• Developed original prototype of AdSense, in spite of Marissa Meyer's naysay • Suggested "Don't be Evil" motto • (Advice on Advice) Limited Life Experience + Overgeneralization = Advice (Note: Paul hasn't founded a startup) • If someone says "That's impossible", you should understand it as "According to my very limited experience and narrow understanding of reality, that's very unlikely" • Quit your job! (...if you're at a boring place) • Secret rule for success: redefine success. "New definition: Learned a lot" • Attempt things with uncertain outcomes... but things you can learn from • Startups are a great place to learn... and you have more opportunities to do things that you aren't really _qualified_ to do + Build something fundamentally new + Not MP3 player (clouds), think "device which holds all of my music, fits in my pocket, and easily syncs with my computer" + How can you tell that a product is fundamentally different? By how people use it + Try adding the words "that actually works" + Try making order of magnitude changes • Disks aren't really random access (compute seeks per dollar) • 100,000 disks for 1 bank of DRAM • Think of disks as sequential access (tape) • Two kinds of data: big data (photos, video, mp3) --> Disk, probably AMZN S3, small data (tags, name, meta data)--> DRAM • DB updates must also avoid causing seeks. How? Updates go to transaction log. All writes are sequential. • (Do databases do this? Google has their own tech)
Q&A: Spent a lot of time overcoming differences in Web Search and (G)Mail -- infrastructure requirements. Left Google a year ago, felt like "it was capturing me".
http://www.zenter.com/ed/view.html?id=226
• 4/8 successful from first Y Combinator batch (ex. Reddit, Loopt) • Rich: 25% long term? • How many of them (Y Combinator founders) would trade for job in a cube? 0% • Why don't more people do it? PG gets hundreds not thousands of applications • 1. Too young (maybe. median age in the world is 27) • (What makes someone seem old? Kid's have flake reflex. How do they response to challenges from adult? If they submit or rebel => Kid) • 2. Too inexperienced (do it anyway.) • 3. Not determined enough (a problem) • 4. Not smart enough (problem? not if you worry about it. not smart enough? write enterprise software.) • 5. Don't understand business (not the hard part anyway. companies buy startups for strategic value) • 6. No cofounder (a real problem) • 7. No idea (no problem!) • 8. No room for more startups (a fallacy) • 9. Family to support (a real problem. Try starting a consulting firm and eventually turning it into a product. Likelihood of success is very low, but you'll always have income.) • 10. Independently wealthy (a legitimate excuse) • 11. Don't fence me in (ok) • 12. Need for structure (i.e. you need someone to tell you waht to do. don't start a startup) • 13. Fear of uncertainty (i can fix that [you will fail]) • 14. Don't realize what you're avoiding (23yo know what they're avoiding, 21yo don't. summer job at a tech company isn't a real job. ask your friends) • 15. Parents want you to be a doctor (all you guys who are laughing, you're not Chinese (Ed: I'm Chinese :) -- treat as a feature request) • 16. A job is the default (relatively new idea, last hundred years. similar change right now? a lot more will do startups, even if the idea seems weird. job as serfdom. how do you get food if you move away from your serfdom to the city? well there are bagels outside ;)
• {Your manager asks you into his office. He asks you to sit, and you sit like a dog. Shows picture of Marissa Meyer (of Google), you have to get permission from this person to release a product.}
• listen to customers. know their top 5 complaints and top 5 fav features. read discussion boards and email lists. foster communication among your own customers • stack rank the top problems core to your success, and get your top people operationally focused there • make frugality and profitability part of your culture • move quickly. make decisions fast! avoid committees, avoid 12-month development projects • have a strong CEO. • F O C U S • hire great people. put hiring and recruiting at the *top* of your priority list (watch the apprentice to practice your evaluation skills.) [Ed: should there be people problem sets? YouTube videos of people talking about themselves, or interviewing, then folks discuss afterwards]
• Don't be distracted by press. PR success not highly correlated to actual access. Don't make decisions just to get press. Top goals of PR are recruiting and strategic partnership • Don't take your company culture for granted • Don't be greedy in business negotions. Heck, let company get bigger piece of the pie, in return for speed • Don't ignore your gut feel about an employee or candidate • Don't: forget to have fun!