A little more economics and a little less statistics would've helped his argument immensely.
Entrepreneurship flourishes when there are diseconomies of scale and dies when there are economies of scale. It's not hard to see why: if things get cheaper as you get bigger, then the first company to get big will gobble up all the smaller companies, and you end up with a monopoly or oligopoly. If things get more expensive as you get bigger, then no firm will grow larger than a certain limit, and after a certain point the most productive employees will leave and start their own businesses, because they can make more money that way. (If costs stay roughly the same, as in most professional services, the industry trends toward many small firms because of the psychological benefits of working for yourself.)
Software is the perfect example of a business with diseconomies of scale. Every experienced software engineer knows that adding people to a project makes it go slower. The only reason why a large software company might be more efficient than a small one is if there's an additional factor of production that has large economies of scale: for example, data centers (Google/Yahoo), logistics & warehousing (Amazon.com), buyer eyeballs (E-bay, Facebook), or retailing (Microsoft, Borland, Ashton-Tate, and all the boxed-software manufacturers of the 80s). Other than that, there's no economic reason why big software companies should exist.
Big companies buy little companies when there's a market that they want to get into and apply their economies of scale to. They can't develop the product in-house, because small teams of programmers are more productive than large teams of programmers and so the startup would continually outpace them. But the resulting merger would be more economically efficient, because the startup could take advantage of the large company's datacenters & warehouses. So they trade money for product; the small company's founders get rich, while the big company buys a product that can make more profit under the large firm than as an independent firm.
It's probably also worth mentioning that technology developments can fundamentally change the structure of an industry. Intel single-handedly destroyed the minicomputer industry, because those firms had relied on their economies of scale in manufacturing to justify their size, and Intel brought greater economies of scale to the thousands of ISVs who could then produce software more cheaply than the big proprietary vendors. The RIAA held a distribution lock on the music industry for years, but when the Internet came along and made distributing music free, power shifted to the independent bands who have significant diseconomies of scale. The Internet looked like it was heading towards a universe of big firms with large datacenters, but then LiveJournal and Flickr published their scalability approaches and made it possible for everyone to scale out for free.
Entrepreneurship flourishes when there are diseconomies of scale and dies when there are economies of scale. It's not hard to see why: if things get cheaper as you get bigger, then the first company to get big will gobble up all the smaller companies, and you end up with a monopoly or oligopoly. If things get more expensive as you get bigger, then no firm will grow larger than a certain limit, and after a certain point the most productive employees will leave and start their own businesses, because they can make more money that way. (If costs stay roughly the same, as in most professional services, the industry trends toward many small firms because of the psychological benefits of working for yourself.)
Software is the perfect example of a business with diseconomies of scale. Every experienced software engineer knows that adding people to a project makes it go slower. The only reason why a large software company might be more efficient than a small one is if there's an additional factor of production that has large economies of scale: for example, data centers (Google/Yahoo), logistics & warehousing (Amazon.com), buyer eyeballs (E-bay, Facebook), or retailing (Microsoft, Borland, Ashton-Tate, and all the boxed-software manufacturers of the 80s). Other than that, there's no economic reason why big software companies should exist.
Big companies buy little companies when there's a market that they want to get into and apply their economies of scale to. They can't develop the product in-house, because small teams of programmers are more productive than large teams of programmers and so the startup would continually outpace them. But the resulting merger would be more economically efficient, because the startup could take advantage of the large company's datacenters & warehouses. So they trade money for product; the small company's founders get rich, while the big company buys a product that can make more profit under the large firm than as an independent firm.
It's probably also worth mentioning that technology developments can fundamentally change the structure of an industry. Intel single-handedly destroyed the minicomputer industry, because those firms had relied on their economies of scale in manufacturing to justify their size, and Intel brought greater economies of scale to the thousands of ISVs who could then produce software more cheaply than the big proprietary vendors. The RIAA held a distribution lock on the music industry for years, but when the Internet came along and made distributing music free, power shifted to the independent bands who have significant diseconomies of scale. The Internet looked like it was heading towards a universe of big firms with large datacenters, but then LiveJournal and Flickr published their scalability approaches and made it possible for everyone to scale out for free.