ai or ia
Hey everyone, it’s Dan back from a hiatus with my perspective on AI use cases and the broader impact.
Today I am reframing the term ‘artificial intelligence.’
More on that further down.
Taboo
The topic of AI can be controversial both inside and outside of the tech space.
And for good reason. It’s taken over as the hot investment.
VC dollars are showered onto startups promising to solve specific problems using AI. Even legacy companies are scrambling to get in on the action.
But for the general public, these innovations can invoke feelings of general uneasiness and dismissal.
It can be scary to think about disruption, let alone research and stay up to date on everything that’s happening in this specific field.
That’s why I’ll do my best to summarize what I’ve learned so far. I’ll also identify a few practical and useful ways I’ve incorporated it into my life.
As always, hyperlinks will be included to dive deeper.
Let’s jump in.
Prompt: “How many possible moves are there in a game of chess?”
The number of possible moves in a chess game is estimated to be around 10^120. This enormous number arises from the branching factor of the game, which is the average number of moves available at each point in the game. With an average branching factor of around 35, the number of possible move sequences quickly becomes astronomical. It is worth noting that this estimate includes both legal and illegal moves, as well as moves that result in the same game position being repeated.
-Notion AI
Man vs Machine
In a 1996 chess match, IBM's Deep Blue, an artificial intelligence system, was able to defeat Gary Kasparov, who was then a grandmaster world champion and an absolute powerhouse in chess.
The event was a spectacle and drew viewership from every corner of the world.
Gary Kasparov vs Deep Blue
It’s important to note that Game 1 went in Deep Blue’s favor in just 37 moves, forcing Karsparov to resign. This event marked a turning point in chess history as this was the first time a reigning world champion ever lost against a computer with tournament conditions and slow time controls according to chess.com
Every move is also mapped here if you’re strategy-curious.
Following the initial upset, Kasparov bounced back, eventually winning the series.
1996 Result: Kasparov–Deep Blue: 4–2
Due to the popular global demand for the matchup, they went head to display again in 1997 atracting even more viewership.
The matches were intense and Kasparov had to focus. He had to heighten his sense of adaptability to handle anything Deep Blue could throw at him.
A big disadvantage for Kasparov was not knowing how much or what kind of data it was trained with.
1997 Result: Deep Blue–Kasparov: 3½–2½
The results of the rematch were both astonishing and shocking for the world.
Following the matchup, Deep Blue's victories were seen as a symbolically important event, indicating that artificial intelligence was becoming more comparable to human intelligence.
Chess and AI have been really close ever since. Nowadays it’s almost impossible to beat the chess AI options out there.
Here is a video explaining why.
Go
After Deep Blue's win, the game of Go became the best example of a game where humans did better than machines.
Go is a popular game in China, South Korea, and Japan, known for its simple rules and many possible moves.
In the past, even players with less than a year of experience could beat the best Go programs. Over time with repetition, these programs got better.
In 2015, Google DeepMind's AlphaGo program defeated the European Go champion, Fan Hui, in a private match.
Then, it surprisingly beat the top-ranked player, Lee Sedol, in the AlphaGo versus Lee Sedol match in 2016.
Go is now the testing focus because it involves:
strategy, creativity, and ingenuity.
“Go was long considered a grand challenge for AI. The game is a googol times more complex than chess — with an astonishing 10 to the power of 170 possible board configurations. That’s more than the number of atoms in the known universe.'“-Google DeepMind
Key difference: Deep Blue used brute computational force to analyze millions of positions, while AlphaGo used neural networks and reinforcement learning.
By building programs and neural networks to play these games against humans, refinement is on a continuous loop. It only gets better with more games played.
Summary
Artificial intelligence is the overarching system.
Machine learning is a subset of AI.
Deep learning is a subfield of machine learning.
Neural networks make up the backbone of deep learning algorithms. It’s the number of node layers, or depth, of neural networks that distinguishes a single neural network from a deep learning algorithm, which must have more than three.
These fundamental concepts play a significant role in the space.
Implications
Is AI perfect? Far from it. As any technology grows, it will have it’s ups and downs.
Reflecting on the ethical considerations surrounding AI is crucial.
As AI technology becomes more advanced and integrated into various aspects of our lives, it raises important questions about privacy, security, and employment.
One key concern is the potential invasion of privacy.
AI systems often rely on collecting and analyzing large amounts of data, which often include user data.
This raises questions about how this data is sourced, used, stored, and protected.
Similar to other industries, clear guidelines and regulations are needed to ensure that individuals' privacy rights are respected.
Another ethical consideration is the potential for bias in AI algorithms.
AI systems learn from data, and if the data used to train them is biased, it can result in biased outcomes.
The impact of AI on employment is also a significant ethical consideration.
As technology advances, there is a concern that it may replace many human jobs.
It’s important to consider the social and economic implications of AI adoption and ensure that appropriate measures are in place to support workers and facilitate a smooth transition.
Use Cases for Varying Professions (AI generated)
Healthcare: Used for medical diagnosis, drug discovery, and personalized treatment recommendations. It can analyze medical images, such as X-rays and MRIs, to assist doctors in detecting diseases and abnormalities.
Finance: Used for fraud detection, risk assessment, and algorithmic trading. It can analyze large volumes of financial data to identify patterns and anomalies, helping financial institutions make informed decisions.
Marketing: Used for customer segmentation, personalized marketing campaigns, and sentiment analysis. It can analyze customer data and behavior to target specific audiences with tailored messages and optimize marketing strategies.
Manufacturing: Used for predictive maintenance, quality control, and process optimization. It can analyze sensor data from machines to detect potential failures and recommend preventive actions, improving overall efficiency and reducing downtime.
Education: Used for adaptive learning, intelligent tutoring systems, and plagiarism detection. It can personalize educational content based on individual student needs and provide real-time feedback and support.
Retail: Used for demand forecasting, recommendation systems, and inventory management. It can analyze customer preferences and purchase history to make accurate predictions and offer personalized product recommendations.
My use case
The name that continues to come to mind is IA, or intelligent assistant. So let’s roll with that.
Here are my top IA tools I’ve been using in 2023 in no particular order.
#1 Notion
Notion is a powerful productivity tool that offers various features to help users organize and manage their work. It recently introduced a new feature like Q&A that’s in beta and I’ve been testing, allowing users to easily search for and recall information within their workspaces, offering more context. The process is known as retrieval augmented generation, or RAG.
Retrieval Augmented Generation (RAG) is a feature that enhances the search functionality within the platform.
It allows users to search for and recall specific information from their workspaces by providing intelligent and context-aware responses.
RAG leverages AI technology to improve the user's ability to find relevant information quickly and efficiently.
It provides users with a powerful tool to navigate and organize their workspaces, making it easier to access and utilize their stored knowledge.
Notion's diverse range of features makes it a valuable tool for organizing information, improving productivity, and enabling creative possibilities.
Notes:
awesome custom templates made by their community
tons of integrations
clean design
team is shipping quickly
#2 ChatGPT (OpenAI)
OpenAI has developed ChatGPT, which is an AI model designed for conversational interactions.
ChatGPT allows users to have interactive conversations with the AI and get intelligent responses.
It is a powerful tool that can be used for various purposes, such as answering questions, providing information, and assisting with tasks.
It’s considered a leading contender in the AI race due to its advanced conversational capabilities and its potential to revolutionize human-computer interactions.
Its ability to provide context-aware responses and its potential for real-time information retrieval make it a frontrunner in shaping the future of AI-driven conversational interfaces.
ChatGPT popular prompts. Try it here.
Notes:
could replace the traditional web search as it becomes more accurate
user-friendly user experience
an interesting mobile app that lets you have a conversation
partnering with everyone
shipping quickly
Grammarly
Grammarly is an AI-powered writing assistant that helps users improve their writing skills. It offers grammar and spelling checks, as well as suggestions for clarity, conciseness, and tone.
Grammarly can be used across various platforms, including web browsers, Microsoft Office, and mobile devices.
It provides real-time feedback and suggestions to enhance the overall quality of written content, making it a valuable tool for writers, students, professionals, and anyone looking to improve their writing proficiency.
Notes:
useful way to have something proofread
easy integration with browsers as an extension
Canva
Canva is a popular graphic design platform that helps users create visually appealing designs. It offers pre-designed templates, customizable elements, and a wide range of images, icons, fonts, and colors.
With Canva, multiple users can collaborate on a design simultaneously and easily share their work. It has a user-friendly interface and drag-and-drop functionality, making it accessible to users with varying design skills.
Canva offers free and premium subscription plans, with the premium plan providing additional features and resources.
With their new suite of AI tools, users have access awesome resources like
Notes:
Dropped Canva AI and made a great product even better
User-friendly interface
Freemium model lets users build confidence before commiting
Duolingo
Duolingo is best language app in the world. It uses artificial intelligence to help users learn languages effectively. Through its AI-powered algorithms, Duolingo offers personalized language courses and adaptive exercises tailored to each learner's proficiency level and learning style.
The platform uses AI to track users' progress, identify areas for improvement, and provide targeted feedback and recommendations. It employs natural language processing to analyze learners' responses and assess their language skills. Duolingo's AI technology enables interactive and engaging language learning experiences. It incorporates gamification elements, such as achievements and leaderboards, to motivate and encourage learners to practice regularly.
By leveraging AI, Duolingo aims to make language learning accessible, efficient, and enjoyable for users worldwide. It continues to innovate and enhance its AI capabilities to provide an effective and personalized learning experience for language learners of all levels.
Notes:
Partnered with OpenAI to build out Duolingo Max
AI-focused from their inception
Apple
Machine learning is crucial for Apple, enhancing user experiences and enabling advanced features.
One area where Apple uses machine learning is Siri, its voice assistant.
Siri uses machine learning to understand user commands and provide personalized responses.
It can set reminders, send messages, play music, all through voice commands. See more commands here.
Machine learning is also used in Apple's camera and photo features. The Photos app uses machine learning to organize and categorize photos, recognize faces and objects, and suggest photo albums and memories.
Apple's News app and Apple Music also utilize machine learning for personalized recommendations based on user preferences and behavior.
Apple prioritizes user privacy by processing data locally on the device with technologies like Federated Learning.
“Many of these AI applications were trained on data gathered and crunched in one place. But today’s AI is shifting toward a decentralized approach. New AI models are being trained collaboratively on the edge, on data that never leave your mobile phone, laptop, or private server.” -IBM
Apple recently announced a big breakthrough in running LLMs on iPhones.
LLM-based chatbots like ChatGPT, Claude, and Bard are very data and memory-intensive.
This poses as a hardware issue for devices like iPhones that have limited memory capacity.
To address this, Apple AI researchers have developed a technique using flash memory (where apps and photos live) to store the AI model’s data.
This is beneficial because it allows the AI models to run up to twice the size of the iPhone’s available memory and translates to a 4-5x increase in speed on standard processors (CPUs) and 20-25x faster on graphics processors (GPUs)
What’s with the tech jargon? I thought this thing was supposed to keep things super simple?
Okay okay, you’re right, I got ahead of myself.
So what does that mean exactly and why is it beneficial to the general public?
This breakthrough means that Apple can run advanced LLMs on almost any hardware device in the ecosystem. Furthermore, it could reinvent the whole experience.
A few examples:
Advanced and very accurate Siri
Bump in Photo and AR features
Real-time language translation
Overall, Apple is well-positioned to successfully leverage machine learning in its products and services to provide enhanced functionality, intelligent features, and personalized experiences while protecting user privacy.
Closing Time
The technology is here, it cannot be un-invented. It’s up to people to understand the implications, and how it can be practically applied to our lives.
If you've made it this far, congrats!
Hopefully these thoughts helped you sparked an interest and learn more.
Don’t forget to check out the resources below and subscribe!
Additional AI Resources
Online Courses:
Free Video:
Newsletters: