3 ms·
> businesses will continue to need humans to differentiate. Honestly, why? If AI actually becomes capable of replacing large sections of the workforce, why wou
by scared_together 4mo ago
> businesses will continue to need humans to differentiate.
Honestly, why? If AI actually becomes capable of replacing large sections of the workforce, why wouldn’t a business composed entirely of AI “employees” outcompete their rivals?
- redwood 4mo agoand why would they need human customers to thrive? They have other machine customers! This is the even more dystopian step two that the essay doesn't explore...
- keiferski 4mo agoI don’t see this happening at all. Even today, a person + AI is vastly better than just an AI. Context is really important.
- jameslk 4mo agoBecause they would all be using the same AI model (in reality a fixed set of them, but let’s say it’s just one for the sake of argument). That isn’t differentiation [0] It’s like if every company hired the same guy named Karl. If everyone is relying on Karl, and Karl is making the same stuff for all these businesses, how is one business going to outcompete another? At that point you need something else to drive differentiation. Branding, strategic partnerships, patents, IP, influencer endorsements, real estate, government licensing, etc. These are either influenced, controlled, or regulated by humans at the end of the day. At the very least you’ll need humans aligning the models for human needs. Humans are the ones being served, they’re the taste makers 0. https://www.mckinsey.com/capabilities/quantumblack/our-insights/from-ai-table-stakes-to-ai-advantage-building-competitive-moats https://www.mckinsey.com/capabilities/quantumblack/our-insig...
- chii 4mo ago> would all be using the same AI model i think this is a very large assumption. What if in the future, AI models are as guarded as nuclear weapons? Because why doesn't this argument apply for nuclear weapons, but does for AI?
- wallst07 4mo agoBecause AI models can create other ones and the weights/training data is what models. So you can hoard data, but not the models.
- ijk 4mo agoHence the drive to control the compute hardware. If your competitors can't run inference at competitive scale and cost, they can't challenge you.