✏️ Why You Might Beat ChatGpt at Math

Have you ever asked ChatGPT to help you with a calculation only to realize it’s just as confused as you are? Well, there’s a reason it acts more like a liberal arts major than a mathematician. Find out more about this vulnerability, and what data scientists are doing to fix this.

What to Expect:

• Why ChatGPT Struggles With Math

• Founder Spotlight: Mark Zuckerberg

• What even is Artificial Intelligence?

• Level Up Your Job Interview Prep

• More AI Tech, Tools, and Talks

AI NEWS
Why are AI Chatbots Particularly Bad at Math?

ChatGPT and other large language models (LLMs) have undeniably revolutionized natural language processing. However, when the conversation turns to numbers, these models often stumble. But, why?

🤓 Deep Dive:

• What they're good at: AI chatbots are built to predict language patterns rather than follow rigid mathematical rules. They are trained to use probabilities to predict the very next word or phrase and generate a response - a strategy that works great for language but falls short for precise tasks like math.

• What they're bad at: Rules-based tasks. While AI is great at finding patterns, it’s not so great adhering to logical, pre-defined rules and formulas that math requires.

• The current hack for it? Well it is simple really: they use calculators on the background, to do the hard work for them. This hybrid approach allows the chatbot to delegate arithmetic tasks to a specialized tool while it focuses on understanding and generating language.

• 🔥 Hot Debate - Silicon Valley thinks the solution is to provide LLMs with more complex math data, believing it could eventually lead to AI thinking for itself through deductive reasoning. However, Yann LeCun, chief AI scientist at Meta, argues that simply adding more data isn't enough - he advocates for a broader approach called "world modeling," which aims to teach AI how the world works, similar to human learning.

But we doubt this will be a problem for long. Google’s DeepMind just announced a model called AlphaProof with advanced mathematical reasoning skills that was able to solve 4 out of 6 problems from the International Math Olympiad, earning it a silver medal.

FOUNDER FILES
From Facemash to Facebook to Meta

Not many people can say they turned a college hobby into a global empire, but Mark Zuckerberg definitely can.

How it started: It’s 2003, and a young Harvard student named Mark Zuckerberg is chilling in his dorm room, coding away on a project called “Facemash.” This cheeky site let students rate each other's photos, a bit like Tinder before Tinder, but without the swiping. Although the site was short-lived due to Harvard execs quickly shutting it down, it inspired Zuck’s next iteration, called TheFacebook, which he rolled out early 2004 (with help from four of his fellow friends) as a way for Harvard students to look up information on other students. What started as a Harvard-only student directory, quickly expanded to other colleges and high schools, until in 2006, they announced it would be open to anyone, cementing themselves as THE global social networking service.

How it shifted: Fast forward to October 2021, and Facebook undergoes a major transformation, rebranding as Meta. Meta’s name change reflects a complete shift in vision for the company. Moving beyond just social networking, Zuckerberg and his team are doubling down on building the metaverse, believing it’s the next big evolution in how we connect online. Zuckerberg envisions Meta as a pioneer in this virtual universe where people can work, play, and socialize in immersive digital environments. Think of it as the internet 2.0, but with VR headsets and virtual real estate.

How its going: Meta is up to ALOT these days:

• Virtual Reality - The company is heavily investing in virtual and augmented reality technologies, like the Oculus Rift and Quest headsets, aiming to make the metaverse a mainstream reality (FYI, they bought Oculus for $2B in 2014 😲). But, that’s just the tip of the iceberg… Meta’s Reality Lab has accumulated around $45B in losses since the end of 2020. But Zuck is in it for the long haul, telling investors to ride the wave, and they will be rewarded.

• AI - Meta’s most exciting AI development is the latest release of their large language model, Llama 3.1 on July 23rd. Why is this so significant? Llama 3.1 can compete with the most advanced models in the market like GPT-4 and Claude 3.5 Sonnet, but that is not what sets it apart: It’s that it is completely open-source unlike its counterparts. This means that anyone can use Llama for free. You heard that right, you can download this powerful LLM on your own machine and use it any way you like. By removing financial and proprietary barriers, researchers, developers, and entrepreneurs will be able to build state-of-the-art AI systems without restrictions - ushering in a new era of AI tools. Check out this beginner guide to get started.

📝 Takeaway: Take risks and don’t be afraid to adapt and evolve your idea.

💧Fun fact: Zuckerberg chose blue as Facebook’s color theme because it is the color he sees best (he is red-green color blind).

CONCEPT
What is Artificial Intelligence?

By now, you’ve definitely heard the words ‘Artificial Intelligence’ and ‘Machine Learning’ thrown around, but you still may be a bit confused about what exactly AI is and what it does. Let’s break it down in simple terms —

🧠 What it is: Artificial Intelligence is the ability of a machine to simulate the intelligence of a human. These systems are designed to think, learn, and solve problems like humans. Think of it as giving a computer a super smart brain. Whether it’s recognizing your voice, understanding your text messages, or making decisions based on data, AI is the tech that powers it all.

🤖 How does it work: Without getting too deep into the specifics, AI is able to mimic human behaviors by doing the following:

1. Collect Data - AI starts by gathering vast amounts of data—think of it as the brain’s memory bank. Just like how our brains store experiences and knowledge, AI stores data. This data can include anything from images and text to sensor readings and user interactions.

2. Identify Patterns - Our brains don’t just learn from reading or studying, they also learn from experience. Similarly, AI utilizes machine learning algorithms to analyze the data it has gathered, learn from it, and then make predictions. It is essentially “training” itself to become smarter and more accurate.

3. Make Decisions - Once the AI has learned enough, it can start making decisions based on the patterns it has identified. It’s able to do this by leveraging a specific ML algorithm, called neural networks - this is a complex, interconnected network (again, just like our brain) that can process information in multiple layers, enabling sophisticated decision-making. It’s important to remember that AI is not following pre-set instructions when making a decision, it is actually learning and adapting on the fly based on the information it has gathered.

4. Iterate Performance - The time old saying, “Practice makes perfect” holds true here. You can’t get better at something if you don’t put in the effort to improve… Well the same goes for AI. AI systems are designed to learn and adapt continuously. The more data they process, the better they become at their tasks.

🤔 Why does it matter?   As AI becomes increasingly integrated into our lives, understanding its fundamentals is essential for navigating the digital world. It won’t be very long until our cellphones are powered by AI! So knowing what goes on behind the scenes can go a long way in helping to make AI less intimidating and more accessible.

PROMPT OF THE WEEK
For: Job-Seekers

I'm applying for a [Job Title] job at [Company]. I'm going to set the context for you by sharing the job description and company profile first:
[Company Profile + Job Description]

Now, I want you to read my resume for further context.
[Resume Text]

Share 7 questions I can expect in the [Job Title] interview at [Company Name] based on the job and give me pointers on how to answer each.

Source : Aatir Abdul Rauf / Substack

TECH TOOLS, TIPS, AND TALKS

📖 What we’re reading: That Will NEVER Work - The Birth of Netflix and the Amazing Life of an Idea
📻 What we’re listening to: Lenny’s Podcast - 5 essential questions to craft a winning strategy with Roger Martin
💻 What we’re using: Poe AI - A chatbot aggregator that gives you access to all the latest LLMs in one place

MORE READING

• OpenAI Announces SearchGPT prototype - Is Google’s Reign Over?

• Apple Releases Sneak Peak of New AI Features - Will You Generate Your Own Emojis?

• ChatGPT Releases Advanced Voice Mode - Find Out How You Can Talk to ChatGPT

We want to empower our readers with actually insightful knowledge so that they are more confident, informed leaders. Because let's face it, AI could be running the world pretty soon... so shouldn't we at least know how it works? If you are curious about a topic and want to learn more, drop us a message below👇🏼