PenLUG Talk: Fast, Cheap, and Out of Control
Download as PPT, PDF3 likes861 views
The document summarizes how Silicon Valley became a hub for technology innovation again after the dot-com bust of the early 2000s. Key factors included the availability of cheap bandwidth, open source software, and low barriers to launching startups with nearly free hosting, storage, and computing resources. This enabled a vast number of experiments and niche companies to be created with very little capital required compared to the dot-com era.
PenLUG Talk: Fast, Cheap, and Out of Control
- 1. Fast, Cheap, and Out of Control How Silicon Valley Got Its Groove Back A completely uncalled-for talk by David Weekly for PenLUG on 11/9/06 Given on a Windows laptop with PowerPoint because I am tactless.
- 2. Why Are We Here? Great Weather Lots of Geeks Full of Possibility (& Hard Work)
- 3. Why Did Silicon Valley Happen? Brain concentration: Stanford & Berkeley Love-fest of Ideas 70’s Hippy Children of 50’s Nerds Homebrew Computer Club (75-77) West Coast Computer Faire (77-83) The beginning of computer affordability
- 4. Money? Wasn’t the big thought, was about sharing Software hadn’t made it big Hardware was either huge (IBM) or hobbyist (Altair) Just make the damn thing, it’ll be cool.
- 5. dotCom Era Widescale computer deployment seeded by the First Wave meets access (AOL). Starts With 3DO IPO (1993) First IPO of a zero-product company Public ownership drive => liquidity Big payouts with hype drove rapid VC dev.
- 6. dotCom Era Both beginning and end marked by public market involvement. Investment was required for innovation Hardware, software, hosting = expensive! No capital, no product.
- 7. dotCom Era Proprietary Thinking NDA proliferation Closed source Why? Crazy availability of capital plus Large, quick liquidity events makes people idea hoarders …and these were MBAs, not hippy geeks
- 8. IPOs Allow for public ownership = good De facto pyramid scheme = bad Enron => Sarbanes-Oxley = dead Post 2002, no easy IPOs $2m/year, personal liability
- 9. post dotCom No more quick money => flakes gone ...but geeks stayed! Hardware continues commoditization Open Source™ software steps up Huge ’99 telco investments => cheap bandwidth
- 10. What’s Amazing Now? Execution is cheap Nearly free storage, CPU, bandwidth Free, Open Source LAMP stack Execution is easy Mature development tools & libraries There was no good answer before Browser JS = new platform for innovation Monetization is possible AdSense, PayPal, CC processing
- 11. What’s Different? Homebrew Computer Club (1975) = New Ideas Every Two Weeks SuperHappyFunHouse (2005) = New Implementations In Six Hours
- 12. What’s Different? West Coast Computer Faire (1977) = Huge Annual Commercial Showcase located in California. BarCamp (2005) = Huge Continuous Showcase, global.
- 13. (Unimplemented) Talk Is Cheap Low friction launches mean lots of experimentation & iteration. Lots of terrible ideas, but ... Cheap, fast failures. Very cost-effective exploration No VC required.
- 14. What If… You can’t IPO and make $100m? You’re not going to get bought for $100m? You can make $2m/year with two friends? “Lifestyle Company” (derogatory)
- 15. Talent Black Hole = You Win No free talent Giant Holes = Google, Intel, Yahoo, NVidia The Failure of Outsourcing A+ coders are not trawling for jobs. They’re probably hired and happy. Money != de facto ability to compete … so don’t freak out about funded competition
- 16. Why Was VC Great? Invest in 20 companies 16 will flop badly 3 will do alright and make a few bucks 1 will make you a huge profit Sucked for an entrepreneur! Try it starting two companies, spending five years each, and you’ll probably still fail. … but now you can launch 20 companies. Launch one this weekend! (Not kidding!)
- 17. Where This Is Going Huge ecosystems around everything Calendars GIS / Maps Text Collaboration: Email, Blogs, Wikis Realtime: IM, SMS Social & business networking Media: Pictures / Music / Movies A million niches, a million companies
- 18. No Opportunity Too Small A $800,000 / year global market is just fine If no operating costs And almost no employees.
- 19. Enabling The Constellation Structured data interchange APIs w/XML, JSON RSS, Atom, webcal Embeddability Google Gadgets, Bitty Browser Cheap Hosted Services GMail, Amazon S3, PBwiki, Wordpress
- 20. Still Missing! Common Data Repositories Easy joins of GIS/calendar/DB data Obvious Multisite Authentication OpenID? MS / Yahoo / Google / AOL logins? Kerberos? Automated API discovery & documentation Seamless idea => cluster tools
- 21. Long Term Challenges Everything To Zero Operations so cheap, can be charity-supported (CCCP) or done for market (Google) May prove hard to compete with Open Source + free hosting. Scalability Not all problems parallelize well CPU bottleneck
- 22. Things For Geeks To Avoid Intellectual Masturbation Writing your own DB / OS / language Overfeaturificationising Things Not as simple as it could possibly be. Perfectionism & Elegance Waiting to launch until the seventh rewrite
- 23. Lessons Leverage Others’ Work Use Open Source Use knowledge from others see Brad Fitzpatrick’s amazing LiveJournal slides! Don’t build it yourself. Make it sustainable Do fun things; find out how to make one of them pay. Be cheap. Use commodity everything. I know OCAML rocks, but damn it, use PHP. Release Early & Often Get over that perfectionist streak. Your users know their needs better than you do. Make Structured I/O Easy w/standards & APIs Just do it.