The Big Strategic Rotation

In November, about two months before all the selfies from JPM 2024 began flooding the zone, UK Biobank debuted the world's most comprehensive health data source derived from half a million people's whole genome sequences. 

The content it holds is derived from 500,000 volunteers who shared a deep dive into their lives over 20 years. More than 10,000 variables were collected on each, giving capitalists and academics extraordinarily rich data to mine. According to The Guardian’s coverage of its official unveiling on November 29, 2023, more than 30,000 researchers worldwide are registered to access the data. These researchers have published more than 9,000 academic papers. Today, writes the newspaper, “the UK Biobank ranks as the world’s most important health database and is arguably the UK’s most significant scientific asset.” 

UK Biobank positions itself as a public good to enable scientific discoveries and improve human health. Per the website, this abundance of genomic data is unparalleled, but what cements it as a defining moment for the future of healthcare is its use in combination with the existing wealth of data UK Biobank has collected over the past 15 years on lifestyle, whole body imaging scans, health information, and proteins found in the blood. "This is a veritable treasure trove for approved scientists undertaking health research, and I expect it to have transformative results for diagnoses, treatments and cures around the globe,” said Professor Sir Rory Collins FRS FMedSci, Principal Investigator, UK Biobank.

Presumably those health improvements, diagnostics, treatments and cures will be made to happen through markets someone develops within the context of a commercial model. In other words, as a business.

And it's probably a safe bet that Google Deepmind drug-development spin off Isomorphic Labs uses Biobank data for the AI platform that just resulted in a $3 billion deal with two of the largest drug companies in the world, Eli Lilly and Novartis. (Isomorphic Labs uses artificial intelligence to predict biochemical structures, which aids in the discovery and creation of new drugs by recommending which potential compounds will have the desired impact on the body.)

Missing from all the keynotes, investor presentations, VC pitches, podcasts and fireside chats at JP Morgan’s annual healthcare conference and Davos 2024 were strategic conversations around provenance (i.e., the origin of things), the changing nature of competition and the value of ‘data equity’ as an engine for growth and economic development.

The uncomfortable conversation that’s arrived for industry and government is this: how to share the wealth created from all those new health products and markets developed from the data that comes through our engagement with search engines, social media platforms, content, loyalty points and an infinite and expanding galaxy of other digital transactions. It also includes the roughly 137 terabytes of data generated every day when the collective we -- consumers-as-patients; patients-as-consumers; HCPs-as-consumers; everybody-as-technologists -- interact with the entire health apparatus in its current form.

Should those 500,000 people whose data is in UK Biobank get a cut of revenue (sales or taxes) from markets based on their data? 

If you ask the family of Henrietta Lacks, a Black woman who had cells taken from her and used for research without consent more than 70 years ago, the answer is yes. They successfully sued JP Morgan attendee Thermo Fisher Scientific, which earned billions from products and services derived from her cells, for exactly that reason. Terms were not disclosed, but the legal team for the family is promising more.

If you ask the U.S. Department of Health and Human Services, who are using the provenance of drugs in their negotiations with the pharmaceutical industry as part of the IRA, the answer is also yes. Among the factors CMS will consider for “the purposes of negotiating a maximum fair price for each of the selected drugs” include “prior federal financial support for novel therapeutic discovery and development with respect to the selected drug.” In other words, did taxpayers help fund drug development and commercialization with taxpayer health data, perhaps as part of an NIH study (the National Institutes of Health contributed $187 billion for basic or applied research related to the 356 drugs approved 2010–2019, nearly all of which were subsequently launched by biopharmaceutical companies).

Perhaps those taxpayers were also military veterans of the United States — there are more than 16 million, of which around 35,000 are homeless — who were part of a clinical study enabled by the VA’s Million Veteran Program.

According to the VA, the Million Veteran Program is the largest such database in the world. It includes not only genetic information but also is linked to the department’s electronic medical records and even contains records of diet and environmental exposure. The VA says its data are available for now only to V.A. doctors and scientists, most of whom also have academic appointments. These academics, the VA is proud to point out, have published hundreds of studies based on data that has already been collected from veterans.

Which is notable not because any of this scientific content is actually read by real people and turned into some sort of clinical practice improvement to scale better care for veterans. It’s notable because this scientific content is now available for reading by the large language models being developed by Google, Microsoft, Datavant, Epic, Truveta, Meta and many others, who in turn position them as big humanity-saving visions that can be pitched to governments and people with big budget at Davos and JP Morgan. It’s a weird feedback loop that seems to discard, omit or assume-way the one thing these models/markets will soon need more than anything else: imagination from human beings.

The problem facing this bigger-is-better approach to AI is that it’s running out of road — Big Tech may soon reach Peak Data.

“We could run out of data to train AI language programs, forcing researchers to get creative to make training data stretch further,” writes Tammy Xu in MIT Technology Review. In recent years, the trend has been to train these models on more and more data in the hope that it’ll make them more accurate and versatile.

“The trouble is, the types of data typically used for training language models may be used up in the near future—as early as 2026, according to a paper by researchers from Epoch, an AI research and forecasting organization. The issue stems from the fact that, as researchers build more powerful models with greater capabilities, they have to find ever more texts to train them on. Large language model researchers are increasingly concerned that they are going to run out of this sort of data, says Teven Le Scao, a researcher at AI company Hugging Face, who was not involved in Epoch’s work.”

Which is exactly what’s bothering the New York Times.

They are suing ChatGPT-owner OpenAI over claims its copyright was infringed to train the system. The lawsuit, which also names Microsoft as a defendant, says the firms should be held responsible for "billions of dollars" in damages because readers can get New York Times content without paying for it -- meaning it is losing out on subscription revenue as well as advertising clicks from people visiting the website.

In his book “Who Owns the Future?” American computer scientist Jaron Lanier, who was named one of the 100 most influential people in the world by TIME magazine and is now a Microsoft employee, makes the case that people should own their data and be compensated if they choose to share some of it.

“No one disputes that big data can be an essential tool in medicine and public health,” he says. “Information is by definition the raw material of feedback, and therefore of innovation. But there is more than one design for integrating big data into society. Because digital technology is still somewhat novel, it’s possible to succumb to the illusion that there is only one way to design it. Is it conceivable to use big data in such a way that people and their economy get healthier?”

Across the arc of human experience, entirely new economic systems worth hundreds of billions are there for the inventing, yet we're stuck defending the past with obsolete arguments about "value" and “cost”.

Until recently, our global economy was basically powered by two forms of value exchange: the first was based on the exchange of goods and services, and then later, the exchange of attention in the form of media and entertainment. Now we have to add a third construct: data equity. This is data that amplifies, informs and powers the commercialization strategies of entire industries and infrastructures, economic systems and subsystems — this includes $100 billion markets for electronic health records, pharmacy benefit management services, employee benefit consultants, data brokers, drug development and marketing, commercial health insurance, technology services, venture capital, and media and content.

Not to mention government revenue.

Perhaps it’s something 23andMe should consider as a way of making new money — for itself, for its shareholders, and for the millions of people contributing data to the struggling business.

CEO Anne Wojcicki said she’s “open to all ideas” in ad­dress­ing in­vestor con­fu­sion and turn­ing around a dis­ap­point­ing start to the ge­nom­ic pi­o­neer’s life as a public company. Writes Andrew Dunn in his conference wrap for EndPointsNews:

“Even as the JP Morgan Healthcare Conference brought a burst of op­ti­mism to start 2024, stoked by a ral­ly­ing biotech mar­ket and im­prov­ing macro­eco­nom­ic cli­mate, 23andMe’s stock dropped an­oth­er 14% last week. The South San Fran­cis­co-based biotech, best known for sell­ing DNA se­quenc­ing kits to mil­lions of peo­ple and now al­so de­vel­op­ing its own drugs, has seen its stock price fall over 90% since go­ing pub­lic in June 2021. It’s tak­en the com­pa­ny’s val­u­a­tion from $3.5 bil­lion at its mar­ket de­but to rough­ly $307 mil­lion.”

Shares of 23andMe are hovering around $0.71, not far from its all-time low. 

An Uncomfortable Conversation

Sandy Pound is the Vice President and Chief Communications Officer at Thermo Fisher, who presented yesterday at JP Morgan. She sat down for a brief interview with her public relations agency to produce a piece of Q&A content based on her thoughts about “why you should embrace being uncomfortable as a communicator” – two excerpts seems salient here:

What communications challenge keeps you up at night?

“Before the pandemic, the biggest challenges were company issues which required traditional crisis communications. We had a robust approach to handling crises through tested scenarios and a playbook of actions. Since the pandemic, the world has been a very different, and at times, difficult place. The unexpected external events that we might face cause me the most concern.”

What is the best advice you’ve ever gotten?

“If you’re not uncomfortable, you’re not growing. This isn’t exactly what I wanted to hear from my husband any time I share something that makes me uncomfortable, but he is 100% right and I appreciate his honesty.  It’s easy to fall into a pattern and as communicators, I think we need to get comfortable with being uncomfortable.”

Indeed.

As Lanier rightfully questioned — who ultimately owns the data around which very, very large amounts of money are being made? And could we reshape the economics of a digital society where there was a fairer treatment of the value of one’s data?

The thing industry and government should probably start having an uncomfortable discussion around is how to make the invisible hand visible, to understand ‘data equity’ as something that can unlock a whole new form of wealth and economic competition.

Hey Google, are you using any VA data to train your model? If so, can you reinvest some earnings into helping state and local governments care for homeless veterans?

You can either get in front of this emerging story and shape it to your advantage, or sit back and become a victim of it, let it wash over you as a series of undulating waves.

Lost in the cloud of humanity-saving potential of generative artificial intelligence forecasting, visioning, agenda-setting and mission-purposing-life-ambition sold at JPM24 and Davos24 is this: When it comes to competing differently or making large-scale wealth creation happen -- the 'next hundred billion' in growth, say, or entirely new commercial models -- the sequence should be economic innovation first, technology innovation second.

The future belongs to those brands that define the rules by which others have to play. It's about building and operating the most inviting and empowering networks, and winning new subscribers with integration models that deliver what people really want: predictable and unhindered access to the goods and services they, and their data, have had a hand in bringing to market.

Call it the Big Strategic Rotation.

 / jgs

John G. Singer is Executive Director of Blue Spoon Consulting, the global leader in positioning strategy at a system level. Blue Spoon specializes in constructing new industry ecosystems.