data collected can now be processed with massive storage power
  • AI super power to analyze data

AI to cure illness

• digitize the human body
  • then understand what changes on our body
• what makes our genes express is our lifestyle
• 600k humans, 700 Trillian datapoints

Health focus on

• Nutrition
• Stress reduction (fight or flight response … shutting down all non necessary functions)
  • find cause of stress
  • meditation
  • food cannot be digested properly
• Exercise - raise your heart rate, use your own body
• Sleep - quality matters (oraring), eight sleep
• find your purpose (they live 10-15 years longer; find something you care and can find happiness) .. what will you do for and life for it
  • love yourself; respect yourself

Recap

• AI is not bad … don’t belief the naysayers
• new level of creativity
• new level of human health
• humans to be more human … it gives back humanity
• DATA-DATA-DATA
  • what data am I accumulating
  • look at Airtable
• It is not all about ChatGPT

Linking

• ➡ Notes from Day 2 of the Mindvalley AI Summit 2023
• MOC Artificial Intelligence
• Interview with Mo Gawdat - AI is worse than climate change based on [[Notes from Mo Gawdat]]

Enjoy this post?