Hacker News Reader

Comment by ilamont | Hacker News Reader

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 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-...

View on HN

This is completely unsurprising. In fact, I bet you could go back and look at old articles from the business press from decades ago that described similar happenings in older technology conglomerates.

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.

I agree. Let's point out that none of these divestitures or strategic allocation decisions would happen without Larry or Sergey.

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.

Google employees don't get stock options anymore, but "Google Stock Units": https://www.quora.com/What-are-Google-Stock-Units-for-new-hi... So although the stock price rising makes everybody happier, there is no such problem of out of the money stocks.

For the rest of the world, we just call them Reserved Stock Units.

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.

Reserved stock unit? Do you mean restricted stock unit? RSU?

Yes, google GSUs = RSUs.

what exactly is RSU? What's the difference between it, and stock you buy from a broker on the market?

An RSU is like an option with a strike price of $0. You receive the option to puchase stock (with vesting).

This is in contrast to actually purchasing a stock on the market.

> to purchase stock [via] vesting

So when the rsu has vested, it is converted to normal stock, where you can sell on the market for cash?

Yes. It's just stock on a vesting schedule.

> 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.

You can sell during blackout periods if you file a 10b5-1 plan a few months (year?) in advance.

http://www.investopedia.com/terms/r/rule-10b5-1.asp

Yes, thanks. Brainfart.

Interesting. This mitigates the problem somewhat (by making it not binary) but not entirely.

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.

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.

Counterpoint: The majority of Google employees aren't anything near "heavily invested" in Google stock. Sure, they get some, but the amount is dwarfed by their regular salary and bonus. For them, stock is "nice to have", but not a big deal. In fact, I'm pretty sure most of them would rather say "Who cares about stock price? Let's do some cool stuff!"

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.

I'm just a lowly Google engineer, but my stock grant is already way over 50% of my base salary. In other words, the stock is somewhere between 1/3 and 1/2 of my total compensation, based on today's price.

That's what's discouraging. Google seems to have made a mess in mobile robotics, getting most of the good people and then screwing up the management of them. Life sciences is a bigger field, and Google isn't a big player there, so they can't trash the whole field, but this seems to be a similar screwup.

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.

One area where Google does seem to avoid this secrecy is in AI: both Google Brain and DeepMind are consistently publishing papers & in many cases putting them up on Arxiv pre-publication.

Perhaps these are divisions blessed with good leadership?

Well AI can do stuff right now. It can help fight spam and rank webpages. Nanobots in your blood can literally do nothing right now.

Presumably, once they actually get results that can be published they'll do so. If they don't, they'll have a lot of trouble gaining credibility in the biosciences.

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.

You might also look at last week's articles about Tony Fadell and NEST.

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.

Were all the various alphabet companies ever supposed to be going for big picture "moonshots"? That's the stated goal of the X Lab, but the other branches are for projects that have "graduated" from labs and are supposed to be producing revenue or at least real products in the near term.

They're probably less interested in expensive moonshots and more interested in shoestring moonshots. More hyperloops and fewer MagLevs, for intercity transport, for examle.

Yes, and the other thing about "Moonshots" is that they have profoundly clear goal. "Moonshot" is used to describe really ambitious undertakings and while that's true, the other part of "a moonshot" is having a specific endpoint which people are working towards.

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.

Microsoft Research is close - I'm working on moonshots there AND there are a lot of smart people getting together and solving interesting problems.

Wasn't there a huge firing there in the past year?

No - they let the whole SV lab go, but it was part of a company wide reset. The SV lab was a blow, but it was a very small part of MSR.

>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.

True, just saying that Bell Labs earned its place in history during a time with very different constraints and that such an environment would be impossible or at least very difficult to re-create today.

That's a shame. I'd like to believe that large companies like Google could throw a bit of cash towards open-ended problems. Realistically it's large companies, governments (please no) and research institutions that can afford to think on large timescales. Governments don't and probably shouldn't, research institutions are fine, but large corporations with nice payscales also feel to me like they should be taking part.

Government augmenting research institutions has been the only consistent source of funding for research that doesn't pay out in a 2-5 year timespan.

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.

Here's a neat example of government support for telescopes from a few years ago: http://www.lettersofnote.com/2016/02/the-galilean-moons.html

"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 ..."

Isn't a lot of medical research funded by governments? Wasn't Pasteur essentially working for the government when he made his discoveries?

In the US, the NIH invests $30B/yr (and other agencies also fund medical research). Note, they're more expansive: much of the research has been into basic biology with no immediate applications (often called pure research, although we were trained to always say the research would eventually leading to curing cancer). England and Japan also do this.

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.

>Governments don't and probably shouldn't

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?

Exactly. The government is the only institution which reliably doles out full-sized grants (free money) for projects with horizons on the order of years. And institutional grants on the order of decades.

In a sense, the Internet is the poster child for this, because the alternatives funded and promoted by the telecom industry was so closed that it never grew. The rapid growth of the internet happened because the government funded it for many years (pre-ARPAnet, ARPAnet/NSFNet/etc, then Internet) before it was really ready to expand rapidly. Then, when it was growing rapidly, it funded people to study tech that allowed it to continue to grow rapidly.

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.

>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 ?

I've studied the history of the internet and have also talked extensively with ppl on both sides of the discussion. At best we'd still be using ISDN which isn't much better than 56kbs

Governments are the only ones who actually benefit from thinking long term in the sense that they are the best distributer of the knowledge that comes out of basic research.

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.

UNIX was... Oh wait.

UNIX was derived from MULTICS, which was partially funded by DARPA grants.

"derived from" in the sense of "complete opposite in every way".

Edit: ok, maybe not _every_ way. Pipes are "inspired by an idea from Multics". https://www.bell-labs.com/usr/dmr/www/hist.html

No, it's not completely opposite. You can see UNIX as a simplification of Multics (with the name being a very nice pun!), dropping most of the really hard-to-implement features and focusing on a simple core. Ken Thompson and Dennis Ritchie were programmers on Multics and they admitted lifting features from Multics to implement UNIX.

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.

> 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.

Interesting. Can you name some?

You missed the shell, and the hierarchical filesystem. UNIX later adopted a number of other Multics features as well.

Out of At&T Bell Labs AT&T was a regulated monopoly for a long time.

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.

Government regulation of Bell Labs as a monopoly strongly incentivized them to invest heavily in research, and basically forced them to sell the results to competitors.

This was not the result of free market competition.

Interestingly, AT&T was prevented from selling UNIX as a consequence of their agreement with the US gov't.

Historically, the governments have been the most effective long-term funders of science. THere is an entire class of scientists in the US (Europe, Japan, etc) who exist because of the government's heavy spending to train students (this has sometimes become a problem, for example the US trained so many bioscientists that when the funding froze a few years later, the market became saturated and many of us left for Industry.

Is it really a problem if government scientists leave for industry?

1. It's a problem for students who invested their time (and possibly money, but certainly opporunity cost) in this training. "Sucks for them, not my problem" is 1) excessively uncharitable, and 2) confuses personal interest with overall social goodness. Don't get me wrong -- it's totally understandable for you to say "is that really a problem that we should try to fix?". And in a lot of cases I probably even agree with you! But it's less understandable for you to simply say "that's not a problem".

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.

Yeah, my plan was always to get a full professor position at a mid-tier university. Had I known the NIH budget would have been frozen, I would have not gone to grad school and instead worked for Yahoo pre-IPO, and today I'd be a many-times-over millionaire funding a new crop of scientists out of my money.

I did the same. Interestingly you can always go back to grad school, get a PhD and even become a professor after you're rich.. Probably easier too.

I might not have made it clear, but these are academic scientists who were funded by government, but were unable to transition from grad school or postdoc to Professor because of a high level of competition due to frozen number of professor positions.

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.

> unable to transition from grad school or postdoc to Professor

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.

In my field- biophysics- and also in the larger bio field, the vast majority of PhD recipients intend to go on to be professors, or research scientists at a national lab (I went the latter route when the former turned out to be too hard), rather than academia. Nearly every single person who graduated in my program went on to do either med school (to be a research MD), or postdoc (path to professor). Now, my school was in SF, so a fair number of people were highly incentivized to start companies (NOT go work for companies- to be founders), but I'd say that's rare and fairly specific to prestigious programs in larger cities/higher profile universities.

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.

I think research md is the field to be in. Md/phd. Plenty of jobs always in demand.

I don't think you can really encourage someone to go for a MD/PhD in good conscience now. It takes too long now and is of questionable benefit. I've interacted with many smart MD/PhDs who were pushing 40 and were still not attendings / lacked their own lab. Your career prospects are severely constrained at that point.

You can always go into practice?

Some limited info for 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.

As a long-time academic, yes. Industry does not do research of the same quality as academia; they will not approach as broad a range of questions. No industry would build a large hadron collider, or investigate obscure metabolic pathways or bacterial coat proteins with no known product development goals. Academic freedom doesn't exist in industry.

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.

That doesn't address the underlying point which is that if there's not enough faculty or research positions, people can't stay in academia to do that great research. That's why government funding is so key.

You'd be writing this on a typewriter and pinning it to a board with a thumbtack in a town square if it weren't for government research into computer technology and the internet.

Government is in fact the only kind of institution that can finance science in the good and bad times. Companies have too much of a short-term profit-driven incentive to fund basic science. They can even do that for short periods, when everything looks rosy and they want to attract talent (examples are AT&T, Google, HP, etc.), but profit is so important that they will cut these investment as soon as necessary to improve other areas of the company.

Reply on news.ycombinator.com