This was a revolutionary year for technology

However, despite the internet’s claims that the world is in peril, this year has seen incredible technological progress. This month, Moderna’s cancer vaccine (US:MRNA), passed stage two trials, US scientists obtained net positive energy from a nucleus experiment, and OpenAI launched the world’s most advanced artificial Intelligence chatbot.

People have been pessimistic about humanity’s ability for innovation over the past two decades. Peter Thiel, a venture capitalist, famously said that while we wanted flying cars, we got 140 characters on Twitter. While internet platforms dominated the 21st-century stock markets, real-world innovation was slowing down, money flowed into them. Thiel, as one of the first investors in Facebook, both profited and cried at his observation.

The pendulum is swinging back to’real-world innovation’ this year. Market cap has been lost by the internet platform companies that have defined 21st century. Meta’s value dropped 65%. Amazon (US) dropped 48 percent, and Twitter’s value plummeted over 50% before Elon Musk had to make a costly acquisition. As Twitter’s advertising revenue declines, the future of Twitter is uncertain.

Network effects are the foundation of all these platform companies. People use Facebook because it’s where their friends are, and Amazon sellers sell on Amazon because that is where the customers are. These effects were so powerful that people believed they could work, and it attracted a lot of interest-free cash. These network effects, like those in shopping malls, are not holding these companies together. According to Pew Research, over the past seven years, US teens who use Facebook have dropped from 71% to 32% and from 33 to 23% for Twitter.

Government Scaling up Investment

Companies that connect people are not the ones that attract investors’ attention, but rather than those that connect people, it is companies that create things that are attracting investors’ eyes. Moderna’s share prices have risen 48 percent in the last six months following its successful phase 2 trials of its mRNA-cancer vaccine. A patient’s tumor is taken and a vaccine is created to target the cancerous cells. It prevents the return of cancer following an operation.

A week before Moderna’s trial result were published, Microsoft (US.MSFT)-backed OpenAI released its amazing ChatGPT program to the world. ChatGPT, an artificial intelligence program, is known for its large language model. It can write code, blog posts, and even poetry. AI has been a technology that was constantly on the horizon for decades. Although there was progress, it was often made in areas that were not visible to the public like life sciences. It has finally made it into the public consciousness this year for the first. OpenAI was valued recently at $20 billion (PS16 billion).

Limitations to AI’s future development are the number and quality of graphics processing units (GPUs), as well as the energy we can produce. Large language models are trained on GPUs. To process large amounts of information, they use parallel computing instead of linear computing. They were created in the late 1990s and are the hardware that enabled AI. However, to train the models, they require enormous amounts of energy. OpenAI chief executive Sam Altman stated that “the compute cost were eye-watering”.

The cost of AI will drop as energy costs fall. Inflation Reduction Act in the USA has provided $369 billion of subsidies to support green investment and lower energy costs. It is already working. Qcells, a subsidiary of South Korea’s Hanwha Group announced that it would build a $1.8bn state-of-the-art solar panel factory in Georgia. Northvolt, Europe’s largest battery manufacturer, is contemplating delaying the construction of its German factory in order to take advantage of US subsidies. This has upset the EU who is now looking to relax its rules on state aid to allow members to invest more in their green energy sectors.

While free market pioneers might question the distortion caused by state intervention, international competition will lead to faster and better energy innovation. The same competition for state aid is taking place in the chip sector. The US and EU passed multi-billion-dollar legislation to increase domestic production of chips. TW:2330, the largest manufacturer of GPUs in the world, is currently building a $40 billion manufacturing facility in Arizona. Warren Buffett’s Berkshire Hathaway, US:BRK) purchased a $4.1bn stake at TSMC in November. This is further proof that even traditional investors are moving towards hard technology.

Brink of the f ourth revolutionary

Jared Kaplan, OpenAI’s founder, published a paper in 2020 claiming that there are “scaling laws” for AI training. The algorithm will function better if it has more information and parameters. Higher energy production and faster GPUs mean better and more powerful AI. Altman has written a blog post called ‘Moore’s Law for Everything’ in which he describes his vision for an AI-driven world. He believes that AI will be the fourth revolution after industrial, agricultural and information. Altman also believes that “the cost of many types of labor will drop toward zero once sufficient powerful AI joins our workforce”, which will lead to a lower standard of living for all.

Altman might be quick to praise the potential of AI, but a positive circle is already underway. Machine learning is being used by companies to locate the metals required to produce solar panels and lithium-ion battery batteries. According to the International Energy Association, demand for cobalt and nickel will quadruple by 2050. In the last two years, the price of lithium has increased fivefold. This price signal will encourage further investment in ways to mine and find the minerals. The cost of creating and operating electric vehicles as well as AI models will be lower if they are powered by cheaper solar panels and batteries.

AI is constantly looking for ways to increase the efficiency of GPUs. DeepMind (US:GOOGL) is an backed AI laboratory that has developed a more efficient way to multiply matrices together. This will reduce the cost of deep learning models. Nvidia (US NVDA) has already announced plans to release a new GPU codenamed Hopper in 2019, which will have a greater capacity to run large-language models.

Green tech is expected to bridge the gap between the development of nuclear energy and fossil fuels. Rising energy prices have given a boost to investment in nuclear fission. The UK government has committed a PS700mn investment to the Sizewell C nuclear power plant. It is expected that it will be completed in about a decade. After the recent US breakthrough at the Lawrence Livermore National Laboratory, California, nuclear fusion is likely to be more commercially viable.

Some technology fails, and not all technologies are guaranteed to succeed. Rising interest rates have wiped out weaker ventures. Meta has placed $36 billion and its reputation on virtual reality headsets. Meta may still be right. However, its falling share price and the biological need to have physical contact with humans, which has been programmed into us over millions of years, suggests otherwise.

The technology that solved a problem that didn’t exist was crypto currency. Bitcoin and ethereum have fallen more than 60% in the past year. The unsavory icing on an industry in decline was the December arrest of FTX’s Sam Bankman Fried. People start to question where their money is, and financial misbehaviour can be exposed by rising interest rates. This hype cycle is over.

Investors are now more wary of socially engineered Ponzi schemes and network effects. More money will follow government subsidies to hard tech. Companies that are at the forefront of AI and energy will prosper. Altman’s vision for the future, after years of stagnation is now suddenly in view. “The cost intelligence and energy will trend rapidly towards zero,” which are two major inputs to the cost of everything.

Technological development often occurs during periods of disruption. Globalization has driven low inflation and productivity since 2000. This low-hanging fruit is now ripe for technology to reap the benefits.