A Tale of Two Exponentials
Use Moore to fight Eroom
Today’s computers are about 1000 times more powerful at the same price than computers from a decade ago. And they are roughly a million times more powerful than they were 20 years ago. This period of dramatic growth from the 1970s to today software that eats the world. Technology is powerful because it rides on Moore’s Law, the ability of the tech industry to dramatically cut costs and improve capabilities over decades.
In contrast, drug design and healthcare delivery Elum’s law, coined from the reverse of “Moore”. These industries have suffered decades of recession. gain of price, which has reached very extreme levels, with healthcare spending reaching about a quarter of US GDP (and rising). Rising labor costs, clinical trial costs, administration costs, increasingly hostile payer-provider relationships, and more are driving the cost of treatment and healthcare on the wrong curve.
Turning services into computing
Given that one is exponentially decreasing and the other is exponentially increasing, the obvious goal is to From Elum’s Law to Moore’s Law. But how is that possible? Human-driven services (i.e., care service delivery) must be replaced by computing (i.e., technology commoditizing services).this is Exactly what we see in AI.
This transformation starts with low-complexity one-off models (usually called machine learning) to perform simple, error-tolerant tasks. For example, Netflix uses AI to recommend shows.
As AI becomes more sophisticated, we are increasingly moving into new categories of possibilities. Generative AI techniques, albeit with mistakes (aka hallucinations), can now not only generate text and images, but also complete complex tasks. For example, chatGPT can generate an English answer to a question, but it fails spectacularly on certain questions, sometimes “hallucinating” a bogus answer.
This advancement, over time, opens up the potential for AI-driven co-pilots for life sciences and healthcare, significantly expanding the skilled workforce or boosting the level of less skilled workers. There is a possibility. For example, AI can suggest answers and ideas, let trained humans choose the best, and handpick results to skip wrong answers. This approach naturally integrates AI into existing workflows.
As time goes on, the proportion of human work will decrease, eventually moving closer to full automation even in areas where human expertise is needed, where even small mistakes can have disastrous consequences, but especially In critical areas it’s probably not possible without a human somewhere in the loop. In particular, mistakes in diagnosis, drug prescribing, and medical procedures are unacceptable. Developing AI that successfully accomplishes these specialized tasks and is not tainted by critical errors is a prominent area for future AI development along the path to AGI, and a natural place for future AI progress. will ultimately have the greatest impact on life sciences and life sciences. health care.
A Renaissance of Algorithms and Computational Power Combined with Advances in Biology and Healthcare
Incredible advances in AI are only part of the story. AI is maturing at a time when life science and healthcare are also transforming, both industries being increasingly driven by engineering. power and opportunity It’s about changing the way we diagnose, treat and manage disease and deliver health.
In the life sciences, advances such as gene editing, cell biology, stem cells, and robotic experiments have allowed scientists to manipulate biology in unprecedented ways. With these advances, biology scale But I found a new Consistencyboth important factor To connect with AI. Furthermore, with the advent of AI, embedded In life science experiments, there is a strong feedback loop in which experiments improve the AI’s predictive ability, resulting in better experiments.
Similarly, healthcare is experiencing a renaissance in the use of technology.With the enormous cost of healthcare weighing heavily on the sector, innovators are I’m hungry for technology This improves outcomes and reduces costs.transition to Value-based payment modelWith active patient and provider engagement paramount, this is also a compelling tailwind for even more of the deep utility of AI in healthcare.
Underlying all of these advances is an enormous amount of computing and data storage that has only recently become possible. For the first time, an algorithmic renaissance was combined with pure computational power to test, iterate, and run these programs.
What it means: Tackle your biggest challenges
Simply put, we have the opportunity to use AI to tackle some of the biggest challenges in healthcare and drug design.
First, medical expenses.The Exponential Rise in Costs is Partially Caused In particular, the cost of skilled labor is increasing far faster than inflation, creating a need for highly trained staff (PhDs, doctors, nurses, etc.). As AI becomes capable of functioning as technical experts, there are opportunities to: develop one’s ability You can use your existing providers to provide care at a much lower cost. Implementing AI with empathy can not only reduce clinician burnout, but also create engagement and maintain adherence to clinical recommendations.
Second, the cost savings help address issues of access (scale) and quality (reduced performance variability). As more healthcare becomes AI-enabled, AI has the potential to democratize healthcare and provide the best healthcare services for all. AI has the ability to expand existing wisdomThis means patients are more likely to receive the correct diagnosis and treatment plan sooner.
Moreover, a significant part of the cost savings and improved outcomes is likely due to the impact of AI in the development of new therapeutics. Here, AI serves as a key driver in our understanding of biology. In the same way that calculus plays a fundamental role in physics, AI will power the unraveling of biology’s complexity, a complexity that certainly exceeds what humans can fully comprehend. We are witnessing AI models of human disease today that pave the way for medicines that are more effective, less likely to fail, and faster to market.So AI can understand biology beyond ability human scientist. This will allow research to scale well beyond the current model, which relies primarily on chance discoveries made possible by hours of human labor in the lab.
Given the above, it is also important to be mindful of potential AI concerns. We recognize the potential for biases and other failures that can arise from training early AI models. data collected by humans. As AI is applied to new industries, scientists, healthcare providers, and regulators must remain vigilant about potentially harmful side effects.
surely, existing Regulatory frameworks in life sciences and healthcare test everything (therapeutic drugs, devices, etc.) for efficacy and side effects. And for those who fear that AI is a black box, we argue that AI is fully explorable, and that any AI can be understood in detail given enough time. Ironically, human reason real black box in healthcare.
A new industrial revolution is now
Clearly, this is not an overnight transition, as healthcare (and biopharmaceuticals) is actually a group of intertwined industries under regulatory oversight. Those who expected the impact of AI to emerge within months may be disappointed or take advantage of the gradual nature of the transition to point out AI failures. Instead, we transition It will take place over perhaps 10-20 years, in a way that all stakeholders are aware of and comfortable with the transition, but at the same time, a large portion of US GDP that has historically been completely unaffected by innovation. changes radically. landscape of technology.
Solving the biggest challenges in healthcare and life sciences requires domain-specific AI, not comprehensive AI that can do everything the average human can do. We envision a series of specialized AI companies designed using specialized large-scale models and built by specialized teams.
Builders should a) use the latest and greatest in AI technology, and (perhaps more importantly) b) use defensible products and go-to-market strategies to develop biopharmaceutical and healthcare products or platforms. You need to understand both how to commercialize As such, teams deep in both disciplines (scientists, AI experts, healthcare builders and operators, product and go-to-market experts) are best equipped to lead and win in this new era. We believe that
What if AI turned every nurse into an inpatient superhero, and every patient had a professionally trained, always-on companion who could talk as much as they wanted for just a few cents an hour? What we are seeing is urging the industry to think about . In terms of therapeutic drugs, we have pursued the development of therapeutic drugs aimed at extending healthy life expectancy. AI-powered research into better antibody therapies to deal with the worst of humanity. Research and development beyond the capabilities of human scientists.
We have been investing in groundbreaking companies like this for years. A complete list of AI and other investments can be found here. And our confidence is growing. A new industrial revolution is upon us and we are excited to play a role in its development.
* * *
The views expressed herein are the individual views of AH Capital Management, LLC (“a16z”) cited and not those of a16z or its affiliates. Certain information contained herein is obtained from third party sources such as portfolio companies of funds managed by a16z. Although it has been obtained from sources believed to be reliable, a16z has not independently verified such information and makes no representations about its perpetual accuracy or suitability for any particular situation. not. Additionally, this content may include third party advertising. a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.
This content is provided for informational purposes only and should not be relied upon as legal, business, investment or tax advice. You should consult your own advisor about these issues. Any reference to securities or digital assets is for illustrative purposes only and does not constitute an investment recommendation or offer of investment advisory services. Further, this content is not directed to or intended for use by any investor or prospective investor and should not be relied upon in any circumstances in deciding to invest in any fund managed by a16z. Do not. (Any offer to invest in an a16z fund is made solely by means of such fund’s private placement memorandum, subscription agreement, and other relevant documents, all of which must be read.) Investments mentioned, referenced, or referred to or any portfolio company, the ones described do not represent all investments in vehicles managed by a16z and that such investment will be profitable or that other investments made in the future may have similar characteristics or results. We cannot guarantee that it will result in A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not given her a16z permission to publish and unannounced investments in listed digital assets) can be found at https://a16z. Available at com/investments. /.
The charts and graphs provided herein are for informational purposes only and should not be relied upon in making investment decisions. Past performance is not indicative of future results. Content is stated only as of the date indicated. The forecasts, estimates, forecasts, goals, outlook and/or opinions expressed in these materials are subject to change without notice and may differ from or contradict opinions expressed by others. . For other important information, please visit https://a16z.com/disclosures.
