Protein-folding, tailor-made drugs and virtual agents create great value long before autopilots are released on the roads

Protein-folding, tailor-made drugs and virtual agents create great value long before autopilots are released on the roads

03 February 2021 from Mikael Syding

Exactly one year ago, the host told MacroVoices that the infection that had spread in China was likely to develop into a serious pandemic. He said the pandemic was the least priced and worst known risk in the history of the financial market. From his perspective, it was clear that the measures taken by China and leaked from the country meant that the infection was much worse than the official picture. There was plenty of supporting data, data which, when subsequently analyzed, were fully sufficient to identify and grade the pandemic and to order reasonable steps to alleviate the effects of the pandemic. Erik Townsend was not alone. There were a lot of investors and politicians who positioned themselves early on for the financial market's upcoming panic reaction. Even an alert and curious person could thus analyze the situation with good results. For an artificial intelligence, an AI system tasked with monitoring pandemics, it would have been even easier and more secure. Unfortunately, the systems are not democratically distributed yet, but in the 2020s, the exponential development in AI will have enormous effects on our lives.

The human brain has an amazing ability to recognize, for example, the category of cats based on a single image or an encounter with a single specimen. It is only in the last five years that neural networks (a special kind of dynamic machine learning where the AI system trains and adapts itself) in combination with extremely large amounts of data have made the machines catch up with us in some crucial areas. This includes the ability to recognize so-called "categories" such as cats in photos, but also to play complicated action strategy games online, the board game Go, or poker. Other areas that are underway are X-ray image analysis, cancer diagnosis and investment decisions. Recently, the company Deep Mind succeeded in solving the otherwise impossible problem of how protein chains fold into their final structure based solely on the molecular sequence of the protein. It bodes well for the development of customized drugs with perfectly tailored features.


Today, it is as easy as it is cheap to develop a service that can give everyone with a smartphone access to a pandemic app that calculates the risk of infection in an area, globally or locally. Unfortunately, the priority has only been to give tips on books, films, news, trends and equity funds over the danger of pandemics. Otherwise, Covid-19 could have been made a small footnote in history instead of as tragic as long shelf meters. Financial companies such as Blackrock and T Rowe have had greater incentives to use the opportunities to serve their customers better than public authorities. No one is surprised, is it? Over time, the AI systems that emerge on business platforms at e.g. Amazon, Facebook, Alibaba, Weibo, NetEase and JOYY will develop into fully-fledged virtual agents who keep track of everything important that happens both in the world and in your own head.

In practice, the agent will be a simulated copy of you and often know better than you what you want to buy, eat, experience in the form of e.g. trips and concerts, or whatever you want to invest in. In many cases, you will gradually let it book and pay for most of your everyday life - restaurant visits, cinema, flights, plumbers, hotels, etc. It can be both faster and easier than you to find out the best combination for you of time, quality, price, appropriate variation or repetition from the last - including taking into account your company and a lot of other factors.

Your digital copy will not least know what events you benefit from or enjoy knowing; including outbreaks of everything from lice to Ebola, winter vomiting and probable or ongoing terrorist attacks; and tell the most important thing to you. You will know exactly what you need before even the news agencies have a clue. If this is important enough, your AI agent may have a mandate to cancel a meeting or even your night's sleep. Sometime during the late 2020s, the dream of self-driving cars may also come true. First on extremely simple city stretches with lots of pre-programmed support points and decision paths. And then gradually in increasingly difficult and complex situations. However, before we have genuine L5 autopilots that master an Indian road junction and also have permission to do so with a sleeping passenger in the car, we are probably well into the 2030s.

And before we get there, semiconductor companies like Nvidia, ASML, Qualcom and Intel, as well as technology conglomerates like Microsoft and IBM, need to raise the underlying hardware a few levels - at least one or a couple more generations. Namely, it is not simply simple to get serial processors and programming languages to fully mimic the human brain's distinctly parallel information processing. New 3D chips and memristors from e.g. IBM certainly bodes well. And advances in quantum computer design at Alphabet, for example, can suddenly lead to even greater quantum leaps in the pursuit of truly competent and human-adapted artificial intelligence systems.

In the professional B2B planning half where companies do business with each other, successful business system companies such as SAP and Salesforce compete with startups such as Workday and ServiceNow to best analyze customers' business data in order to optimize their value-creating processes. Properly handled, there is no defense against a good AI system. It can easily turn trillions of dead data points into strategically crucial business decisions. The greater the amount of data regarding input and end results that the neural networks practice on, the better the action proposals for the customers will be: “Implement a campaign like this for them. Follow up as follows "," This business area must change, e.g. like this or like that ”,“ Issue an instrument like this. Recall these ” etc.

AI systems also create life. The Israeli company Embryonics has taught its AI system to recognize at an early stage, before implantation in the uterus, whether the development of an IVF embryo signals survival or not. The initial studies have given clearly better results than for the human expert panels. This applies in particular to the ability to recognize weak embryos, but also the ability to find strong embryos where the panel members wanted to discard the embryos and rather produce new ones.

The AI development of the 2020s promises a more automated, more efficient and at the same time more individualized and happier world. Advanced hardware, software and intelligent use of these on large social and professional game boards are driving the development of better AI. And the resulting AI systems, in turn, can be used to develop better hardware, software, and user platforms in an ever-accelerating AI spiral.

This is where Vontobel's AI tracker that follows the Solactive Artificial Intelligence index comes in. It is made up of the 14-20 companies in AI-relevant industries that are ranked best in five years of growth, this year's free cash flow and have at least USD 750m in market value. The index is rebalanced every six months so the holdings are equilibrated before each new half-year. This semester started with 20 holdings with 5% weight each, but has now spread apart to at least 3.4% weight for SAP and at most almost exactly double the weight, 6.8%, for ASML (they develop lithographic tools to develop the next generation of semiconductor products).

Solactive Artificial Intelligence Index


@Mikael Syding

Important legal information

Legal notice

This information is in the sole responsibility of the guest author and does not necessarily represent the opinion of Bank Vontobel Europe AG or any other company of the Vontobel Group. The further development of the index or a company as well as its share price depends on a large number of company-, group- and sector-specific as well as economic factors. When forming his investment decision, each investor must take into account the risk of price losses. Please note that investing in these products will not generate ongoing income.

The products are not capitalprotected, in the worst case a total loss of the invested capital is possible. In the event of insolvency of the issuer and the guarantor, the investor bears the risk of a total loss of his investment. In any case, investors should note that past performance and / or analysts' opinions are no adequate indicator of future performance. The performance of the underlyings depends on a variety of economic, entrepreneurial and political factors that should be taken into account in the formation of a market expectation.

02/03/2021 17:36:03


Write a Comment






* Required fields need to be filled in