Hacker News Reader

Comment by kulkarnic | Hacker News Reader

As someone who's worked at MSFT himself, the author certainly knows what he's talking about. However, there's a couple of factors one must reflect on:

1. It's a HUGE company (80k employees?). When you get to that size, it's hard to ensure all great ideas get to market. For a very long time in its history, the company was guided by a dominating spirit (computers on every desk) and a similar leader (Bill Gates). With both outmoded now, it's easy for business-groups to take local decisions, rather than follow the war-plan.

2. BUT, they've made HUGE investments in the future. Consider Microsoft Research. MSFT isn't going to be IBM anytime soon (huge company, but better known for sales folks than tech enterprise). Though they'll possibly not be as sexy as Apple, they'll definitely have the tech/IP if/when they want to utilize it.

View on HN

I think you hit on their biggest problem now. They have (mostly) succeeded in their original goal to get a computer on every desk. What's next though? It seems to me that they need a new grand plan for the company as a whole.

Exactly right, and now a company like Google which makes almost all of its money off advertising can have a driving mantra like "to organize all the world's information" and utilize that network. Furthermore, Google can pump money into open source and new devices to bring people online - things that make the web more important to people's lives - and it will benefit their core business. Most of the world's population isn't even online, Google could even make a case to buy out ARM and licence it for free to other companies (inter-licencing with other processor companies might be difficult) just to expand the market to nearer to its full potential (6 billion and rising). Whereas Microsoft faces having to sell software products to different markets, like games and movies do, which basically means price-fixing and changing the product (be that through crippled "lite" versions of Windows for the developing world, or the price-fixing that is involved in region-encoded DVDs and regional sales of games).

Their current mission statement is a bunch of crap, something like "Enabling businesses to reach their full potential".

But the other mission statement-y phrase that gets thrown around a lot is having an integrated experience between "three screens and the cloud". The three screens being TV (Xbox), computer, and cell phone. That gives a little direction, but as much as I'd like to see them be competitive in mobile, I'm not optimistic.

Based on their latest presentation at CES, they have even lost that tiny sense of direction. Now the strategy has become "Many screens and a cloud" to encompass netbooks, tablets, etc. that weren't easily anticipated a year or two ago.

Maybe I'm cynical, but I feel like a long-term mission statement shouldn't be changing every 12 months.

They really should go back to "Information at your fingertips."

I fully agree with you. For past few years they are all over the place. From digital homes, car entertainment they are pursuing everything. And now retail stores. They are forgetting the fect that they still have not delivered an OS that will replace XP in corporations.

I am not so sure about that. Win 7 is so far, as good as or better than XP. I am now blessing it for my business clients.

New machines, granted. But that, unfortunately, is de rigueur in this industry.

I don’t think your example in (2) is particularly effective. At least through 2000 or so, IBM spent more on R&D than just about anyone, Microsoft included. Others now spend comparable amounts, but IBM still got the most patents of any company last year: and they have for 17 years straight.

Big R&D budgets, with lots of interesting advanced research, DARPA grants, etc. etc. can certainly be helpful, but they don’t save a company that lacks a broader product strategy.

I'm somewhat familiar with the research models afforded by both companies, and have a slightly different view.

Research at IBM is mostly product-group driven. Microsoft affords more flexibility- allowing more academic research, for example. This puts them farther into the future, IMO, than solving immediate product research needs. What I meant by investment above is more than monetary- it takes a lot to get a good research team going.

I agree that nothing can save a company with shut eyes. A strong research group is a way to keep them open.

That may be true of IBM today, but I’m not so sure it’s true of IBM 15 or 20 years ago, which is, I think, what you were trying to make your point about, above. For decades, IBM made more than half of all the CPUs in the world, and really had by a considerable margin the most R&D of any tech company, including all kinds of extremely forward-looking “basic research” type stuff. More recently, they’ve switched to a much more services/consulting “solve particular companies’ problems one at a time” type approach, and correspondingly sold off their PC business, among other consumer-oriented parts of the company, and ditched a lot of the most pure academic research. But that’s only been after they stopped being the dominant computer maker.

Anyway, I think you could possibly be right; I just think you need a better example than IBM. :-)

They still make the CPUs for the XBox 360, PS3, and Wii, and they do a fair amount of collaboration with AMD, so they're still making an impressive number of CPUs.

Perhaps, but consider:

* Agile practice indicates that experience with users is the best way to keep project on a useful path. Research groups necessarily act without such input, which leaves them prone to going off into the weeds, including producing interesting technologies without a total appreciation for the practicalities, or a clear path to delivering it to users.

* One of Jobs' first acts on returning to Apple was to disband the research group (http://news.cnet.com/2100-1001-203996.html). Why? I would guess because he wanted research to happen in the course of and for the purpose of innovative product development. There you can get the benefit of a user-focused development process, with a defined end result.

* As with any institution (e.g. NASA http://news.ycombinator.com/item?id=1092972), without accountability from strong leadership or users they're prone to unproductive efforts. After all, how do you judge the quality of their efforts? How can you motivate them to strive - day to day at all levels - to make an impact, when they have these cushy unaccountable ivory tower positions?

Or, more practically, what has MSR produced that's active in changing the world, compared to a handful of startups acquired by Apple?

Or, more practically, what has MSR produced that's active in changing the world, compared to a handful of startups acquired by Apple?

the premise of funding basic research (as opposed applied product-driven research) is that only by giving researchers freedom from short-term demands can it be possible for them to do something truly innovative and 'out of the mold'. a 'classic' example is all the work done by brilliant researchers at Xerox PARC in the '70's that was way ahead of its time; of course, Xerox execs have been kicking themselves in the rear for not capitalizing on those technologies (Apple and Microsoft did), but the hope is that Microsoft will be smarter than Xerox was and actually capitalize on basic research created in MSR. Giving researchers freedom will breed lots of supposedly-useless work (that's only useful or interesting to academics), but it provides a greater possibility of doing something totally game-changing.

Since you bring up Xerox PARC, don't/didn't they suffer from the same problem? Lack of/poor marketing?

You raise some great points there- the best is qualifying what kind of research is actually a good thing.

It may (I'm skeptical though) that an agile business can thrive without investing in research.

But then, who invents the future (say the Fusion kind of 30-year-future pointed at above)?Our fallback is an underfunded university system, but it lacks market insight.

At this point, I don't really know if product-embedded research is the way to go. Apple seems to be pretty good at this-behold multitouch everything-but I'm curious how they work- How do they hire researchers for instance? If someone knows, please share.

You raise some great points there- the best is qualifying what kind of research is actually a good thing.

It may (I'm skeptical though) that an agile business can thrive without investing in research.

But then, who invents the future (say the Fusion kind of 30-year-future pointed at above)?Our fallback is an underfunded university system, but it lacks market insight.

You raise some great points there- the best is qualifying what kind of research is actually a good thing.

It may (I'm skeptical though) that an agile business can thrive without investing in research.

But then, who invents the future (say the Fusion kind of 30-year-future pointed at above)?Our fallback is an underfunded university system, but it lacks market insight.

> Or, more practically, what has MSR produced that's active in changing the world, compared to a handful of startups acquired by Apple?

I see a lot of great stuff coming out of MSR; their Simons are almost single-handedly responsible for GHC's internals and over the last 2 decades making Haskell a usable, nay, fast, language. But what OSs is Haskell principally developed & used on? Linux or OSX.

for me it looks like Microsoft is doing a lot of research into the equivalent of 'Fusion' (always fifty years away) - which is not of much benefit for their current customers...

I agree with both points. In addition, from the sound of the article, it seems like the biggest problem with having 80k employees is the jealousy and feudalism at the management level when it comes to actually implementing the great ideas across multiple teams—the ClearType and Tablet PC incidents being examples of this.

Is the problem “jealousy and feudalism” in general, or is the problem efforts by the leaders of the Windows and Office units to preserve their turf in particular?

I got an offer from MSR a while back. I remember asking, "so... how do you guys ship?" and they actually thought this was funny, laughing and laughing. "That's so cute!"

Because for them, it was an impossibility.

Why would you expect a research lab to ship a product? Seems like you were interviewing with the wrong group.

A research lab shouldn't directly ship products, but there should be a path from research to productization. MSR seems to mostly lack this path.

It depends on the group. When I interviewed there recently, one of my interviewers was there in exactly this capacity (taking research into products).

It's also a non-goal. I used to work there. It was hard to see stuff ship since researchers seemed incentivized on some value of 'prominence' (research community and MS-internal).

Even in the cases where they productize it - the Program Manager on the product team is not going to get a bonus by taking on some new project in MSR in addition to the commitments filed last year.

So researchers get on just fine without slaying dragons to ship. More cool papers to publish anyway.

A true story: a CS professor promotes Java widely and even writes undergraduate textbooks on Java.

Goes off on sabatical with Microsoft Research, and becomes a .NET evangelist. Stops writing Java books and starts writing .NET books.

Is there a quid quo pro? Probably not. But it does illustrate another purpose for Microsoft having a research organisation: winning over academics, and, by proxy, their students.

Reply on news.ycombinator.com