Library > Talks

The Internet Paradigm: Talk by Tim O'Reilly

July 2003, at Microsoft UK

Notes summary:

Paradigm shift #1: Hardware

Prior to 1982, hardware ruled.  Then IBM released specs for building PC computers.  Didn't seem that important to the industry at the time, they were 'toys'.  Took a decade to really take off, but along cam Dell, Compaq bought DEC...

Paradigm shift #2: Software

Software was now decoupled from hardware.  Lock-in and competitive advantage moved to software.  IBM had given away the future... to Microsoft.  Hardware became a commodity.

Paradigm shift #3: The Internet

Applications and information decouple from both hardware and software.  Lock-in and competitive advantage moves to the data and customer relationships.  Software becomes the commodity,..  Think Amazon and eBay - the application will stop working without people.  It's the participation age.

Where is Linux really successful?  Not as a traditional operating system.  It's Google, Amazon, Yahoo.  (and you don't get access to their source code, only the APIs, sound familiar?)

3 trends that matter:

1, Software as a commodity:

• Amazon switched from Unix to Intel to save costs (10x saving running Linux on Intel)
• Apache means web serving is not a revenue opportunity
• MySQL threatens to do the same to databases

2. Customisability:

• Software is built for use in delivering services
• Internet-era apps are updated daily, not yearly  (e.g. www.salesforce.com = 12 updates between 1999 and 2003)
• 'Info-ware' - interfaces built with dynamic data, scripting rules

3. Network-enabled collaboration

• 'Ad-hocracy' - people just get together to fix things, distributed internationally (skills, costs, timezones), like-minded devs find each other
• Power shifts from companies to individuals
• Users help build the application

More people have contributed to Amazon than have contributed to Linux... 

Small pieces loosely joined:

• An architecture of participation means that your users help extend your platform
• Interoperability means that one component or service can easily be swapped if a better one comes along (e.g. Google data centre)
• Lock-in occurs because others depend on the benefits from your service, but you are not in control. 

What does this mean for Microsoft?

Got to change at some point, not going to see the same margins as in the past (IBM had to get used to this one).  Could MSN be a big part of the future?  Currently focused on consumer, but what about business?