Google Life Sciences Exodus
statnews.com
statnews.com
This reminded me of the Bloomberg article about Boston Dynamics: (1)
But behind the scenes a more pedestrian drama was playing out. Executives at Google parent Alphabet Inc., absorbed with making sure all the various companies under its corporate umbrella have plans to generate real revenue, concluded that Boston Dynamics isn’t likely to produce a marketable product in the next few years and have put the unit up for sale
Is it safe to say that despite Google's founders' expansive visions of the future, the company is not really interested in expensive "moonshots" that might take many years to realize? It sounds like the robotics and medical innovation labs are really more focused on inventions that will bring in cash in the next few years.
1. http://www.bloomberg.com/news/articles/2016-03-17/google-is-...
Theere's plenty of waste in the government, but you'd be hard-pressed to find the typical types of government waste or incompetence in for example the NSF. Unless you believe basic science is wasteful I guess.
"Galileo Galilei most humbly prostrates himself before Your Highness ... to write of having decided to present to Your Highness a telescope that will be a great help in maritime and land enterprises. ... This telescope has the advantage of discovering the ships of the enemy two hours before they can be seen with the natural vision and to distinguish the number and quality of the ships and to judge their strength ..."
Much of the most impactful research has taken decades to play out and the original research didn't necessarily have medical goals, but turned out to be necessary to making progress.
Some companies have high enough profit margins to dedicate large budgets to research with no immediate applications. Genentech is the first one that comes to mind, but there are others. It takes a great deal of courage to fund this sort of research when you are a CSO or CEO because of constant calls for immediate profits.
I've seen this play out for so long- in biology, in physics, and IT- that I've become inured to the idea that most good ideas take decades to play out until the profits arrive.
I am not entirely certain whether it's correct to say Pasteur was working for the government. When a scientist reaches his level of stature, she pays her own ticket.
The extreme irony of you writing this comment on the Internet should not go unnoticed. Are you really unaware of the extent to which the government underpins all of the interesting long range research in the country?
AT&T and ITU and others never would have been able to do this with the tech they promoted- we'd all be using 56Kbps modems to call BBSes today if they had won.
This is such a ridiculous statement I can't tell if it's sarcasm (from the post I guess not - but that's the kind og argument I'd make if I wanted to make a strawman argument on this), extreeme hyperbole or you really beleive it ?
Quantum Physics was more or less a governmental sponsored project. The Internet was a governmental sponsored project.
Knowledge for the sake of knowledge is the ultimate risk taking and governments not private institutions should perform this as government are the ones who benefit the most.
Edit: ok, maybe not _every_ way. Pipes are "inspired by an idea from Multics". https://www.bell-labs.com/usr/dmr/www/hist.html
Many of the things that Multics had eventually made their way back into various UNIX derivatives, because they were great features that a modern OS wants to have. Multics was way ahead of its time in many ways.
Interesting. Can you name some?
The point is not that government is better at innovating than private organizations it's that they are better suited for actual research in things that doesn't have a market yet.
Private companies on the other hand are much better at taking knowledge which can be applied to the market to then run with it. But up until that point I think it's fair to say that we would want to sponsor research into basic science for the good of the whole society.
Most major discoveries are accidents, it's those accidents which leads to market application.
Instead history shows us that you can't find something by looking for it. I.e. the most important research we have done haven't had a market to begin with but rather have let the pursue of knowledge be the primary motivator.
This was not the result of free market competition.
2. Suppose you don't care about providing steady scientific employment for individual scientists but do value scientific progress (I find this position contradictory to the point of absurdity, but let's roll with it). Even in this case, cyclical funding cuts are still a source of waste:
a) society essentially wasted money subsidizing the education of people who are now not using their training (some might be, but probably many aren't).
b) funding freezes are often temporary, but by the time spending rebounds all those scientists you spent billions of dollars training have moved on to indsutry (research or no). So now you have to double down and train a new generation in order to make progress.
It's not clear to me whether it's bad for academic scientists to leave for industry- that includes me and a number of people I have recruited from academia to work in industry. It's worked well for me, but it was very unpleasant for my cohort to expect to become professors, when they couldn't, due to the NIH budget freeze.
But were they all intending to do that in the first place?
Do you think most people who go to graduate school are intending to become academics? Lots of people say that, but it's not what I see in PhD students - many want to go into industry from the start.
I don't know a single person in my program who actually said they wanted to work in industry (other than the company founders).
Perhaps it's different in other fields.
For the engineering PhD programs (chemical, biomedical, electrical, and mechanical) I'm familiar with at (what was once termed) Carnegie R1 universities, the split ends up being around 50/50 for academia vs industry track after graduation. For bio at the same schools, I saw the same as you, near unity for academia. Of my engineering PhD cohort, I was the only 1 of 14 who founded a startup.
Furthermore, industry is notoriously tight with the results it produces; even if, internally, it's producing good science, the fact that this science isn't being broadly shared with other people is a huge drag on the pace of progress. Academic openness is indispensible to scientific advancements.
The goal of Verily is to "tackle big mysteries of disease prevention and health". That's rather vague and they even admit that they will "evolve the mission" over time.
Bell labs was perhaps the closest thing to having a bunch of smart people getting together and solving interesting problems-- unfortunately, today, in the era where results-oriented project managers run the show, magical places like Bell labs are a thing of the past.
To be fair, also in an era when Ma Bell doesn't have a virtual monopoly over telecommunications in the United States--may have worked for the time but really wasn't such a wonderful thing. Bell Labs did do a lot of wonderful things but they resulted in part from the fact that they were funded by an organization which was essentially insulated from market competition.
When I learned corporate finance it was pretty much a given that a conglomerate should trade at a discount.
Now Google is somewhat insulated from share pressures because they have two classes of stock. This makes the owners think they can use it as their personal playground.
But Google will eventually succumb to shareholder pressure. Why? Because so many employees are soooo heavily invested in Google stock. So even if Larry and Sergey want to do all this cool stuff their executive staff is incented to follow the market. And not only the executive staff but employees at almost all levels. Just so much comp is tied to share prices.
Just think if you were a Google exec with your net worth tied up in Google stock and you're charged with making strategic decisions about where to invest Google's capital budget, you're going to be pretty tempted to direct it towards revenue generating opportunities over and above "science fair" projects. Sure, you'll do your best to kowtow to the CEO's personal projects, but ultimately you and your fellow execs are going to behave more like hedge fund managers. And I wouldn't blame anyone who does.
As you allude, Google's ability to deliver depends on the share price. While they are free from activist shareholders, they aren't free from competition for talent. A rising stock price keeps employees happy. Employees with lots of out of the money options get jealous when their friends experience IPOs.
The OPs point still stands regardless; unless for whatever reason you think you'll get 1) a windfall of stock grants that are 2) calculate from some actual monetary amount divided by the current stock price, you want the value of your stock to go up.
This is in contrast to actually purchasing a stock on the market.
So when the rsu has vested, it is converted to normal stock, where you can sell on the market for cash?
With the smaller Google-specific "well but" limit that insider trading rules restrict employees from trading in the company stock except for narrow trading windows after the company's quarterly reports.
Starting up interesting long-term projects and then canning them at the first sign of a cloudy quarterly forecast is not how you compete for top talent. It just makes the entire company look spastic and irresolute.
Google/Alphabet is in a good position to build the next Bell Labs. Instead, they're building the next Lucent. That's a bummer.
I think most Google employees would fall into either (1) those who don't have enough stocks to care much about stock price, and (2) those who have too much money to care.
Boston Robotics and Nest were both heavily characterized as having issues working with Google proper. And both Nest and now Verily have both been recently characterized as having issues with leadership.
Apparently Tony Fadell told the Dropcam founder that he hadn't 'earned the right' to report to him.
It seems Alphabet companies have a lot of trouble brewing these days.
Google's secretiveness is hurting them. They don't get public feedback when their management screws up. This allows screwups to continue longer. If their R&D people were presenting papers at conferences and filing patents, as AT&T, IBM, and Xerox did in the days when they had big labs, this problem would have been caught earlier.
This is a downside of trying to protect intellectual property by secrecy, rather than patents. You lose external feedback.
That said, in science bad management is kind of the norm. Many big labs in academia are horribly run yet consistently publish huge results. In industry the counterbalance is the ability to easily move on to a different employer, which seems to be what's happening. If true, the bad management should correct itself. One hopes.
Full disclosure: I work at a company that could be seen as in the same area as Verily.
There is a reason pharma and biotech companies are slow lumbering giants. It's so they can absorb the impact of R&D failures.
The fact is, it doesn't work that way in biology. When developing new biotech, you'd be lucky if you understood 5% of what a new drug does in the body (sometimes it's 0%!). Serendipity plays a huge part in new drug discovery, although most companies don't like to admit it.
Also, I'm not surprised that Google has matured to the point where "moonshots" don't work so well. You see it in biopharma as well. Huge companies, with billions of dollars of R&D spend fail to produce new innovation on a regular basis. It's the small biotech companies that seem to come up with the innovations, often on a much smaller budget.
The reasons for that are multiple, but it does come down to a big company worrying about how they will grow their $100 billion company by 10% (finding $10B in business isn't easy) and the fact that with large companies comes structure, process and other things that stifle innovation.
I went from academia to industry to learn how to do what Andy said: make biotech copy tech. In many ways, I was extremely successful: I built Exacycle, which took Google's style of warehouse computing and made it available to a number of biology researchers. The scientists who were successful with it basically had to drop a lot of the preconceived notions about how to solve problems- they were accustomed to spending a huge amount of time setting up their simulations or drug docking (months or years of prep work before running sims for a year). Exacycle converted this to a rapid cycle where they could test an idea in a day on 600+Kcores/sec, make a few tweaks, and run again. This greatly sped up the productivity of scientists.
Ultimately, these ideas weren't well received in biotech until we published a bombshell paper on GPCRs that spoke to them in their language. Even then, it wasn't a product that I think any biotech would want to buy (this is the part where Andy was spot-on). I think Google/Verily paid attention to what we had to say and hopefully, they'll be able to succeed in ways that biotech does not.
At this point, though, I think the nail is in the coffin for rational drug design as a multibillion dollar investment for pharma. It's just not productive enough. Instead, I think pharma needs to solve the inverse problem: for every drug that is already approved for something, what diseases can it impact (the RDD problem is "for a given disease, find a drug that modulates the disease").
The fact is, it doesn't work that way in biology. When
developing new biotech, you'd be lucky if you understood
5% of what a new drug does in the body (sometimes it's 0%!).
What’s rather alarming is the sequence of this current
minimal cell: of those 473 genes, 149 of them are of
unknown function.
The drug made it all the way to market until some folks realized, it doesn't work via GABA at all. It actually impacts the glutamate neurotransmitter.
So much for rational drug design!
that's how you get Theranos.
Nothing wrong with that personality type it's just totally not a fit with Google or frankly with managing any sort of technical people.
this is the only way to deal with a perpetuum mobile of a money printer they've got in house and the resulting 70B+ cash pile problem, and to avoid it growing into 200B problem that Apple is battling (when did you hear anything new from Apple? - yep, this is how left unchecked small issue grows info a big problem). So, either moonshots (probably a lot of them and god forbid one turns into a money printer too) or pay the dividend like IBM.
Well disrupting markets is cool, disrupting people not always.
From experience, usually managers who like to disrupt have some insecurity about them (they know they lack technical knowledge or other personal things) or they are under high pressure, which translates to going around disrupting things because it feels to them and to those above them that work is being done -- "look at all this work, we've re-organized practices 2x this year, switched to agile and so on isn't that great!" kind of stuff.
But I've never been at a place where turning the wheel (a -> b -> c -> a) wasn't a big part of what management did.
He tried it with Andy/Android, Vic/G+, and Conrad/Verily, among lots of other places. It doesn't work very well, but it looks like it has great 'velocity' so he's happy.
Someone has already made the comparison to boston dynamics, but doesn't this also sound just like the recent problems within dropcam/nest?
Does anybody have a counterexample of somebody with this personality leading a team to greatness?
It feels like a personality type like this is mentioned more often than not when my technical friends leave companies. Somebody has to frame the memo that guys like this (they're always guys) will kill your org.
Marissa Mayer might qualify.
I don't think it's a guy thing - I think it's a corporate culture thing.
You get failure and disaster whenever you have overpaid and undertalented management. Gender isn't particularly relevant.
You especially get failure and disaster when upper management runs a corporation as its own personal cash cow and makes decisions for reasons of self-promotion, self-image, or plain greed, instead of for sound growth-oriented reasons - never mind employee welfare.
In short, he wants “to defeat Mother Nature,” Conrad told STAT in a brief interview last fall when the Verily name was unveiled.
I mean, welcome to every doctor and bio researcher ever. Bio is hard, very hard. Think of it this way, your code is actually alive and trying to kill you back. Doyah' think that might increase the complexity and timelines for debugging a bit, maybe?
* And its researchers have published just one scholarly article — about a spoon for people with hand tremors.*
I actually saw some of this too. It's a neat idea, The Bodine lab at CU Denver does some cool stuff with this too, using rubber-bands and not electronics that run out of batteries. It's cool stuff, really, but its a 'if all you have is a hammer' myopic outlook for the problem. I concede that the leapfrogging would be great from there.
A new car has up to 400 different sensors. You know the oil pressure. You know how much air’s in your tires. But we don’t do that for people,” Conrad says on a promotional video posted on Verily’s website. “Instead of episodic, reactive health care, we should provide preventative and proactive health care,” personalized to the individual.
Let's ignore the privacy issues that places like 23andme are dealing with when it comes to this level of detail. The cost alone and the time needed to validate these things are non trivial. If you have, I dunno, black skin or freckles, how much more complex did you just make the problem for an optical sensor? Yes, I can hear you already, you start putting in a case struct in your code, blammo, done. Ok, now for every single sense you add, you have to do this all over. Then you have to train this on the staggering variety of healthy human geometry. The time and more importantly, the power, needed to do this increases factorially (I think) with each new sensor paradigm. Oh, and then you sweat a little bit or go outside. In short, this is really hard.
Verily says it has two wristbands in the works: one to read diagnostic nanoparticles that would detect cancer or other ailments at very early stages, another to continuously monitor skin temperature, pulse, and other heart activity.
Ok, this is a pop-press article and I will give a pass to the band that is essentially a fit-bit. Journalists are low-paid compared to their hours and utility to society and, as such, can make very big differences seem mundane. However, the cancer screening band sounds too good to be true. As such, it likely is. I admit, I have no idea what this thing is or how well it works, but there is a significant test it needs to pass: is it better than a dog? (https://en.wikipedia.org/wiki/Canine_cancer_detection)
Ok, look, I make it seem like these guys are fools. They are not, they are all very smart and very hard working people that want to see the world as a better place. However, the SV bubble that includes YC and their ventures into biotech, is just foolish here. \
Bio is Hard, full stop.
Bio wastes colossal amounts of cash in failed ventures, more so than in tech, though I have no data to back this up. Bio is not something that you 'hack', it hacks you (see any disease or disability). I get it, VCs see the profit margins on Viagra and their mouths get wet. Here is what you need to know though, in biotech the road to success is very very difficult, but once you get 'there' you essentially have a monopoly due to that barrier to entry. But don't confuse that monopoly with the the reality that many grad students and VCs imploded in failed ventures too.
Mad respect to those folks working under this guy, but SV needs to understand that debugging yourself is neigh impossible.
There may be more synergy than you think.