There are many possible futures, but the one that will come to pass is the one we collectively allow to happen. To a degree, of course.
A future that I like to imagine is one where LLM scaling laws hit a plateau. Plateau-level LLMs become commodities. They get distilled into small sizes. Everybody can have plateau-level local LLMs. People discover that it’s better to be a local-powered independent centaur than to be a reverse centaur for a mega-corp.
In the West, decline of living standards and continued decay of institutions. Probably very few organizations actually benefit from LLMs, mostly bankers.
But what if a decentrilized internet came along that offerd 90/10 split 90% gos to user (,node) and 10% gos to upkeep and maintenance of decentrilized Internet. That alone would change the dynamics of all industrys.
I believe there will be a massive shift from a lot of white-collar jobs to blue-collar jobs, as these jobs will be the ones most affected by AI. I believe Western countries will be poorer than they are now, and countries in Asia and Africa will become richer as more money will flow into their manufacturing and Agricultural industries.
Greater wealth inequality. More authoritarian policies as climate refugees from the global south put pressure on classically liberal governments.
Less freedom of movement. Most of the Western world living in a surveillance state.
Climate breakdown manifesting itself with bigger and more frequent wildfires, floods and hurricanes. Potentially the first sign of the AMOC collapse.
An economy begining to become untethered from labor. Feudal conditions from a rent seeking class of supranational corporations . Social stratification across ownership: most people don't own anything at all.
You can predict by using one of several techniques:
1. Naive prediction: it will be similar to now
2. Linear regression: look at several parameters associated with LLM (speed, quality, accuracy, etc.) and create linear regressions from these parameters
2a. Polynomial regression: same as above, but fit a polynomial
3. Group prediction: select 20 friends, ask them to make predictions, then find the average(s)
...and many more with increasing levels of complexity
A future that I like to imagine is one where LLM scaling laws hit a plateau. Plateau-level LLMs become commodities. They get distilled into small sizes. Everybody can have plateau-level local LLMs. People discover that it’s better to be a local-powered independent centaur than to be a reverse centaur for a mega-corp.
Elsewhere? who knows
Less freedom of movement. Most of the Western world living in a surveillance state.
Climate breakdown manifesting itself with bigger and more frequent wildfires, floods and hurricanes. Potentially the first sign of the AMOC collapse.
An economy begining to become untethered from labor. Feudal conditions from a rent seeking class of supranational corporations . Social stratification across ownership: most people don't own anything at all.
1. Naive prediction: it will be similar to now
2. Linear regression: look at several parameters associated with LLM (speed, quality, accuracy, etc.) and create linear regressions from these parameters
2a. Polynomial regression: same as above, but fit a polynomial
3. Group prediction: select 20 friends, ask them to make predictions, then find the average(s)
...and many more with increasing levels of complexity
And for LLM & AI, I predict that the tech bros overestimate the pace of change, while the average joe underestimate the new capabilities.