This episode, we debate the cultural acceptance of AI inference being good enough to replace us. Off the back of developing lightning fast inference hardware, should we care that outputs are 90% right rather than the full 100?
The last decade has shown us that what we see, learn, know even… depends on who built the software that we use. Does it matter who builds the hardware too? What if they’re the same entity?
We go heavy with the social observations in this one, so please, enjoy.
Sean’s links:
Ben’s links:
Otherwise, you can find Promise on your favorite podcast platforms:
Share this post