Person
Daphne Koller
Daphne Koller is a computer scientist and the founder of the drug discovery company insitro, and in three podcasts (Leadership Next, NEJM AI Grand Rounds and The Long Run with Luke Timmerman) she discussed her career and how machine learning can be applied to medicine. She said she finished college at 17, was the first machine learning hire in Stanford's computer science department, and later co-founded Coursera.
She told the podcasts she then joined Calico, left in 2018 and launched insitro, a name she says combines in silico and in vitro. She also spoke about drug failure rates, saying candidates fail more than 90 percent of the time in clinical trials, and about insitro's data, partnerships and programs.
4 episodes · 97 quoted claims
Daphne Koller says calculus gave Newtonian physics a formal way to predict how systems such as rolling balls and pendulums behave.
“calculus was that for Newtonian physics. Whether it's a ball rolling down a hill or a pendulum swinging, we were able to make predictions about what a system will do.”
Leadership Next (Fortune) · 0:06:55
Daphne Koller says the name insitro reflects the integration of in silico (computer-based) work and in vitro (lab-based) work.
“insitro is the synthesis, the deep integration of in silico, which means in silico on the computer, and in vitro, which means in the lab.”
Leadership Next (Fortune) · 0:10:44
Daphne Koller says she felt humbled and unworthy when she received her MacArthur award.
“when I did get that MacArthur Award, I actually felt very humbled and unworthy”
Leadership Next (Fortune) · 0:09:03
Daphne Koller describes her co-founding of Coursera as bringing the web to education in a meaningful way.
“And you can think about what we did in Coursera as the kind of bringing the web in a truly meaningful way to education.”
Leadership Next (Fortune) · 0:10:44
Daphne Koller says her parents were academics and she expected to remain an academic until retirement.
“My parents were academics. I was convinced I would retire as an academic, be a professor emeritus like my father.”
Leadership Next (Fortune) · 0:11:30
Daphne Koller says she got into computers at age 12 while her father was on sabbatical in the United States.
“I got into computers actually when we were in sabbatical, when my father was here on sabbatical in the United States. And I was 12 years old.”
Leadership Next (Fortune) · 0:12:42
Daphne Koller says the machine learning revolution began in 2012 while she was at Coursera, and that by 2016 she noticed it was not yet making a big impact in life sciences.
“because I'd actually been at Coursera when the machine learning revolution in 2012 started, in 2016, I lifted my head over the trenches.”
Leadership Next (Fortune) · 0:15:12
Daphne Koller says she spent time at Calico after Coursera, where she first learned how drug discovery works.
“I spent some time at Calico, which is a great company and learned about drug discovery for the first time.”
Leadership Next (Fortune) · 0:15:48
Daphne Koller says drug candidates fail more than 90% of the time once they enter clinical trials.
“And then you push that into the clinic and it fails over 90 % of the time.”
Leadership Next (Fortune) · 0:16:19
Daphne Koller says UK Biobank enrolled half a million people who agreed to have their DNA and other trait measurements recorded.
“they got half a million people to agree to have their DNA measured, as well as a whole bunch of other what we call phenotypic measurements, which are various traits.”
Leadership Next (Fortune) · 0:20:43
Daphne Koller says the UK Biobank dataset was very Eurocentric and that governments with more diverse populations could build similar resources.
“access to more diverse populations, because the UK Biobank was very Eurocentric”
Leadership Next (Fortune) · 0:22:28
Daphne Koller says ALS has four approved drugs that extend life by two to three months, and patients typically live three to five years.
“There are four approved drugs. They extend lifespan by two to three months. The patients have a three to five year lifespan.”
Leadership Next (Fortune) · 0:24:59
Daphne Koller says Eli Lilly makes a particular type of molecule for insitro because Lilly has capacity and capabilities insitro lacks.
“We have actually an kind of inverse collaboration with Lilly where they have extra capacity and capabilities in making a particular type of molecules that isn't within our sweet spot, and they're making that for us.”
Leadership Next (Fortune) · 0:29:42
Daphne Koller says regulatory toxicity requirements for a drug candidate take about a year to complete and are hard to speed up.
“there are a certain set of requirements that regulators require, for example, on toxicity that just take as long as it takes a year to do these experiments.”
Leadership Next (Fortune) · 0:27:36
Daphne Koller says roughly 50 to 60 percent of startups in the United States have been founded by immigrants.
“even startups as a whole, something on the order of 50 to 60% have been founded by immigrants.”
Leadership Next (Fortune) · 0:32:57
Daphne Koller says the U.S. Food and Drug Administration has embraced AI much more than in previous years.
“I mean, even the FDA has moved to embrace AI much more now than it's had in the past few years.”
Leadership Next (Fortune) · 0:34:34
Daphne Koller attributes the Bay Area's success partly to Stanford and Berkeley, along with a talent pipeline and available investment money.
“The Bay Area has been the success that it is because it is home to two of the greatest universities in the world, Stanford and Berkeley.”
Leadership Next (Fortune) · 0:36:13
Daphne Koller says insitro's stated values include 'create together' and 'engage with respect'.
“Create together is one of our values. Engage with respect is another one.”
Leadership Next (Fortune) · 0:38:40
Daphne Koller says insitro grew to about 50 people just as the pandemic began, which distracted the company from adding decision and governance processes.
“I will use as an excuse that when we hit 50 was also when the pandemic hit.”
Leadership Next (Fortune) · 0:39:41
Daphne Koller says AI used as a therapy aid has no emotions or consciousness and works like a sophisticated mirror for the user.
“There's no emotions on the other side. There's no sentience. There's no consciousness.”
Leadership Next (Fortune) · 0:14:10
Daphne Koller argues that AI harms should be handled by regulating specific uses rather than the technology itself.
“it's not about regulating the tool. It's about regulating the use case.”
Leadership Next (Fortune) · 0:49:44
Daphne Koller says insitro's 'pipeline through platform' strategy aims to build a repeatable engine that produces multiple drugs, not just one.
“we are committed to something we call pipeline through platform. And that means that we don't just make a drug that makes people better.”
Leadership Next (Fortune) · 0:50:51
Daphne Koller says she started studying computer science at age 12 or 13 and finished her undergraduate degree at 17.
“So when I was 12 or 13, and I completed my undergrad degree at 17.”
NEJM AI Grand Rounds · 0:04:57
Daphne Koller says her master's and PhD work focused on multi-agent systems and game-theoretic models of them.
“a lot of that was about multi-agent systems and game theoretic models for multi-agent systems.”
NEJM AI Grand Rounds · 0:05:56
Daphne Koller says she was the first machine learning hire in Stanford University's computer science department.
“I was the first machine learning hire into Stanford's computer science department.”
NEJM AI Grand Rounds · 0:07:06
Daphne Koller says that at a 1985 AI conference panel, most panelists except Judea Pearl argued numbers and probabilities had no place in AI.
“And most of the panel, with the exception of Judea Pearl, who I consider to be one of the foremost leaders in this new vanguard, all said that you cannot use numbers in AI.”
NEJM AI Grand Rounds · 0:08:01
Daphne Koller did her postdoc with Stuart Russell, who she says favored the shift toward probabilistic methods and machine learning in AI.
“i did my postdoc with stuart russell and i know that he's a very big proponent”
NEJM AI Grand Rounds · 0:09:01
Stuart Russell asked Daphne Koller during her postdoc what she would build from her thesis, and she realized the answer was nothing, which pushed her toward applied work.
“If I gave you a group of really talented undergraduate computer science majors to work with you and you could code together something from your thesis, what would that be?”
NEJM AI Grand Rounds · 0:12:47
Daphne Koller says early microarray data drew her into biology because it let machine learning study gene interactions and their links to phenotypes.
“the first microarray data, where you could actually start to think about gene-gene interactions and relationships between genes and phenotypic consequences.”
NEJM AI Grand Rounds · 0:14:27
Daphne Koller's first biology project was with a tuberculosis expert, applying machine learning to tuberculosis transmission data.
“So the first project that I did was actually with a wonderful colleague who was a tuberculosis expert.”
NEJM AI Grand Rounds · 0:15:04
Daphne Koller says each of the first three Stanford massive open online courses attracted about 100,000 learners within weeks.
“where we had 100 ,000 learners in each of those courses in a matter of weeks”
NEJM AI Grand Rounds · 0:29:57
Daphne Koller left Stanford University around the end of 2011 or early 2012 to work on Coursera, just before the deep learning revolution began.
“I left Stanford at the end of 2011, early 2012, which was just when the machine learning revolution was starting in 2012”
NEJM AI Grand Rounds · 0:31:21
Stanford University capped Daphne Koller's leave of absence at two years, and she resigned her endowed chair to stay at Coursera.
“Stanford basically came and said, well, we have a cap of two years on your leave of absence.”
NEJM AI Grand Rounds · 0:30:57
Daphne Koller stayed at Coursera for about five years in total.
“I stayed at Coursera, I think, for a total of about five years.”
NEJM AI Grand Rounds · 0:31:21
After Coursera, Daphne Koller joined Calico, an Alphabet-affiliated drug discovery company, where she worked with leaders including Art Levinson and Hal Barron.
“it was an incredible opportunity to work with unbelievably talented leaders like Art Levinson and Hal Barron and others.”
NEJM AI Grand Rounds · 0:32:51
Daphne Koller left Calico in February 2018 and launched insitro.
“I ended up leaving Calico in February of 2018 and launching a company in situ”
NEJM AI Grand Rounds · 0:33:56
The name insitro combines 'in silico' (in the computer) and 'in vitro' (in the lab), according to Daphne Koller.
“it's the synthesis of in silico, which means in the computer and in vitro, which means inside the lab”
NEJM AI Grand Rounds · 0:33:56
Daphne Koller says insitro started as a one-person company funded with about $100 million from investors who had been seeking a machine learning drug discovery company.
“they said, yeah, here's $100 million, do something.”
NEJM AI Grand Rounds · 0:34:37
Daphne Koller attributes the growth in AI companies working on large-molecule protein and antibody drugs to the success of AlphaFold and its successors.
“driven by the successes of AlphaFold and its follow-on successors that allow us to design proteins very effectively.”
NEJM AI Grand Rounds · 0:39:10
Daphne Koller says insitro relies heavily on unsupervised and self-supervised learning because there are too few successful drugs to use as training examples.
“We also have a much greater reliance on unsupervised and self-supervised machine learning algorithms where the need for supervised examples is much, much lower.”
NEJM AI Grand Rounds · 0:40:44