Report: Smart contact lens and other ambitious projects at Verily are floundering

https://www.statnews.com/2016/06/06/google-star-trek-fiction/

by ravichandran_s • 10 years ago

71 58 10 years ago

As an IT person with only a degree in CS and historically little interest except for IT and maybe some engineering topics the best time I invested in decades was to start - completely by accident - on the medical track.

Hundreds of lecture hours and a plethora of courses in chemistry, org. chemistry, drug development (chemistry), statistics (with focus on medical topics), anatomy, physiology, (some) bio-chemistry, and neuroscience later I have learned sooooo much.

All simplistic/optimistic (or in the case of antibiotics pessimistic) mass-media headlines are almost complete bullshit. In that field you MUST look at the details. Which of course require some knowledge to understand - but actually not all that much, just some basics. I've also met quite a few doctors by now, some of them in research, and see a common path from youthful optimism to disillusionment. Not as in "giving up", just getting rid of baseless optimism about what is achievable and with what effort, and resetting goals (to much lower levels).

For an easy example, how many IT people think the brain works like a computer? That it's "binary based"?

For illustration, here are two serious (but fun) papers written by a biologist and by a neuroscientist, respectively:

"Can a biologist fix a radio?—Or, what I learned while studying apoptosis" - http://www.cell.com/cancer-cell/fulltext/S1535-6108(02)00133...

"Could a neuroscientist understand a microprocessor?" - http://biorxiv.org/content/early/2016/05/26/055624

I think engineers benefit greatly from studying some (freshman university level) biology, bio-chemistry, physiology. In engineering everything is so much simpler, so one tends to try and use ones experience in ones own field when looking at those fields. In reality it's a great mess. If you had to debug source code written by evolution over more than a billion years you'd have no chance - no chance at all - with any of the methods you used. Even the most disgusting project messed up over decades by rotating hoards of unqualified programmers who in addition all couldn't care less and managers who magnified the mess could not remotely create a software project as messed up.

Imagine Microsoft Office Suite source code run through several obfuscation tools. On top of that, the test suites only test an incredibly tiny amount of potential inputs (the amount of "inputs" where a biological system fails exceeds the ones when it works by several orders of magnitude) - and they go "green" as long as the software starts and runs at all for some very small amount of time, even if it runs badly. An it's still easier to debug than even a tiny biological system.

paulmd• 10 years ago

What exactly is sensationalized about antibiotic headlines? My understanding is that they are indeed massively under-funded in research and development while also being massively abused by the livestock industry (particularly overseas, where drugs-of-last-resort are being used on livestock) and general practitioners (although a tiny factor in comparison). I will grant you that the US is trying to take steps to crack down on livestock use, at least.

I don't actually know any IT people who think the brain works like a computer, or that it's "binary based". I know people who think that neural nets attempt to approximate the neural activity of the brain, and also that it's a pale imitation at best, with many modes of action not properly modeled. I know people who think that free will is a myth and that everything we think is just chemical and electrical reactions in the brain, which is likely correct.

LionessLover2• 10 years ago

Honestly, your reply doesn't appear to be meant to be constructive. I just have to casually scan major news sites for headlines and what is the first one I find when asking google for two search words (not the first search result but the first one relevant to the topic - I only used two words, as I said):

BBC, 2015:

Analysis: Antibiotic apocalypse
A terrible future could be on the horizon, a future which rips one of the greatest tools of medicine out of the hands of doctors.

Note the "apocalypse" and the "terrible" future. Of course, not sensationalized at all...

As for part 2 of your comment all you have to do is read comment sections.

Yes, those who ask are told that is not so, but how many ask?

Even "scientists": "The scientist planning to upload his brain to a COMPUTER: Research could allow us to inhabit virtual worlds and 'live forever'" - http://www.dailymail.co.uk/sciencetech/article-2879803/The-s...

"The Singularity Is Near: Mind Uploading by 2045?" - http://www.livescience.com/37499-immortality-by-2045-confere...

I'll stop here. And that's from people who actually thought about that subject, in comment sections you find replies from normal people who don't make headlines.

Remember the excited comments when the Human Brain Project (HBP) was announced? I just looked at some comments on this very website when the announcement was being discussed... Or a random brain related Slashdot headline: "Why We Should Build a Supercomputer Replica of the Human Brain" (https://slashdot.org/story/13/05/15/1926252/why-we-should-bu...) - actually another person who even looks at the topic a lot and still comes up with such ideas (never mind the comments below).

return03• 10 years ago

Kording is actually a data scientist (thus an "IT person"). His paper is a fun read, but not to be taken too seriously. It doesn't point out what is hard about understanding the brain.

mattfishburn• 10 years ago

If you want to learn more about the history of non-invasive glucose monitoring mentioned in the article, I recommend "The Pursuit of Noninvasive Glucose: Hunting the Deceitful Turkey" by Smith:

[pdf] http://www.mendosa.com/The%20Pursuit%20of%20Noninvsive%20Glu...

T_Fri• 10 years ago

I think that instead of negatively reinforcing being "quixotic"(overly idealistic) and stumbling when taking on a complex project, let's positively reinforce the fact that they had the fucking balls to try.

In future, people will be able to use this project (including it's shortcomings) as a learning tool.

Kudos to the team for biting off more than they could chew, for now we all have a better of idea of what size bite to go for next time.

rtpg• 10 years ago

I agree with the sentiment, but still it is a bit worrying when we see so many hyped startups where a lot of scientists are like "yeah this won't work at all."

I mean I know that there are doubters for good ideas too but what's the line to draw in terms of skepticism?

CPLX• 10 years ago

I've never really understood this logic, primarily because as a theory it's not really falsifiable. If the concept could be applied to anything then it has lost all normative powers.

Should we positively reinforce people who jump of cliffs convinced they have a novel scientific theory of gravity that will allow them to fly? Presumably not. Then presumably we should take a similar view of those who attempt things that are similarly scientifically impossible, even if said impossibility is less obvious to a lay observer.

embwbam• 10 years ago

Maybe the first person to do so, sure. One of the reasons you know that never works (besides physics :) is that people have tried it.

WWKong• 10 years ago

Think science vs engineering. The scientific method is to make an observation, hypothesis,validation through experimentation. Even if you prove your hypothesis false it is a win for science. These bleeding edge projects are as much about science as engineering.

w1ntermute2• 10 years ago

The reason they were able to try these inane ideas wasn't because they had the "balls." There are plenty of foolish, arrogant, and naïve people out there who think they know better than the experts on any given topic you might care to name. The reason they were able to do this was because Google was showering the proceeds of their search advertising down on them.

etendue• 10 years ago

I view the article differently (although I agree it is a learning tool): I see it as showing yet another example of what I think of as Silicon Valley hubris.

I've spent about 15 years in research on colloidal nanoparticles, much of that for biological applications of magnetic and fluorescent nanoparticles. There are countless barriers to creating the proposed "tricorder", each significant enough that overcoming even a single one would make one a "made man" in academia, e.g., "solving" RES uptake.

Look at this statement from the article, for one example of what I see as hubris,

    "The particles were so safe, [Conrad] said, that animal testing could be skipped."

This statement contradicts advisory statements by the FDA, guidelines issued by the FDA, the general understanding of the frameworks put forward by the NNI (National Nanotechnolgy Initiative), general trends in the field, and (regulatory) common sense. This "tricorder" would be a combination of a hardware diagnostic platform, a pharmaceutical contrast agent, and some sort of software platform: it doesn't get any more complicated than that. It would take months for the FDA to figure out how to regulate this device. I am completely unable to conceive how anyone with any experience in this area could think that animal testing could be skipped.

Also, your statement that we should "positively reinforce the fact that they had the fucking balls to try" reads to me as if you are unaware that there are many of us in the biomedical world trying, very hard, fighting over scraps of funding and trying to move the ball down the field. We just aren't in the public eye or consciousness. It hurts to see a group like Verily get celebrated for trying when the rest of us get ignored (or worse, dismissed) when we're working on some pretty challenging stuff under some lousy conditions.

petra3• 10 years ago

Another point with the hubris - why go the most complicated thing - continuous, in-body sensing , instead of half/quarter-yearly blood testing ? Sure it's somewhat better, but not that big of a deal, especially since you lose a lot in accuracy, cost, complexity, etc.

It's also actually something we see in the story of theranos, and it's quite puzzling - as almost as if you need to sell the project to the public and investors the same way you sell some mobile game.

gist• 10 years ago

I am glad that they did this but I don't think this rises to the level of "fucking balls to try". What is the down side of doing this with other people's money?

nl• 10 years ago

Well I'm thankful that Google is prepared to take big bets in health science, even when they don't work. Much better to spray money at this problem than - say - Nest.

Hopefully these are some of Edison's 9999 lightbulbs that didn't work.

dpcx• 10 years ago

"even when they don't work".

While they may not "work" (read: they may not have the outcome that was expected from the beginning), they still "work" in that they help Google (and others) learn how to improve other things later.

It may not be the most fiscally responsible thing to do, but learning isn't cheap when you're trying to solve the problems they're trying to solve.

PLenz2• 10 years ago

I'm beginning to think that Alphabet (read: Google) just isn't very good at things that aren't related to advertising. There have just been so many failures (Verily, Nest, Glasses, etc.) as they have tried to get away from their core business - it almost looks like IBM back in the 90's as they tried (and mostly floundered at) becoming a services company over a mainframes company.

You can argue that they are trying 1000 things and 999 can fail as long as that 1 succeeds - but I don't think this is generating a good public view of them.

jacquesm2• 10 years ago

Just wait until they get into building cars and then abandon the project after a few thousand have been delivered to consumers.

The whole attitude that goes with software-as-a-service development does not go well with developing hardware. Many companies have tried, IBM, as you quote went the other way and made plenty of mistakes. Google (sorry, alphabet) is still seeing this as diversifying but that doesn't change much, I don't see how they will pull this off if they can't do a thing like 'glass' without damaging their reputation.

petra3• 10 years ago

With regards to the self driving car - they seem to take a very serious, risk minimizing approach, far from software-as-a-service.

Just compare their slowness and deliberation vs Elon Musk's carelessness and speed with recent stuff.

unprepare• 10 years ago

has their car been a failure? arguably one of the best economic upsides is being the first one to have a fully road-ready autonomous driving system.

chromebooks seem to be gaining in popularity (though that may be driven by shortsighted education purchases like the early iPad)

their broadband services seem to be incredibly popular in the areas they are servicing. (this is just anecdotal, i havent seen any hard numbers)

But i agree, they have had some high-profile failures in recent history, and its possible thats a sign of the organization growing to large/complex/whatever to be successful at exploring new product categories.

ocdtrekkie2• 10 years ago

Their car will be a failure/footnote. Car manufacturers are already doing far more interesting things than Google is in the space. While Google is obsessed with replacing human drivers, and telling people to just trust the car, companies like Volkswagen have worked on self-driving UI that works with the user, telling them before the car makes a maneuver and such. While Google hasn't tested in snow, other car makers (like Ford) have.

Google's really only power in the self-driving car space is that Google is rolling out Apple-level marketing on the topic. People associate "self-driving car" with "Google" by default, because that's where all the media hype is.

But when it comes down to market time, car manufacturers know how to make cars. They know how to make vehicle interfaces people are comfortable and familiar with. They know how to built reliable, relatively bug-free products that don't need a software update every week to work correctly.

tjl• 10 years ago

It's not really road-ready. Right now it requires a very precisely mapped set of roads. Plus, they've only been testing it in a state with very good weather. I'm very curious to see how it behaves in a harsh winter. Even then, there's the pretty public mess up with the bus a few months back.

return03• 10 years ago

Being the first one doesnt guarantee they 'll be the successful one! It's the last one that always gets the market.

asimuvPR4• 10 years ago

You can argue that they are trying 1000 things and 999 can fail as long as that 1 succeeds - but I don't think this is generating a good public view of them.

That is a real problem. Organizations (not exclusive to Google) seem to quickly jump out and announce something before anything real is in their hands. Anytime you announce something a countdown starts in the minds of people. I learned this the hard way, of course.

cmelbye• 10 years ago

Google seems to have a hard time with this. It seems like a majority of announcements at I/O are half a year away (or more) from being released.

bhewes2• 10 years ago

I think the start of the count down clock is intentional. I have been on the opposite end where with out the countdown clock nothing real happens. And given the fiduciary duties of a publicly traded company talk first, results second is par the course.

Fricken• 10 years ago

To be fair, nobody else has been successful at any of the moonshots Alphabet is going for either. It's not like they're getting trounced by the competition or anything, for the most part there is no competition.

I'm not sure how one would quantify the effect that the Google Glass' faceplant has on Joe Blo's decision to choose between ios or Android, but I'm guessing it's negligible.

With Alphabet's space camp subsidiaries there does seem to be some dissonance. While pursuing far-sighted pie-in-the sky projects is laudable, the expectation that these projects be monetized within visible horizons is a bit naive. Lately Alphabet has been dealing with a reckoning of sorts. They kind of have to decide whether or not they're committed to throwing money at these projects indefinitely.

Larry and Segei have allowed themselves to get swept away by their grand visions for the future.

ocdtrekkie2• 10 years ago

To be fair, Google has also failed at many of its attempts to compete with established players. They haven't made a dent in Amazon for shopping, Facebook for social, etc. I suspect they're strongly looking to create new markets because they're aware of how difficult it is to break into established ones. And Google doesn't seem to be happy settling for "third biggest social network".

adwf3• 10 years ago

Email, maps, phones...

They are tremendously successful in areas beyond search ads, but it's easier to remember the failures.

Also, reducing their successes down to their monetisation strategy isn't exactly fair. Search is the product, ads are the business plan.

I'd say that what Google aren't good at, is anything that can't be easily quantified. Anything where they can't just examine reams of data and A/B test to success. If it involves some sort of intuition or informed design, that is where Google fails.

EDIT: I might also add that future speculative profits are also something that can't be easily quantified. Hence all the myriad project cancellations.

PLenz2• 10 years ago

Email, Maps, Android - those are all in existence to feed data to Google's core product - ads.

return03• 10 years ago

They don't need to be. Search advertising is their golden goose (and i don't think they are trying to get away from it!). The question is whether these failed experiments are informative, if they have contributed a body of knowledge that is proportional to the investment taken. Their car experiments seem to have sparked a bout of incremental improvements that is probably gonna create a wave. As for the rest, i am not sure if there are lessons to be learned here.

djrogers• 10 years ago

I think the biggest problem here is that we have too much media coverage and scrutiny on stuff like this today. 24x7 financial news, 24x7 tech news, and 24x7 financial/tech news!

Here's the deal - these are HUGE bets, with a HUGE risk of failure, and a HUGE reward. As a society we should be focusing on the willingness to take these bets, and allowing them to fail without pointing an laughing like a bunch of school kids at a science fair.

Scientific history is full of failure, and examples of failure leading to discovery and advancement. Let's embrace it and repeat after me - "Failure is Always an Option".

bravo222• 10 years ago

True. However Google PR tells the media about these bets in the first place. It used to be the case that R&D was kept under wraps. I'm sure a number of projects fail at Apple too but we never hear about them because Apple doesn't leak that info to the press.

scholia• 10 years ago

A lot of research comes out of universities, where it gets publicized because people want their research funded.

Google's contact lens project for example, originally came out the University of Washington and Microsoft Research got publicity from it before it switched to Google. In fact, Microsoft Research still features the project: Medical Sensing via a Contact Lens (1)

This kind of thing is not a problem if you present it as blue-sky research. It can become a problem if reporters think you're moving from research to commercialization because then they expect a product launch.

Google's car came out of Stanford University's AI Lab. Google took it over after Stanford's car won $2 million in the 2005 DARPA Grand Challenge competition (2). Again, it was great publicity for Stanford as a winning AI project, but it's a potential disaster for Google if it never turns into a good product.

Most of the major car manufacturers were, of course, doing similar research work, but generally keeping it fairly quiet....

(1) https://www.microsoft.com/en-us/researchconnections/science/...

(2) http://news.stanford.edu/news/2005/october12/stanleyfinish-1...

beambot2• 10 years ago

Apple is a bad example; their secrecy means that humanity learns virtually nothing from most experiments. Most experiments fail -- either abject technical failures or merely commercially unviable at the time of their development. Hiding these away is a shame. People could learn from failures (not spending resources pursuing bad paths) or unviable explorations (revisit them later when conditions change).

I think a more useful example is Bell Labs and their Technical Journal. The journal involved a more-significant level of rigor without the PR and media hype.

Unfortunately, when the media follows every little happening (eg. patent publications)... big megacorps are left with two bad choices: let people draw erroneous, hyperbolic conclusions ("OMG! Ads in my eyes!") -or- address the PR and messaging directly head-on.

Publication may still be the solution, but the timing issues are acute. From a business perspective, you need to be filing IP as soon as you have reasonable evidence of a proof of concept. But by the time you have sufficient material for top-tier journal peer review, those patents are already being published. The solution might be (again) something like the Bell Labs Technical Journal... but that provides a lot of fodder for competitors and patent trolls. So it's a tough balance!

Either way... I applaud large organizations taking big technical risks and bets.

sandworm1013• 10 years ago

It's not the amount, but the quality. How many of the media people at the event had the faintest understanding of biology? The interns that are sent to these briefings don't have the skills to see the snake oil. Send a medical student. Send an engineer. Send a lawyer. Send anyone with experience in innovation within highly-regulated space. And listen to them. Do something more than photocopy the press release onto your "news" feed.

michaelbuckbee• 10 years ago

Completely agree, it seems like there is this consistent drumbeat of "nobody is working on big things", "devs are working on widgets instead of curing cancer" and here you have a group _trying_ to make some impact on core quality of life things for people and the criticism is essentially that it's not as easy (aka "floundering") as knocking out a Facebook app.

ModernMech• 10 years ago

[Conrad] promised a working prototype of the cancer early-detection device within six months. That was three years ago.
Conrad rolled out the Tricorder to tech reporters in 2014, describing its scientific basis as proven. He predicted that high-risk patients would begin wearing the device within a few years, followed by widespread adoption.

What to take huge risks? Fine, but don't promise results are around the corner when they're not. Headlines with the words "arrogance" are completely appropriate here.

realworldview2• 10 years ago

I, for one, am not surprised. Are there tax advantages associated with this type of failure?

realworldview2• 10 years ago

Down vote and no reply to a valid question. What a surprise...

dang• 10 years ago

Please don't comment about getting downvoted. There's more than one HN guideline about this: https://news.ycombinator.com/newsguidelines.html.

In this case I imagine the downvotes were because "I for one am not surprised" sounds glib and dismissive (and kind of braggy) without adding information. The question that follows it is fine, although to make it seem like less of an insinuation you could have added some (neutral) context for why you were asking.

_ihaque• 10 years ago

Stat has covered Verily in detail in the past; this article from March argues that the problem stems from the leadership:

https://www.statnews.com/2016/03/28/google-life-sciences-exo...

  Former employees, however, characterized Conrad in less complimentary tones.
  They said he exaggerates what Verily can deliver, launches big projects on a
  whim, and rashly diverts resources from prior commitments to the next hot 
  idea that might bring in revenue. This has led to what they describe as
  difficult meetings with business partners, and resignations by demoralized 
  engineers and scientists in the face of seemingly impossible demands.

Report: Smart contact lens and other ambitious projects at Verily are floundering

Related Stories