Hacker News new | past | comments | ask | show | jobs | submit login
John Hennessy named chairman of Alphabet/Google (ieee.org)
455 points by teklaperry on Feb 2, 2018 | hide | past | web | favorite | 101 comments

This guy is one of my heroes, I could not be happier. Go him. The H&P book on computer architecture is a book that changed my life. He is like Knuth except that I understand what he says. The Knuth stuff, I get it, it's profound, but I never got it. It was too hard. H&P, they laid it out in ways that I could understand. I'm not stupid but I'm not that smart, I really loved that book, it was written in a way that not so smart people get it.

If there was a way I could say "go you", yeah, I want to do that. He's a great guy, has done great work, go him.

And all that said, I bet this is not a "go him" job, I bet he makes a difference, maybe a huge one, maybe not so huge, but I bet he makes a difference. Let's come back in 5 years, I bet he will have made stuff better.

I too loved "Computer Architecture: A Quantitative Approach", and read it all the way through before the semester even started.

Each chapter had a pithy quote at the start, and my favorite was the excerpt from the EDSAC instruction set:

Z -- Halt and Ring Bell: Stop the machine and ring the warning bell

Meh, it has a lot of fallacies and pitfalls.

Can you elaborate a bit? Unsubstantiated drive-by insults aren’t very useful to other readers.

Edit: thanks for the explanation. Note to the several downvoters: you aren’t accomplishing much here by “punishing” innocent unfamiliarity with an in-joke...

It's a joke. Every chapter in the book has a section titled "Fallacies and Pitfalls" describing commonly held beliefs that should be avoided for the topic of that chapter. See for example page 58 here: https://books.google.com/books?id=cM8mDwAAQBAJ&lpg=PP1&pg=PA...

Can you recommend a book that would serve as a useful prerequisite to CA: AQA? Based on your recommendation, I flipped through the first few pages on Amazon, but it felt a lot like Algebra 2 without knowing Algebra 1.

At my university CA: AQA was used as the text for the second course in computer architecture. The first course in computer architecture used this book: https://www.amazon.com/Computer-Organization-Design-Fourth-A....

If you do get around to reading CA: AQA, you can follow along with this Coursera course (it's taught by the professor I had when I took it at Princeton, and he does a great job teaching it): https://www.coursera.org/learn/comparch

I don't have enough patience for courses. I sort of scan the material for what I really need or want to learn and ignore what's not interesting to me. I can make consistent A's with the benefit of Adderall, but I don't feel comfortable with that.

Your recommendation was very useful and interesting to me, and I bought the paperback. I was surprised that it was cheaper to buy a used paperback than to rent it via Kindle for a month.

It's another book by the same author - Computer Organisation and design : the hardware/software interface - http://a.co/aQYK7DG

If you want a more in-depth look at the book a quick google found me this on the first page results.


Seconded, that's a great book.

Also, Hennessy was also one of the minds behind the MIPS architecture.

I found reading the Art of Computer programming books wasn't that bad. If you ignore the quantative analysis for the algorithms and just look into the algorithms themselves then there are a lot of gems in here that you just don't find anywhere else. If you want to learn how to prove the algorithm runtime then there is obviously a lot more to the books, but for finding other algorithms and data structures it is a great tome of options and its organised in a way to just dip into when you have a particular problem.

Don't ignore the books on the assumption its too hard as that wasn't my experience of them.

I see a new 6th edition came out in December 2017

Are the new editions worth it? My library lets me get an e-book of the 4th edition from 2007 for free, so are the last decade of added material worth buying it?

Computer organization and design was by far my favorite class (at UMich). It really played well to the puzzle of engineering that makes me love the field in the first place.

Eecs 370 or 470?

They both use one of Hennessy's book. Although I think 370 used Tanebaum's years before that

370, the exams for that course were just a blast, only exams I've had genuine fun taking in my life. Would have done 470 but I graduated a semester early (that and programming languages courses are ones I retrospectively regret not having taken). Computer Organization and Design 4ed was the one at the time I think - it is Hennessy.

I love computer science now a lot more than I did at the time, heh, but that's part of why I bothered finishing early. The professional environment suits me well.

Just curious, what kind of work do you do now?

Software engineer. Same username on github will lead you to linkedin if you're curious about my (relatively short) career timeline there.

C'mon, all you are doing is selling yourself short

Hennessy was an incredible steward at Stanford, and I'm 100% sure that he will be a difference maker at Alphabet. I don't see him taking the position if he didn't have an incredible vision for the future that he could drive.

The intro to this video talks a bit about his accomplishments as president.


This video is probably more in line with the interests of the community here (Stanford Seminar - Entrepreneurial Thought Leaders: John Hennessy of Stanford University):


I was in the audience when he announced the launching of the Knight-Hennessey scholarship. It was an incredible speech. He was always an incredible fundraiser, but putting together a $750M fully endowed scholarship program is a heck of a way to say goodbye to the university.

> I don't see him taking the position if he didn't have an incredible vision for the future that he could drive.

How much influence does a chairman really have on the the vision though? Isn't that more of the CEO's business?

I suppose the board picks the CEO, but that spot doesn't appear to be changing anytime soon.

"How much influence does a chairman really have on the the vision though? Isn't that more of the CEO's business?"

In an ideal situation the chairman can function as the intellectual sparring partner of the CEO. Not many can do that, as some things cannot be discussed outside of the company personnel, and going downwards in the chain of command always has this gradient of authority which can function as an intellectual inhibitor.

The Chair also acts in the last resort to sack the CEO if the company is going off the rails - soothing which the Chair at HP should have done a few years ago.

His CV page [1] is clean and quite up-to-date. Not sure if he handles the edit himself or not, but I find the page's design clean compared to a lot of the professors' out there. I just think it's interesting to mention here.

[1]: https://web.stanford.edu/~hennessy/cv.html

    <html xmlns:v="urn:schemas-microsoft-com:vml"
    xmlns:mv="http://macVmlSchemaUri" xmlns="http://www.w3.org/TR/REC-html40">

    <meta name=Title content=cv.html>
    <meta name=Keywords content="">
    <meta http-equiv=Content-Type content="text/html; charset=unicode">
    <meta name=ProgId content=Word.Document>
    <meta name=Generator content="Microsoft Word 14">
    <meta name=Originator content="Microsoft Word 14">
    <link rel=File-List href="cv_files/filelist.xml">
    <link rel=Edit-Time-Data href="cv_files/editdata.mso">
    <!--[if !mso]>

I wonder if Google is going to offer incentives for him to switch to Google docs, like free 100mb quota upgrade on his Google account

Whoa, pump the brakes there. Nooglers need to earn that upgrade.

I do love how these kind of CV sites always seem to abhor any CSS styling that would not be understood by a twenty year old browser.

EDIT: and now contrast1970 has solved the mystery!

Yeah but this one is particularly neat though.

Most profs don't want to maintain an HTML CV.

Impressive! Only negative: the "Slides from Recent Talks" section is all broken links.

So now both H & P working at Google? (Though I don't know enough about corporate board stuff to tell if chairman is "working for")

A "Board of Directors" is the executive unit of a publicly owned company [https://en.wikipedia.org/wiki/Public_company]. It is separate from Google's internal corporate hierarchy.

Pretty crazy! Hopefully they get Patterson to work on some novel computer architectures. TPUs are very cool but they are not for consumers.

I know what you mean, but e.g. every time the Google Assistant talks to you, the consumer, you're using TPUs.


They're getting there: https://cloud.google.com/tpu/

Who are "H & P"? I haven't been able to track down who the P is.

Computer Architecture: A Quantitative Approach is a seminal text in computer science by John L. Hennessy and David A. Patterson.

From what little I've read of it for my classes in the subject, it's well written and accessible.


I'm a CS PhD student, but my undergrad and master's are in physics, so I didn't go through all of the same steps many did in their CS curricula.

It is usually in graduate courses, I guess maybe some better schools' undergrad programs.

Agree with gp, very well written text. Probably the only class textbook that I have actually read entirely. (10 years ago or whatever)

I’m sure H&P is the CLRS of architecture.

Patterson. Computer Architecture is their landmark book

To me, Patterson is the real superstar of the duo.

He was also President of Stanford for a bunch of years.

Weird.. In a time when people are extra tuned to the relationship between technology and society and whether just because we can build something, should we (information bubbles, etc), it's strange that Alphabet would pick someone so out of touch with their impact to the long tail of humanity.

At least that's the impression I got of Hennessey from Malcolm Gladwell's podcast http://revisionisthistory.com/episodes/06-my-little-hundred-...

I'm glad he got his books updated with the RISC-V editions before this

Haha, while I'm happy for him and excited for what this means for Alphabet, I had the same thought. Something tells me this guy would be able to juggle a new edition nonetheless

>>RISC pioneer John Hennessy’s side hustle gets serious

I couldn't help but get a chuckle out of the subtitle.

One day I hope my side hustles involve helping to run an $800B company.

I have no issues with John Hennessy, but I’m surprised they didn’t give the chairman position to Larry Page, who’s CEO and a Co-Founder. I wonder what’s preventing him from taking over full control...

Sarbanes-Oxley Act strongly encourages non-Employee Chairman of the Board. It’s the direction that well-governed companies are moving.

Paige and Brin have enough voting power to pretty much make any major decision.

How much power exactly? Or do they own some special shares that don't need majority to rule?

Together they control over 50% of the voting power. They both own substantial amounts of Class B shares, which aren't available to the public and carry 10 votes per share.

What’s going on with Schmidt? Was this planned?

Schmidt has always been one of the people politically greasing the wheels for Google. You most often see his involvement where the politicians are. It very much sounds like in his "technical advisor" role, he'll continue to do that. He said he was interested in 'science and technology issues' and 'philanthropy'. That reads as "politics" to me.

He stepped down at the end of 2017 in order to focus on his charities(according to the press release). But, his daughter did die last year as well, perhaps that is the cause for him refocusing his life

It's also worth mentioning that he's still on the board, just no longer as chair.

Could be as simple as when you’re ludicrously rich and have been doing something for awhile you can get bored, especially if you’re not the founder or otherwise heavily emotionally invested in it.

Most likely he was cast aside so Google can cozy up better with the Trump administration. Schmidt was close with the Clinton campaign, which Trump himself noted when they met:


I think this conceptually holds as much water as a colander.

Schmidt has been on his way out for a couple years - it appeared to have been part of a planned transition.

For some context, this article in the New York Times presents essentially the same theories/discussion: https://nytimes.com/2017/12/21/technology/eric-schmidt-googl...

I think he was on his way out to a post in the Clinton administration as Google’s man on the inside. He’s mentioned his political ambitions in several interviews/profiles.

When that didn’t pan out, they didn’t see much use in keeping him around, especially in light of Trump’s penchant for grudges.

Suggesting, as that article does, that dating outside your marriage "in the age of #MeToo and the spotlight on sexual harassment in the workplace, might be problematic for Alphabet" is a heck of a leap. Leading a possibly non-monogamous lifestyle is sexual harrassment/assault? Wow.

The New York Times is happy to make that "heck of a leap" as well, albeit with slightly more cautious wording: https://nytimes.com/2017/12/21/technology/eric-schmidt-googl...

I don't necessarily agree but or at all but the downvotes have made this alternative opinion too light to read. Chill out people.

Google isn't some benevolent entity. They've recently been on a hiring spree for Republican lobbyists (which most here consider much more evil than democrats). Personally I think both parties are not that different.

>Personally I think both parties are not that different.

It frustrates me greatly to see comments like this. It's worth taking a moment to even briefly skim the official platforms of the two major parties. They are profoundly different on a enormous range of substantive issues and policies.

Perhaps there are some issues on which the two parties agree, and certainly there are many political positions which neither party represents. That hardly means that the two parties are "not that different."

The thing that makes them "not that different" is how they both make jokes of their official platforms by their actual decisions.

That's true for many issues like taxes, entitlements, gun control, immigration, healthcare, abortion, environment, etc.

However, it probably fair to say both parties have most candidates skewed to the middle on many of those issues. Lining up strongly with the party on all of those things doesn't win elections usually. And it's also fair to say both parties have members that will betray those base positions for wealth or power.

I can see why someone would call them similar based on frustration with a two party system that rewards moderates.

The Platform of the party is worth less than used toilet paper. It literally is meaningless as an example of what the party actually stands for.

I kind of agree, in the sense that they are similarly ethical/reasonable. Their platforms couldn't be more different, but I don't see one party being evil and corrupt and the other party being good, which is what often seems to be the opinion online.

It's not about policies. Policies are entirely different but equally self serving.

I'm astounded that there are blacks living in ghettos but Nancy Pelosi shuts Govt down for people who are pondered to by speaking Spanish but paraded as having known no other culture and assimilated.

I don’t mind the downvotes or anything, but it struck me as strange that on the same day that it’s reported Google is pursuing partnerships with an entity as corrupt and demonstrably awful as the Saudi regime, people see this as outside the range of possibility.

Are you serious? Why can't it be that the CEO under performed and couldn't manage their increasing costs to pay Apple and Mozilla. Also, that they can't seem to sell consumer electronics unless they are low margin phones?

Because Schmidt isn't the CEO. (note that I also don't really think its politically motivated either, but that's certainly more likely than thinking that Alphabet is underperforming).

There were some rumors.

Duckduckgo is your friend.

Does this mean good things for RISC-V?

Nope. That's Patterson.

Yep. And as someone engaged in the RISC-V community, I've seen zero evidence of Google involvement outside of Patterson. Normally if they had something brewing, you would start to see signs on the forums, email, or Github.

Amazing and Awesome. Congrats.

Great choice.

Is this a conflict of interest?

Correct me if I'm wrong but didnt Hennessy, while President of Stanford, agree to give away university-owned page rank IP to Google for $1mm (a company he was conveniently personally invested in) thereby lining himself up for massive personal gains at the expense of the University?

Right, that's why Google is still paying Stanford for the license, and gave the university shares that were worth $360M in 2005 already? [0]

"In 2012, Google paid Stanford $80,000 for the license, a sharp decline from the $400,000 paid in 201a and only 1/10 of what Google paid in 2010. Since Google began disclosing these payments, the company has paid Stanford around $3 million in licensing fees. Of course, that’s in addition to the $1.8 million shares that Google issued to Stanford, and which the school sold for $336 million in 2005.Google made up some of the difference by donating $3.4 million to Stanford for scholarships and what the proxy describes as “other philanthropic endeavors”, compared with $3 million in 2011."

[0] https://www.footnoted.com/new-details-in-googles-proxy-on-st...

Stanford sold for an average of $187 per share for the 1.8mm they received. That was years later at the IPO beyond. At the time they received them the value was far far lower (likely on the order of what I suggested). And the license amounts.. the article says it was 80k in 2012. That's an awful licensing agreement by any measure. This is a company that Stanford's own president - a genius by any measure - was an investor in so clearly believed in it's value, and yet he agreed to sell the IP essentially for a pittance to himself.

The annual fees to use the algorithm is probably not substantial and a quirk of the agreement, and the substantial part of the licensing agreement is the 1 million or so shares that they received.

The share grant must have been fairly early on the history of the company (or else Google won't be able to use PageRank, which was fundamental to Google Search), probably pre-Series A. Google itself wasn't worth that much back then, so the 1M shares was probably bigger slice of a much smaller pie that eventually got diluted through later funding rounds.

It was reported that Standford made more than $300M off their shares and they get yearly royalties. I think they did ok on that deal.


That makes it "only" the second most profitable Stanford patent ever. If they had only waited a little longer before cashing out...

Create more value than you capture.

According to [1], Stanford received 1.8 million shares of Google stock (I'd imagine for PageRank and other rights), and they sold it over time for $336 million.

[1] http://www.redorbit.com/news/education/318480/stanford_earns...

If they hadn't sold it in 2005, it would be worth $3.6B now (10x in 12 years), more than 10% of Stanford's total endowment.

If they hadn't sold it then should they sell it now?

Always sell half, and keep half.

$3.6B figure looks about right, but the growth was more like 20x. Don't forget the 2:1 stock split in 2014.

Is he not the author of the famous book - Computer Architecture : A Quantitative Approach by Hennessy Patterson?

Indeed the same.


Wrong Hennessy, wrong drink.

Unless we're talking Ballmer Peak, in which case Hennessy can be the correct answer to either.

I was disappointed too that this guy doesn't have anything to do with the cognac ...



victory of the meritocracy! if only this principle was applied to politics (no flags please)

Registration is open for Startup School 2019. Classes start July 22nd.

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