Hacker News new | past | comments | ask | show | jobs | submit login
Day in the life of a Googler (Matt Welsh) (matt-welsh.blogspot.com)
121 points by siddhant on Dec 26, 2010 | hide | past | favorite | 49 comments



This article reminds me of those weight loss infomercials:

- In 'fat' mode, the video color is bad, and the person is frowning, sad.

- In 'skinny' mode, the video color is clear, the person is smiling, and their complexion is better.

This gives the illusion that there's a dramatic change in the person's weight, when often they've only lost 10 pounds (which is great, just not dramatic). In the same way, there doesn't seem too big a change in Matt's schedule, there's just a shift in attitude. It probably wasn't a requirement to "Groan at the amount of work I have to do before the onslaught of meetings in the afternoon", or to "spend next 45 minutes reading Engadget, Hacker News, and Facebook". In the end, he had about 3 hours to work at Harvard, and did about 4 hours of work at Google.

It's interesting to dig into what's actually changed between jobs, because you might not completely know what you want in a work environment.


Nitpick: Matt was at Harvard, not Stanford.

With respect to what has changed, one of the things that comes through is that his day at Google is more self-directed and less interrupt driven. Fewer meetings, no petitioning others for grants or industrial money, no need to get up in front of a class, and no meetings with students. Instead there are three or four projects that need to make progress, but how exactly that happens is up to him.


Nitpick noted, and duly edited!

It just seems weird that he would compare schedules to make the point that he likes the kind of work he does better at Google. Lectures and office hours are a productive part of being a professor. He had some meetings about grants, but it was not much longer than his meetings at Google.


Grant writing is stressful and has an impact beyond just the meetings directly discussing the grant. I haven't had the pleasure of writing them myself, but I know several people who do. There is intense pressure because the grants pay the stipends and tuition for your students and postdocs. If the grant does not succeed, you can't support them. In some cases you need grants to pay your summer salary, as well (rules here are complicated). Add this to low acceptance rates from most major funding agencies and long lead times between submission and the decision, and you have a pretty stressful situation.

Most of this work is spent in polishing the grant and figuring out how to best present your research agenda. This is not writing code, so if you'd rather be writing code, this isn't really a good fit for you.


Grant writing is stressful and has an impact beyond just the meetings directly discussing the grant. I haven't had the pleasure of writing them myself, but I know several people who do.

Many people who are good at whatever activity the grant is supposed to support are terrible writers or at least terrible grant writers (which isn't surprising because grant writing can often be extraordinary boring).

From what I can gather, most scientists in the sense of working at universities as profs aren't selected for their ability as writers until they actually become profs, or at least post-docs. Which means that many will be something like 30 before they realize writing grants is important, but that they may not be well-suited to doing so, since few people are.

My family's consulting business does grant writing for nonprofit and public agencies (you can read more about this at http://blog.seliger.com if you're curious), and many of our clients hire us precisely because they're not good at grant writing but are often good at providing whatever service they're providing. We mostly write human services grants but sometimes do technical work. I'm actually surprised that more profs don't find ways to outsource or partially outsource grant writing; they could do for in the neighborhood of $5 – $10K / proposal, which is relatively small relative to the value of their time.


In other words, being a professor has far more sales and management responsibilities than the average software developer's gig.


Yes. It is all about writing business plans in advance: i.e. have the poker-face to promise in advance to the funding agencies' excel-wrangling paper-tossers, that you will make that given discovery [WTF!].

It is like raising VC capital: an agency is likely to be more friendly, if it sees another take some of the initial funding risk. So you sell your idea, and manage the project -- not unlike a feudalistic landlord.


Realize that I have to give lecture in half an hour. Pull up lecture notes from last year. Change "2009" to "2010" on the title slide. Skim over them and remember that this lecture was a total disaster but that I don't have time to fix it now. [..] Give lecture on cache algorithms to 70 or so somewhat perplexed and bored undergrads.

I'm sure there's a little hyperbole but this was outlining a typical day? It's not a surprise the students would be perplexed or bored if taught by someone whose passion wasn't for education. Some colleges get this right; many get it wrong. While it's good to have experts in your faculty, you need good educators first.

Many colleges seem only too happy to coerce expert non-educators into giving lectures against their will, but it makes as little sense as pushing good educators into doing large research projects..


I think it's too soon to tell how things will go at Google. He seems like a dopamine junkie (I can relate) so maybe after a few months he'll be checking HN and Engadget from Google as well.

For a very smart guy like Matt, chances are boredom will set in after a while... it will really be a test of Google to see if it can capture his imagination for 7 hours a day after he's worked there 6-9 months and all the novelty is gone.


He can't be too bad a dopamine junkie considering he got tenure at an ivy league institution.


True. It makes me wonder how he tolerated all the earlier steps of academia only to find being a tenured professor drudgerous.

One hypothesis is that since there is so much competition to become a faculty member at Harvard, there is very little support b/c someone else will be ready to step in and take the job.


What's a dopamine junkie?


Someone who enjoys the constant stimulation of new bits of information.


I clock out for lunch, is this not normal at Google?


We don't have time clocks here, so no.


I don't think time clocks are normal in our industry. Every W-2 programmer I know is salaried.


I'm salaried as well but I'm also expected to be at work 40 hours a week. We have a simple time tracker which I'm referring to as the "clock" but I've honestly never heard of a company that doesn't have at least some form of time tracking if only for HR (vacation, etc).


As faculty he was,

1. Interacting with people who are interested in commercial viability of his research ideas

2. Interacting with students, interested in pursuing research careers and seeking guidance

3. Teaching. Forming concepts in otherwise uninformed minds

4. Coaching grad students who mostly will get inspiration from him to pursue reasearch careers

5. Preparing for talks.

I think the author is under-estimating his contributions back in college.

While he laments the student whose minds are un-prepared, he likes to debug/test the code which too in the same analogy is going thorough curative process.

IMHO, if you compare the long term outputs (ROI), his work in college would far exceed that his does in private enterprise. Hope at some time, he could return back to campus re-invigorated.


Sounds like the author likes hacking but loathes being a professor. Can't help but assume he's doing a huge disservice to his students by being so disinterested in teaching that he forgets about/doesn't even try to improve his lectures from the year before.


FWIW, I sat in on some of his lectures on distributed sensor networks and thought he was great.

Even if he wasn't, he gave up his tenured faculty position to hack full time at Google (IIRC, less than a year after getting tenure). See http://matt-welsh.blogspot.com/2010/11/why-im-leaving-harvar... for some background.


His RateMyProfessors page: http://www.ratemyprofessors.com/ShowRatings.jsp?tid=961898

At least one person thought he was good (though the lack of more than one review may be a sign).


Harvard doesn't use RateMyProf much (there is an internal rating system called the 'q guide').

Chosen at random, his 2009 CS61 class has generally very positive reviews (see: http://imgur.com/cSWsq.png).


I'm sure he felt his lectures were no good. Most smart people are that way. They underestimate the impact of what they do. I've read that Leonardo da Vinci felt he had accomplished nothing when he died.


It sounds like he didn't have enough time to accomplish all his duties as a professor.


That's a tad judgmental, I'm sure Harvard doesn't just hand out tenure to anyone who punches a time card.


You're right in that it's judgmental; I'm not at all familiar with him. Perhaps he's just gotten burned out on professorship. But this is coming from the horse's mouth.


While geting tenure at harvard is indeed no small feat, I doubt teaching has much to do with it. Teaching sadly plays little role in getting tenure, especially for prestigious institutions. Grants and papers are the most significant points.


Teaching has nothing to do with it (see Brian Harvey at the bottom of the ranks).

Papers have nothing to do with it. Really few papers are read these days. I would guess, that an average IEEE author _writes_ the same number or more papers than reads (write-only memory doesn't make sense).

Even less papers have an actual academic impact in CS in the age of the internet, where bits come at the price of caffeine and inspirations of ideas of somebody from a blog/news site (peer review comes in the comments). Not much old-school papers, journals or libraries. As always, there are few exceptions, but you are certainly better off playing the lottery.

Non-virtualizable science is a different matter (wet lab, electronics, physical experiments etc.) But bureaucrats prefer in silico, because it is cheaper.


I may be wrong, but it does not seem like you have much experience in academics.

I don't know any university where papers (and more especially their impact, as measure by h index and co) do not matter a lot when getting tenure (or any kind of post PhD position, really).


They do matter, I think our experience does match. This is part of the problem I was alluding to.

All I am saying is, that I do no longer take for granted the paper/IF (& other Thomson ISI Corp.) science metrics as the golden standard, because it is being actively abused (clique self-referencing, grant business etc.). This is not a new controversy, and I think it is much like the RIAA/MP3 shift, which is not only taking place in publishing (web daily news, kindle etc), but also in scientific publication (NIH publication policy etc).

arxiv.org seems to be one of the best technical compromises, but unfortunately, it doesn't fit very well in the establishment's established authoritative scheme (IEEE, ACM etc), which is a racket by anyone's standards. And it can be misleading too (there was an ARS report on this, about the level of trust in the result being published).

Long story short: I, for one (and I think you too) are very unlikely to get tenure based on papers on arxiv.org about volcanoes, transportation, urbanization and JIT. No matter how good we are in Linux sysadmin an distributed systems.

Though back on topic, publication per se is what early stage training is about (and not PhD), in my experience. Of course, one can always ''downgrade'' an EST to a PhD, but then one has to go through all the post-docs, to get to be staff again. Short of luck, of course. For those with tenure, it is very rare to actually write papers to publish (though it does exist) -- that is what students, assistants and fellows are kept around for.


Oh, sorry about the confusion. I realized afterwards that you may have actually described how it should instead of the how it is. Maybe my English is faulty, but I think it was important to stress this for people outside of academia, it could be misleading otherwise.

Otherwise, I agree almost entirely with you - I myself left academia mostly for those reasons. Whether tenure professors do write papers depend quite a bit on the field, though - I know many professors in statistics who still contribute significantly to papers written by co-authors (mostly grad students, post docs), for example.


No, it is a little worse. He mentions the usual grant-squirrel-wheel.

But he remains silent on the BS part, which I guess he got bored with. I think he got disgusted the day he realized, that writing ''scientific'' papers about virtual economies and pseudo-medical sensor devices and about volcanoes -- as serious they might sound to an administrative bureaucrat handing out pennies for the researchers -- it is just plain BS, even if he may have some facts right. It is making life worse, by diluting the cool stuff with useless junk (fail often mantra, but for its on sake).

If one stops admiring the Emperor's new clothes and breaks ranks on BS groupthink, there is no surprise in exclusion and ostracism. It is a no brainer necessity to leave.

And even [one might even cynically write especially] Harvard is certainly no stranger to ''Scientific misconduct'' -- just a recent example: https://www.nytimes.com/2010/08/21/education/21harvard.html


This reminds me of a quote I read from a lecture a few years back by a history prof at Colby:

"How long will you need to find your truest, most productive niche? This I cannot predict, for, sadly, access to a podium confers no gift of prophecy. But I can say that however long it takes, it will be time well spent. I am reminded of a friend from the early 1970s, Edward Witten. I liked Ed, but felt sorry for him, too, because, for all his potential, he lacked focus. He had been a history major in college, and a linguistics minor. On graduating, though, he concluded that, as rewarding as these fields had been, he was not really cut out to make a living at them. He decided that what he was really meant to do was study economics. And so, he applied to graduate school, and was accepted at the University of Wisconsin. And, after only a semester, he dropped out of the program. Not for him. So, history was out; linguistics, out; economics, out. What to do? This was a time of widespread political activism, and Ed became an aide to Senator George McGovern, then running for the presidency on an anti-war platform. He also wrote articles for political journals like the Nation and the New Republic. After some months, Ed realized that politics was not for him, because, in his words, it demanded qualities he did not have, foremost among them common sense. All right, then: history, linguistics, economics, politics, were all out as career choices. What to do? Ed suddenly realized that he was really suited to study mathematics. So he applied to graduate school, and was accepted at Princeton. I met him midway through his first year there--just after he had dropped out of the mathematics department. He realized, he said, that what he was really meant to do was study physics; he applied to the physics department, and was accepted.

I was happy for him. But I lamented all the false starts he had made, and how his career opportunities appeared to be passing him by. Many years later, in 1987, I was reading the New York Times magazine and saw a full-page picture akin to a mug shot, of a thin man with a large head staring out of thick glasses. It was Ed Witten! I was stunned. What was he doing in the Times magazine? Well, he was being profiled as the Einstein of his age, a pioneer of a revolution in physics called "String Theory." Colleagues at Harvard and Princeton, who marvelled at his use of bizarre mathematics to solve physics problems, claimed that his ideas, popularly called a "theory of everything," might at last explain the origins and nature of the cosmos. Ed said modestly of his theories that it was really much easier to solve problems when you analyzed them in at least ten dimensions. Perhaps. Much clearer to me was an observation Ed made that appeared near the end of this article: every one of us has talent; the great challenge in life is finding an outlet to express it. I thought, he has truly earned the right to say that. And I realized that, for all my earlier concerns that he had squandered his time, in fact his entire career path--the ventures in history, linguistics, economics, politics, math, as well as physics--had been rewarding: a time of hard work, self-discovery, and new insight into his potential based on growing experience."


This is a great story, but it suffers from a certain measure of survivorship bias. Ed Witten certainly had the fortitude and introspection to jump from field to field yet still keep in mind the ultimate goal of his explorations (which eventually led to great success) but my guess is that most people that hastily quit most endeavors that don't immediately strike their fancy do not end up nearly as successful.


So long as you're making an honest effort, and actually working, there's no shame in stopping if it appears to you that your effort will not be rewarded. Witten, from that summary, spent at least a semester on all of his failed endeavors, and if three months is too hasty, we humans may be too short-lived to be properly deliberative.


I agree with your assertion but not really the justification you give. I don't think he needed an ultimate goal; he followed what seems like a random walk but ultimately found something that worked, hugely. There's definitely survivorship bias, but heck, if we discounted all stories by that criterion the world would be a much less interesting, albeit perhaps more realistic, place.


I believe there's another lesson here: the importance of failing early. Ed didn't spend a frustrating career before quitting politics or fours years before quitting math. It can often be far easier to continue doing something you don't truly enjoy than to quit and try something else.


One of the more interesting "I work at Google" posts, supplanted by a superb HN comment :) Thanks.


Comments like this are what keep me coming back to HN...at the expense of my productivity ;)


Interesting. But mostly concerned at the amount of soft drink consumed.


I was especially perplexed by him drinking a Red Bull to get through the rest of the day, and immediately following that with shots of Scotch.


When I glanced over the schedule he had at Harvard, I thought it was from PHDComics.


Stop drinking that Diet Coke!!! It's bad bad bad. If you're trying not to gain weight - don't drink diet drinks. It tastes like sugar and makes you wanna drink more sugar.

Get some tea instead, you can drink hot tea, cold tea, you can add lemon or sugar, and it's much better than ewhhh - diet coke.


So now that he is at Google it looks like he is only really working from 9:00 to 4:00 (7 hrs.) at Google vs. over 8 hrs. as a prof, but in those 7 hours he is getting much more meaningful work done.


It was to 5, just sounded like the day he was recalling they had some drinks in the last hour, assuming that doesn't happen every day. Was also some after hours work, not sure if that was related to his day job though.


I really liked this post. I know it has to be taken with a grain of salt, but I found it a great reality check.

I also have a procastination and web-surfing ( hackernews :) ) problem with a job that doesn't motivate me.

A guy who got a PhD, a tenure and then a job at Google had the same problem. I feel a bit better.


This post should be called "Day in the life of a Harvard professor". I was more surprised with that description. The typical day at Google is pretty much what I would expect.


I'm picking up on some thinly-veiled hostility from several commenters here; does Matt have some poor reputation in the tech world that I'm just unfamiliar with?


only a percent believable! But nicely written :)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: