His early work was at Bell Labs, and that experience shines through in his views on management in tech. My first work experience was in working with a computer scientist who had been at Bell, and what stayed with me was his reverence for the institution and it's culture.
It may be fair for say we inhabit the world that Bell Labs (and PARC) built.
We definitely inhabit the world that Bell Labs & PARC built, but to Yann's (and Arnold's) point, it is worth understanding why neither of these organizations or their parent companies were able to exploit the inventions that they made.
I find that my critics are always more right about me than I would like. Which means I can learn by listening to them. The only thing - they don't have any actual incentive to improve me.
For academia, we talk about the three legs of the stool, but academia is actually very good and very practiced, historically-speaking, at preserving and transmitting knowledge. The prestige associated with 'creating' knowledge has been coopted and corrupting to academia. It's not a place to make new things. It never has been. Making new things is actually a problem - more new things made in academia are WRONG than right, and the rigor to test them is only available outside, in industry, where wrong ideas very quickly turn into failures. New ideas need to grow within industry (or with hobbyists who are insanely focused and self-critical, and this only in rare cases) and then come back to the academy to be disseminated.
More things in ANY research are wrong than right. Ever heard the origin of the lightbulb? Anyway, academic research is far from optimal but I wouldn't dismiss it either.
1 - In the last couple decades, lots of academics have started companies based on their research and done quite well.
2 - If we look at the DARPA Challenge, it would seem academics took driverless cars from impossible failure to something successful in that challenge and made it clear a road-worthy product was achievable. That's far from the only example but surely a clear one.
Marc's comment is a gross oversimplification - large companies that have been around for a 100 years or more aren't just ossified bureucracies - free markets ensure constant adaptation & innovation. As you say right at the end, you need both - in competitive markets.
Peter Drucker wrote about the advantages in a start-up of having two or three co-equal founders, or nearly co-equal. Hewlett & Packard were like that, "As HP grew, Bill and Dave created independent divisions to encourage innovation and accountability, while the CEO drove overall strategy."
Innovation, making things better by making changes. Accountability, what managers try to provide to their human resources who are doing the work in the way the managers want, and watch (either closely or not).
The huge huge improvement in management at profit oriented companies is based on revenue from the production of stuff customers actually pay for, and the costs to make that stuff. Measured in dollars, called profit -- and quite objective. Unlike most govt, or non-profit orgs. The creatives create improved products, the managers keep the costs low.
On the prior post idea Mob vs Individual, the creatives are the sovereign individuals, being held accountable, tho not quite managed, by the Mob. Very much not good, right now, but consistent with an increasingly popular idea that non-owning stakeholders should have more control over companies -- more Mob rule. Which I'm increasingly against, in some ways, yet I'm also more angry with and negative towards Big Tech, and supporting the idea that "something must be done" (because now I'm in the Mob that wants something done).
Freedom to act, especially as an individual, so often in conflict with Security of oneself and of others around you.
I'm reminded of this Twitter thread by Shreyas Doshi, which explicates three types of product leaders - the Visionary, the Operator, and the Craftsperson.
That creativity bureaucrat trade-off is very tricky to handle. Government institutions evolve into bureaucratic monopoly institutions that fail new challenges, often with real silly stupidities.
"In fact, I am conscientious about keeping promises, but I rarely follow up on my ideas. I would rather keep coming up with new ideas than pursue any one of them to final implementation."
I have a hard time believing you rarely follow up to completion but if so that would be a problem in academia.
I look forward to reading the pieces you link. I may have already read one of them.
I'm a creative type that ended up in a stability industry because I needed stability in my life (my health was unstable enough).
On the one hand, it's been a good fit financially (I've done well because I've got a skillset most don't). And if your the only person with drive and vision you can often get your way.
On the other hand it can be very frustrating on a day to day basis. Constantly being told we can't innovate because of "office politics."
The saving grace is its all been work from home. I can see why some creative types, especially single ones in good health, would enjoy collaborating with other creatives in person. I think that is why Musk hates work from home. But if you're a creative type that would be driven "Office Space" mad by having to physically do that shit in person every week WFH is a huge blessing (also, anyone with kids). The fact that I can just ignore 90% of corporate bullshit working from home keeps my sanity.
I’d argue that Microsoft (and Facebook) are both in a third category not covered here because it’s mainly a management skill called “fast following”. Neither of these firms truly disrupted the world with their own innovation but rather, bundled others marketable concepts for mass market very well. That’s certainly a “kind of” innovation but not truly disruptive.
Some say that Apple.(and Steve Jobs specifically) shipped other peoples ideas but I disagree and unusually, they’ve continued to innovate without Steve fairly well.
All of these companies have research divisions and the products you mentioned from Microsoft are results of that work but Microsoft essentially meets existing demand rather than successfully creates it. The list of their products that fit that description seems endless. Same thing with Facebook but NOT true of Apple or Tesla/SpaceX, and it used to not be true of Google (at all) who truly created new demand but has clearly become lost in the wilderness with good, smart, but now seemingly ineffective managers.
Amazon on the other hand stated out balancing an existing, rapidly growing demand (I don’t think they created demand) but then succeeded later in “leaning into” the demand “satisfaction” business and as an infrastructure provider … which, when it comes down to it, was slightly more fast following than a truly creative process. People forget that when they first got online (if it was after 1998), the first place the went was often this mythical “Amazon” where they likely bought something just to see how the WWW might work. Implementations have always been super key to Amazons success. Work with enough software engineers (especially pony-tailed Perl programmers) and you know what I mean.
Kling did an amazing piece about a year ago on why innovation escapes large, successful companies due to relative risk aversion at scale. If I can find it, I’ll repost it here.
While I also think MS overpaid for both, I notice a lack of an alternative valuation. $4 $2 $1? Billion.
Had they developed Skype well, Zoom would not have taken off. That’s the MS implementation problem, not the M&A, nor price. Had they made Skype 2 as good as Zoom, it would have been great.
Similarly, MS stupidly wanted windows on a phone, rather than a phone OS that works well with Office, as a third phone choice.
Great examples of management failure to integrate or innovate effectively, partly due to size.
Reminds me of successful skunk works in military development, with some Director protecting the project from financial suits & auditors, so as to be successful.
Interesting comments from Yann LeCun on this topic, here: https://twitter.com/ylecun/status/1773706233985282119
His early work was at Bell Labs, and that experience shines through in his views on management in tech. My first work experience was in working with a computer scientist who had been at Bell, and what stayed with me was his reverence for the institution and it's culture.
It may be fair for say we inhabit the world that Bell Labs (and PARC) built.
We definitely inhabit the world that Bell Labs & PARC built, but to Yann's (and Arnold's) point, it is worth understanding why neither of these organizations or their parent companies were able to exploit the inventions that they made.
I find that my critics are always more right about me than I would like. Which means I can learn by listening to them. The only thing - they don't have any actual incentive to improve me.
For academia, we talk about the three legs of the stool, but academia is actually very good and very practiced, historically-speaking, at preserving and transmitting knowledge. The prestige associated with 'creating' knowledge has been coopted and corrupting to academia. It's not a place to make new things. It never has been. Making new things is actually a problem - more new things made in academia are WRONG than right, and the rigor to test them is only available outside, in industry, where wrong ideas very quickly turn into failures. New ideas need to grow within industry (or with hobbyists who are insanely focused and self-critical, and this only in rare cases) and then come back to the academy to be disseminated.
More things in ANY research are wrong than right. Ever heard the origin of the lightbulb? Anyway, academic research is far from optimal but I wouldn't dismiss it either.
1 - In the last couple decades, lots of academics have started companies based on their research and done quite well.
2 - If we look at the DARPA Challenge, it would seem academics took driverless cars from impossible failure to something successful in that challenge and made it clear a road-worthy product was achievable. That's far from the only example but surely a clear one.
That’s a really good insight.
Marc's comment is a gross oversimplification - large companies that have been around for a 100 years or more aren't just ossified bureucracies - free markets ensure constant adaptation & innovation. As you say right at the end, you need both - in competitive markets.
Some are, some aren't but it is true that a company has to do a few things right to last many decades.
Peter Drucker wrote about the advantages in a start-up of having two or three co-equal founders, or nearly co-equal. Hewlett & Packard were like that, "As HP grew, Bill and Dave created independent divisions to encourage innovation and accountability, while the CEO drove overall strategy."
Innovation, making things better by making changes. Accountability, what managers try to provide to their human resources who are doing the work in the way the managers want, and watch (either closely or not).
The huge huge improvement in management at profit oriented companies is based on revenue from the production of stuff customers actually pay for, and the costs to make that stuff. Measured in dollars, called profit -- and quite objective. Unlike most govt, or non-profit orgs. The creatives create improved products, the managers keep the costs low.
On the prior post idea Mob vs Individual, the creatives are the sovereign individuals, being held accountable, tho not quite managed, by the Mob. Very much not good, right now, but consistent with an increasingly popular idea that non-owning stakeholders should have more control over companies -- more Mob rule. Which I'm increasingly against, in some ways, yet I'm also more angry with and negative towards Big Tech, and supporting the idea that "something must be done" (because now I'm in the Mob that wants something done).
Freedom to act, especially as an individual, so often in conflict with Security of oneself and of others around you.
As a semi-insider what is your explanation of the Fed just not keeping/quickly returning inflation to its target level in 2008-2009?
Locklin on science had a piece today that hit on some of these themes.
https://scottlocklin.wordpress.com/2024/03/28/against-the-nerds/
I'm reminded of this Twitter thread by Shreyas Doshi, which explicates three types of product leaders - the Visionary, the Operator, and the Craftsperson.
https://twitter.com/shreyas/status/1375491623308550144?lang=en
That creativity bureaucrat trade-off is very tricky to handle. Government institutions evolve into bureaucratic monopoly institutions that fail new challenges, often with real silly stupidities.
"In fact, I am conscientious about keeping promises, but I rarely follow up on my ideas. I would rather keep coming up with new ideas than pursue any one of them to final implementation."
I have a hard time believing you rarely follow up to completion but if so that would be a problem in academia.
I look forward to reading the pieces you link. I may have already read one of them.
I'm a creative type that ended up in a stability industry because I needed stability in my life (my health was unstable enough).
On the one hand, it's been a good fit financially (I've done well because I've got a skillset most don't). And if your the only person with drive and vision you can often get your way.
On the other hand it can be very frustrating on a day to day basis. Constantly being told we can't innovate because of "office politics."
The saving grace is its all been work from home. I can see why some creative types, especially single ones in good health, would enjoy collaborating with other creatives in person. I think that is why Musk hates work from home. But if you're a creative type that would be driven "Office Space" mad by having to physically do that shit in person every week WFH is a huge blessing (also, anyone with kids). The fact that I can just ignore 90% of corporate bullshit working from home keeps my sanity.
I’d argue that Microsoft (and Facebook) are both in a third category not covered here because it’s mainly a management skill called “fast following”. Neither of these firms truly disrupted the world with their own innovation but rather, bundled others marketable concepts for mass market very well. That’s certainly a “kind of” innovation but not truly disruptive.
Some say that Apple.(and Steve Jobs specifically) shipped other peoples ideas but I disagree and unusually, they’ve continued to innovate without Steve fairly well.
All of these companies have research divisions and the products you mentioned from Microsoft are results of that work but Microsoft essentially meets existing demand rather than successfully creates it. The list of their products that fit that description seems endless. Same thing with Facebook but NOT true of Apple or Tesla/SpaceX, and it used to not be true of Google (at all) who truly created new demand but has clearly become lost in the wilderness with good, smart, but now seemingly ineffective managers.
Amazon on the other hand stated out balancing an existing, rapidly growing demand (I don’t think they created demand) but then succeeded later in “leaning into” the demand “satisfaction” business and as an infrastructure provider … which, when it comes down to it, was slightly more fast following than a truly creative process. People forget that when they first got online (if it was after 1998), the first place the went was often this mythical “Amazon” where they likely bought something just to see how the WWW might work. Implementations have always been super key to Amazons success. Work with enough software engineers (especially pony-tailed Perl programmers) and you know what I mean.
Zuckerberg made a huge investment in virtual reality, and got weak results in an area that no customers wanted in the first place.
Kling did an amazing piece about a year ago on why innovation escapes large, successful companies due to relative risk aversion at scale. If I can find it, I’ll repost it here.
Big Firms Need to be Sluggish: https://open.substack.com/pub/arnoldkling/p/big-firms-need-to-be-sluggish?r=8068&utm_medium=ios
While I also think MS overpaid for both, I notice a lack of an alternative valuation. $4 $2 $1? Billion.
Had they developed Skype well, Zoom would not have taken off. That’s the MS implementation problem, not the M&A, nor price. Had they made Skype 2 as good as Zoom, it would have been great.
Similarly, MS stupidly wanted windows on a phone, rather than a phone OS that works well with Office, as a third phone choice.
Great examples of management failure to integrate or innovate effectively, partly due to size.
Reminds me of successful skunk works in military development, with some Director protecting the project from financial suits & auditors, so as to be successful.
You don't think MSFT is doing pretty well in Cloud stuff?
The distinction doesn't seem important to me but ok.
I don't even see the relevance of these questions, much less what answers I'm supposed to find.