Tyler’s concern is that bottlenecks (maybe energy, land, social trust, governance, etc) would emerge like hard stone after the soft stone's been eroded and slow multi-factor growth. But this assumes that artificial intelligence interacts with bottlenecks passively or linearly. If intelligence can reshape the system--finding shortcuts, parallelizing tasks, and manipulating human institutions--it becomes much harder to imagine bottlenecks being constraints to the same degree that they've been historically. There are limits of course. Some constraints are impossible to compress beyond a point. When the hard stone has been eroded, the very hardest stone is easier to trip on. However, a sufficiently strategic intelligence could also anticipate such constraints earlier and begin solving them proactively, or reorganize systems to minimize their impact.
Guardrails / governance are friction like any other except that they're not self-enforcing. When rules stand in the way of what you want, the effectiveness of those rules entirely depends on the motivation and capabilities of you vs those of the rule enforcer.
We've operated so long with human motivations & capabilities being dominant that we mistake that as the entire possible range. It isn't.
What a great interview, it’s great to hear someone taking the other side of AI’s explosive growth predictions, who is in Gary Marcus, and comes at it from a different perspective. Way to go fellas
Your podcast is super fun, so rare I. This space. It's cutting edge/fun/positive/informing. And I have a quantum /next gen compute company called Entanglement, and that world is bland. You spice it up
The solution is simple and is left as an exercise for the reader.
Tyler’s concern is that bottlenecks (maybe energy, land, social trust, governance, etc) would emerge like hard stone after the soft stone's been eroded and slow multi-factor growth. But this assumes that artificial intelligence interacts with bottlenecks passively or linearly. If intelligence can reshape the system--finding shortcuts, parallelizing tasks, and manipulating human institutions--it becomes much harder to imagine bottlenecks being constraints to the same degree that they've been historically. There are limits of course. Some constraints are impossible to compress beyond a point. When the hard stone has been eroded, the very hardest stone is easier to trip on. However, a sufficiently strategic intelligence could also anticipate such constraints earlier and begin solving them proactively, or reorganize systems to minimize their impact.
Guardrails, then, emerge as a bottleneck if we assume a level of intelligence that circumvents such constraints.
Guardrails / governance are friction like any other except that they're not self-enforcing. When rules stand in the way of what you want, the effectiveness of those rules entirely depends on the motivation and capabilities of you vs those of the rule enforcer.
We've operated so long with human motivations & capabilities being dominant that we mistake that as the entire possible range. It isn't.
Saving it, despite Tyler looking way exactly like Modi here. ☺️
What a great interview, it’s great to hear someone taking the other side of AI’s explosive growth predictions, who is in Gary Marcus, and comes at it from a different perspective. Way to go fellas
Your podcast is super fun, so rare I. This space. It's cutting edge/fun/positive/informing. And I have a quantum /next gen compute company called Entanglement, and that world is bland. You spice it up
what is the context of this? Which group was Tyler speaking to?
Just 10 minutes into it and already loving it 💚 🥃