Horowitz founded his VC firm in 2009, and wanted to fix 2 dated ideas in venture capital

In the past, VCs provided money but no service to entrepreneurs. (Note he means investors can only provide ideas that generally work, but should NEVER run the company)
Industry believed there were only 15 new companies a year to hit over 100M in revenue, but a16z predicted it would be 200 companies because of software.

scaling firms

You need centralized control so your firm makes decisions without grid lock.

The firm splits into smaller groups to maintain “high-fidelity” conversations, which works best with group of seven.

To prove their model to skeptics:

invested 25% of a 300M fund into a skype buyout
eBay owned Skype but not the IP. a16z bet that the founders connection to the product would stop them from shutting it down

Their fast decision making and actually knowing the founders made it a lot easier

the importance of Network Effect

Designing the firm as a network effect driven entity. Value of a network increases exponentially (n-squared) as it grows. two strategies:

Reinvesting Fees: Instead of taking large salaries, they spent the money hiring people to build relationships with engineers, executives, and corporations.
The HP Briefing Center “Hack”: To gain access to large corporations, they leveraged their connections at HP Enterprise Briefing Center. They would track which major companies were visiting HP and invite those executives to their own briefing center to meet their startups

When a16z introduced its service-based model, established VC hates it.

Buying companies and running them depend on efficiency and firing people. They focus on existing ideas and growth

but VCs only focus on funding new ideas and growth.
But this private equity model is still a good model as AI gets better and make it more efficient.

You cannot catch a competitor by throwing money at a problem

because massive amounts of GPU/Data can catch a 2 year lead
there is unlimited demand for AI products, since it works better than everything else

you don’t need to climb the corporate ladder (the people that climbed it will struggle)
there is unlimited money for good ideas (good idea = something world needs that doesn’t exist)
start a company only by solving a small personal problem over large ambitions

AI is a toolset to master. Apply it to whatever you like. Prove you are modern.

Do not listen to friends for career advice

Stop dreaming and build something

solve a problem sized to YOU. (dropbox started because of hatred for USBs)
subsequently increase scope in every project
the path to a company / frontier lab is never a straight line. its always an accident, starting with smaller boring easy projects, or even someone developing skills (zuckerberg)

many talented, well-funded teams fall apart within 6 to 12 months because they fail to set the right cultural standards from the beginning. Find the correct “cultural initialization”.

Bitcoin failed because it had no voice in Biden era - Gov required permission for any GPU sale

Now AI will grow because of backing from government.

The founder was noted as a ‘best distributed systems prof in academia’

Wall street only follows narratives

it is a voting machine in the short term (follows narratives over fact)
real earnings override negative narrative once market realizes AI cannot replace real-world relationships
So building relationships will be even more valuable as people start dropping them because of AI

Wall street wrongly thinks its a SaaS apocalypse.

Navan is software that manages travel for business (largest expense)
Navan needed global supply chain relationships with every country/hotel rather than code
It helps out travel managers (specific role, dedicated to managing travel for company)
Wall street thought Navan is garbo since its Saas