Venture Fund a16z Backs GenML to Combat Eroom’s Law
To improve your local-language experience, sometimes we employ an auto-translation plugin. Please note auto-translation may not be accurate, so read original article for precise information.
GenML, an investment thesis by venture fund a16z, aims to reverse Eroom’s Law by combining algorithms and compute power in life sciences.
AI has the potential to tackle challenges in healthcare and drug design, such as rising costs due to the need for highly trained staff and addressing issues of access and quality.
Key factors supporting GenML’s potential include GPT-4, AlphaFold, and RNA therapy projects.
The tech industry has long been familiar with Moore’s Law, which states that the computing power of computers grows exponentially while the cost of computing decreases. However, there is another law, less known but equally impactful, called Eroom’s Law. This law describes how the rate of innovation in an industry slows down each year, accompanied by an exponential increase in the cost of new products. One particular field where Eroom’s Law has made its presence felt is the development of new drugs.
To move from Eroom’s law to Moore’s law, human-driven services must be converted into compute. This transformation begins with simpler, one-time models (typically machine learning) that perform simple, error-tolerant tasks, such as Netflix using AI to recommend shows. As AI advances, we are entering new realms of possibility, such as generative AI methods producing text and images or completing complex tasks with errors (aka hallucinations). This progression opens the door to the possibility of AI-powered co-pilots in life sciences and healthcare that can greatly scale skilled labor or uplevel less-skilled labor.
The incredible progress of AI is only part of the story; there is also a renaissance in algorithms and compute power, as well as advances in biology and healthcare. Engineering-driven advances in life sciences have resulted in significant advances in gene editing, cellular biology, stem cells, robotic experiments, and other areas, allowing scientists to manipulate biology in previously unheard-of ways. These advancements have enabled biology at scale as well as with newfound consistency, both of which are essential for connecting with AI. Furthermore, incorporating AI into life science experiments creates a strong feedback loop in which the experiments improve the AI’s predictive power, which in turn improves the experiments.
In an attempt to combat Eroom’s Law, the venture fund a16z has recently published an investment thesis focused on the intersection of AI and Biotech, known as GenML (Genomic Machine Learning). This thesis suggests that GenML has the potential to reverse Eroom’s Law, bringing about a change in the industry and unlocking substantial opportunities for startups and investors.
Underlying all of these advances is an immense amount of computing and data storage, which has only recently become possible. For the first time, a renaissance in algorithms has been married with the pure compute power to test, iterate, and run these programs.
AI has the opportunity to tackle greatest challenges in healthcare and drug design. First, the cost of healthcare is rising due to the need for highly trained staff, particularly PhDs, MDs, nurses, and others. As AI becomes increasingly able to function as a technical expert, there are opportunities to extend the abilities of existing providers to deliver care at a much lower cost. If implemented with empathy, it can engender engagement and maintain compliance with clinical recommendations, as well as mitigate clinician burnout. Second, with reduced cost comes the ability to address issues of access (scale) and quality (reduction in variance of performance). As more care becomes AI-enabled, AI has the potential to democratize healthcare, giving the best healthcare services to everyone.
Several key factors support the belief that GenML could break through the barriers imposed by Eroom’s Law:
- GPT-4, a non-specialized model developed by OpenAI, has shown promising results in drug discovery. Even OpenAI acknowledges the potential risks associated with this capability in the GPT-4 model.
- AlphaFold, an AI model developed by DeepMind, recently made headlines by successfully unraveling the complex 3D structures of proteins—a challenge that has confounded scientists for half a century.
- AI-assisted projects in the field of RNA therapy have demonstrated significant potential in finding cures for previously incurable diseases. By harnessing the power of AI, researchers can now explore treatment options that were once unimaginable.
- The success of AI in various domains heavily relies on the quality and scale of the available datasets. Open Data initiatives and the emergence of crowdsourced research datasets are facilitating the expansion of knowledge and enabling more comprehensive AI-driven solutions.
A key part of both the reduction in cost and improvement in outcomes will likely come from AI’s impact in the development of new therapies. AI serves as a key driver in understanding biology, allowing research to be scaled far beyond the current model, which primarily relies on serendipitous discovery enabled by hours of human labor in the lab.
However, it is important to note the potential concerns around AI, including embedded bias and other failures that may arise from training early AI models on data collected by humans. As AI is applied to new industries, scientists, healthcare providers, and regulators must remain vigilant for potentially harmful side effects. The existing regulatory framework in life sciences and healthcare tests everything (therapeutics, devices, etc.) for efficacy and adverse effects.
The New Industrial Revolution is now underway, and while some may expect AI’s impact to occur overnight, we look forward to a gradual transition that will likely occur over time. These developments in GenML offer a glimpse into a future where Eroom’s Law might be overcome, not only in drug development but also in other industries.
Read more about AI:
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.