Skip to content

CONTRIBUTORS

Matthew Moberg 

Portfolio Manager
Franklin Equity Group
United States

We have long argued that we are in the beginning stages of the Fourth Industrial Revolution—a structural transformation of the entire economy. This revolution is being built on a series of new, powerful technologies which includes Artificial Intelligence (AI), Genomics, and Robotics. An economic shift of this scale and complexity proceeds in fits and starts—but a sampling of innovations from this quarter demonstrates significant progress in the right direction.

AI can diagnose tuberculosis

Researchers recently built a model called HeAR that can listen to humans and identify early signs of potential disease. They trained the model on 300 million sounds, including 100 million coughs. The model was able to diagnose patients with tuberculosis based on the sound of their cough alone.1

Why it matters: AI diagnosis by sound alone is a novel advancement. Around three million cases of tuberculosis go undiagnosed each year—often because people lack access to health care. Although it can be treated with antibiotics, if tuberculosis is untreated, one in three people will die. This new model will be deployed across India to provide a location-independent, equipment-free respiratory health assessment. Other potential use cases for “bioacoustic” models are myriad. Importantly, other models trained on HeAR performed well even with less training data, which is essential for health care applications in which data can be scarce.

A new wheelchair climbs stairs

South Korea’s Institute for Machinery and Materials recently unveiled an ingenious new design for wheelchairs. Inspired by the surface tension properties of water droplets, the team created a “smart chain” structure which enables the wheel to conform to the terrain around it. The chain is attached to spokes that change tension in response to different terrains.2

Why it matters: Patients in wheelchairs will have much greater mobility. Using the smart chain, the wheelchair can climb stairs and navigate rocky terrain with ease. This could be a game-changing invention for the ~130 million wheelchair users worldwide,3 perhaps rendering the ramps that currently accommodate wheelchairs a thing of the past. This sort of specific technological breakthrough is a great example of an innovation made possible on top of the general-purpose technology of robotics.

Exhibit 1: Illustration of AI in Health care

Specially trained AI models used in health care are already improving outcomes for patients

For illustrative purposes only.

Phone screens made of transparent wood

Researchers at the University of Maryland have created transparent wood. Wood is “made up of countless little vertical channels, like a tight bundle of straws bound together with glue.”4 To make it see-through, scientists first remove the glue, which is called lignin, and is responsible for wood’s brown color. The wood is then filled with a substance like epoxy resin that renders the wood transparent.

Why it matters: Americans crack more than 50 million phone screens a year, resulting in nearly US$3.5 billion in damage costs.5 Transparent wood could be a much stronger and less breakable material for phone screens in the future. It can also be used in color-changing windows and in glowing light fixtures.

New AI model Strawberry can “think”

Since ChatGPT took the world by storm in late 2022, AI models have continued to make gains. Strawberry, ChatGPT’s latest release, is the most powerful yet. Strawberry, rather than providing an immediate answer to a query, will “think” through questions. Effectively, this means that the model will generate multiple answers, review them, and select the most accurate and relevant.6 When asked to take a qualifying exam for the International Mathematics Olympiad, Strawberry’s predecessor, GPT-4O, correctly solved only 13% of problems vs. 83% for Strawberry.7

Why this matters: The release of Strawberry is a sea-change for AI models. The first iterations of large-language models (LLMs) were trained on huge datasets to memorize answers. In contrast, Strawberry has been trained to memorize “reasoning,” an approach that makes the model a much more adept problem-solver and less prone to error. The result is that Strawberry is now capable of acing difficult math tests, solving the Saturday NYT crossword puzzle, and even writing an entire codebase for a new videogame.

A new material may replace solar panels

An Oxford team recently developed an “ultra-thin material” that has 27% efficiency in converting sunlight into usable energy. With further progress, the team believes that this new photovoltaic device could exceed 45% efficiency. This is not a “panel” in the traditional sense—this material is thin enough to cover any common object and can therefore be used in a variety of settings.8

Why it matters: The current most commercially available solar panels peak at 23% efficiency. Further, the silicone used in these panels is far more expensive than the thin-film perovskite which underpins this new material. Because it is both cheaper and more effective, this new material shows remarkable potential in clean energy, with immediate applications in electric vehicles and smartphones. In a world where 82% of global energy still comes from oil, gas, and coal, this is welcome progress.

Endnotes:

  1. Source: “This AI model is helping researchers detect disease based on coughs.” Google: The Keyword. August 19, 2024.
  2. Source: Siberling, A. “TechCrunch Minute: This robotic wheelchair can climb stairs.” TechCrunch. September 5, 2024.
  3. Source: “Wheelchair Users.” Physiopedia.
  4. Source: Coleman, J. “The Surprising Possibilities of See-Through Wood.” Smithsonian Magazine. December 19, 2023.
  5. Source: Gilmour, J. “Americans break two smartphone screens each second, costing $3.4 billion a year, report says.” Miami Herald. 
  6. Source: Samuel, S. “The new follow-up to ChatGPT is scarily good at deception.” Vox. September 14, 2024.
  7. Source: Fried, I. “OpenAI releases ‘Strawberry’ model with better reasoning.” Axios. September 12, 2024.
  8. Source: Turek, S. “Oxford scientists make remarkable advancement with material that generates solar power without traditional panels: ‘The future is brighter’.” MSN. September 15, 2024.

 

WHAT ARE THE RISKS?

All investments involve risks, including possible loss of principal.

Equity securities are subject to price fluctuation and possible loss of principal.

Investments in fast-growing industries like the technology and health care sectors (which have historically been volatile) could result in increased price fluctuation, especially over the short term, due to the rapid pace of product change and development and changes in government regulation of companies emphasizing scientific or technological advancement or regulatory approval for new drugs and medical instruments.

The opinions are intended solely to provide insight into how securities are analyzed. The information provided is not a recommendation or individual investment advice for any particular security, strategy, or investment product and is not an indication of the trading intent of any Franklin Templeton managed portfolio. This is not a complete analysis of every material fact regarding any industry, security or investment and should not be viewed as an investment recommendation. This is intended to provide insight into the portfolio selection and research process. Factual statements are taken from sources considered reliable but have not been independently verified for completeness or accuracy. These opinions may not be relied upon as investment advice or as an offer for any particular security.

Any companies and/or case studies referenced herein are used solely for illustrative purposes; any investment may or may not be currently held by any portfolio advised by Franklin Templeton. The information provided is not a recommendation or individual investment advice for any particular security, strategy, or investment product and is not an indication of the trading intent of any Franklin Templeton managed portfolio. Past performance does not guarantee future results.