9 ms·

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.

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.

lispm 17y ago
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...
jacobolus 17y ago
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. :-)
- krakensden 17y ago
  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.
Empact 17y ago
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 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 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?
- gwern 17y ago
  > 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.
- kulkarnic 17y ago
  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.

lispm 17y ago

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...

jacobolus 17y ago

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. :-)

krakensden 17y ago
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.

Empact 17y ago

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 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 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?

gwern 17y ago
> 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.
kulkarnic 17y ago
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.