Embracing AI Innovation: How McKinsey, A Century-Old Firm, Boosts Productivity with AI

Mornin’ miners⛏️,

Happy Thursday!

Welcome to "Digger Insights" - your daily 7-minute tech enlightenment! In our enjoyable and easily digestible dispatches, we break down the latest tech trends, giving you a quick and comprehensive view.

Join us and in just a few minutes each day, you'll gain an advantageous perspective on the ever-evolving tech landscape. Ready to dive in?

Let’s get to it!

Today’s Highlights:

• Economist Jeremy Siegel Stays Optimistic with Big Tech Boom 👍

• Researchers from MIT Develops Self-Learning Language Model 🤓

• Almost 50% of McKinsey Employees Embrace ChatGPT 💻

Wharton Professor Jeremy Siegel Stays Optimistic Regarding Big Tech Boom and Bubble

It is no question that the tech industry has been going through quite the AI craze. The demand for AI-powered chatbots and high-powered graphics processing units has increased, resulting in investors pilling into certain stocks.

Professor of finance at Wharton School at The University of Pennsylvania, Jeremy Siegel, was invited to speak on CNBC’s “Street Signs Asia”, discussing the rise in AI tech stocks and how it could possibly affect the economy.

Nvidia Rise

One example of an AI tech stock rise is in Nvidia AI chips. The demand for Nvidia chips used in AI reached an all-time high and brought Nvidia’s valuation to nearly $1 trillion. CEO Jensen Huang stated that Nvidia’s shares were up 166% year-to-date.

On May 28, Nvidia unveiled its new large-memory AI supercomputer to enable the development of next-gen models for generative AI applications. Huang stated that generative AI, large language models, and recommender systems are the digital engines of the modern economy.

Stock Bubble

The innovations companies like Nvidia bring to further improve AI technology, along with its rising demand, are feared to cause big perturbations such as the stock market bubble.

Concerns for a bubble similar to the late 1990s dot-com stock bubble have been appearing, but Siegel has stayed optimistic. Siegal remarked that “it’s not a bubble yet” and that the rise in AI stocks can bring efficacious results instead. Its rise has helped lift the S&P 500, possibly preventing a banking crisis.

This notes that AI seems to be a pioneering and progressive aspect of the tech industry, and though individuals, entrepreneurs, and businesses are certainly welcome to keep innovating with it, it is always safe to pay attention to the impact of what you’re tinkering with.

Researchers from MIT Innovate with Scalable and Self-Learning Language Model

With the growing amount of large language models, such as GPT models from OpenAI, and LaMDA from Google, MIT researchers Hongyin Luo, James Glass, and Yoon Kim aimed to scrap the preconceived notion that smaller models embody limited capabilities, and have succeeded by creating their own algorithm.

This MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) research team has developed SimPLE (Simple Pseudo-Label Editing), a language model algorithm that utilizes self-training. With this technique, the model has the capability of learning from its own predictions and adopting natural language understanding.

Smaller but Better

The CSAIL team has argued that LLMs, although revolutionary and dominating, have distinct limitations when it comes to comprehending tasks.

To prevent their small model from having similar limitations, the model is trained to grasp the core principle of language understanding; contextual entailment. Contextual entailment represents a connection between two sentences, where if one sentence is true (the premise), it is probable that the other sentence (the hypothesis) is true as well. The research team trained the model using a model that understands this notion.

This training enhances the model’s comprehension of diverse language tasks as well as its versatility and adaptability. The researchers also employed a natural language-based logical inference dataset which enables the model to ascertain and adapt to a broad spectrum of tasks.

SimPLE’s Safety and Sustainability

Traditionally, language model training requires utilizing LLM APIs or humans to annotate manual data. This method could compromise and expose sensitive data and information. SimPLE’s method allows every dataset to be annotated without sharing any sort of data with the annotator, with templates describing each task, ensuring information safety and privacy.

By developing a smaller model that is potentially equally as powerful as large models, this research team has paved the way for future AI technologies that prioritize scalability, sustainability, cost-efficiency, and privacy preservation.

The CSAIL team’s next plan is to further enhance their self-trained model’s capabilities by training them to align between claim and fact/moral principles, which would detect misinformation, hate speech, and stereotypes.

So, depending on what you need, smaller language models might be the better option, especially for professionals and business owners who appreciate less computational costs and privacy risks in handling sensitive data.

Read the paper regarding SimPLE, here!

McKinsey and Company Employees Embrace ChatGPT

Consulting firm McKinsey and Company has been around for almost 10 decades, first founded in 1926 by the University of Chicago’s James O. McKinsey. You would think that being long in the tooth warrants the struggle to keep up with the times, but that is far from true for McKinsey.

With 30,000 employees across 67 countries, McKinsey’s senior partner, Ben Ellencweig, claims that nearly 50% of the firm’s workforce embraces ChatGPT and similar technology.

McKinsey’s AI Usage

Being the leader of alliances and acquisitions at McKinsey’s AI consulting arm, QuantumBlack, Ellencweig has allowed McKinsey employees to use generative AI with certain guidelines. He ensures that McKinsey employees do not upload confidential information and are simply utilizing AI services for more efficient work.

McKinsey’s use of generative AI depends greatly on what its consulting customers and businesses need. McKinsey had done research on how they use generative AI, which includes what Ellencweig calls “the four Cs”; coding, customer engagement, creative content generation, and content synthesis.

By using ChatGPT and similar tools, McKinsey’s client software developers have had about a 35-55% increase in coding productivity. Generative AI is also used to offer more personalized customer interactions and streamline content generation processes, refining audience segments in the process. Lastly, generative AI has also shed new light on how businesses can combine different data points and services.

Generative AI Safety

Global leader of QuantumBlack, Alex Singla, provided a five-step framework for McKinsey employees to ensure the secure use of generative AI. He shared aspects to consider when approaching generative AI, which include paying attention to the IT stack and infrastructure and use of structured or unstructured data, choosing the right AI model, the UI and UX, and the ability to manage the changes generative AI could bring to a company.

This framework and McKinsey’s openness to generative AI may inspire and be applicable to entrepreneurs and professionals with businesses that have plans to integrate generative AI into their day-to-day management.

Favorite Product of The Day

Are you interested in knowing more about the tech industry? You want to learn about AI and no-code tools but don’t know where to start (aside from us, obviously)? Well, let us introduce you to The ProHuman AI!

The ProHuman AI is a newsletter designed for the everyday person. It provides simplified but in-depth information regarding AI, technology, no-code tools, and side hustles.

It breaks down the latest advancements for you so you can stay ahead without giving yourself too much of a headache, and shares resources and tips so you can level up your side hustle game!

You can check out and subscribe to The ProHuman here.

Meme & AI-Generated Picture

Job Posting

• System1 - Senior Data Engineer - Marina Del Rey, CA (Remote/Hybrid)

• Invitae - Senior Full Stack Engineer, Patient Data Network - San Francisco, CA (Remote/Hybrid)

• Grand Studio - Senior Research Designer (Contract Pool) - Chicago, IL (Remote/Hybrid)

Promote your product/service to Digger Insights’ Community

Advertise with Digger Insights. Digger Insights’ Miners are professionals and business owners with diverse Industry backgrounds who are looking for interesting and helpful tools, products, services, jobs, events, apps, and books. Email us [email protected]