Why does the VC model of seeds, saplings, trees not work within an organization?

Organizations are about consensus.

Drucker: to have innovation internally you have to wall them off and cannot judge by the standards of the existing organization, and only once they could stand on their own you try to integrate back in.

Rule of thumb: don't aim for consensus (which leads to blandness), aim for everyone agreeing that this investment won't kill the org, and one person thinks it could be great.

If you try to do seed bets internally they need to have freedom to attack the host.

Executives often want to pull weeds as they sprout up to not cause problems or eat away at the coherence.

An anti-evolutionary pressure!

Kills the antifragility of the system.

Cisco has a whole spin in strategy that created external startups that would be re-integrated if they went well.

It was killed off partially because the winning teams undermined social trust by being "unfair"

Is it possible that any company that had a successful venture farm internally would tear itself apart?

For a company to cohere, it has to see itself as one thing.

For an evolutionary system to work, the things must be in competition with each other, to see each other as separate.

Large organizations are a tenuous balance between being "one thing" and "a swarm of competing things".