Recently, two articles have been very popular.

Charles I. Jones (Stanford GSB + NBER, a top economist specializing in long-term economic growth, writing for the academic community) in his paper "A.I. and Our Economic Future" offers the judgment: AI will significantly change the economy, but the process is likely to be gradual.
His framework emphasizes "weak links"—economic growth depends on the most difficult-to-automate parts of the system, so even if AI boosts the efficiency of many tasks by 10x/100x, GDP may not immediately surge in tandem.
Growth will come, but slower than market expectations.
On the other hand, Citrini Research (a U.S. macro hedge fund research organization targeting traders/fund managers; the article mentions research analyst and macro strategist Alap Shah) in "The 2028 Global Intelligence Crisis" (subtitle: "A Financial History Thought Experiment from the Future") uses a completely different time scale: looking at asset pricing, employment shocks, credit transmission, and liquidity risks over the next 24 months.
Its core warning is that if AI replacement speeds outpace the buffers of labor markets, policies, and financial systems, a phenomenon called "ghost GDP" may occur—booked output rises, but income does not flow to ordinary consumers. Demand weakens instead, potentially transmitting from white-collar employment pressures to SaaS revenue, private credit, insurance capital, and ultimately the fragility of the entire financial system.
What’s most worth noting when reading these two articles together is that they discuss the same reality but target different audiences and answer different questions.
Jones writes for economists, concerned with long-term growth paths and equilibrium over 20–50 years; Citrini writes for market participants, concerned with how corporate profits, asset prices, and credit risks will reprice over the next few quarters to two years.
So, their conclusions may seem opposite, but they are more like an overlay of "long-term growth logic" and "short-term shock logic."
The real disagreement isn’t whether "AI will change the distribution pattern," but rather "how fast." Both sides agree that AI will increase capital returns and compress certain labor income shares; the debate is whether this is a gradual structural change over 50 years or a sharp re-pricing within 5 years.
If it’s the former, society still has time to buffer through retraining, policy shifts, and new industry absorption; if it’s the latter, many institutions (employment, credit, social security, asset valuation) may not react in time.
This is also why there’s a potential blind spot in today’s market: many people only price in "AI efficiency gains → corporate profit increases → stock price rises," but do not seriously consider "where does the demand come from." Those replaced by AI are precisely part of consumer demand.
Rational decisions at the corporate level (layoffs, efficiency improvements, increased AI investment) combined could turn into a macro-level prisoner’s dilemma.
"A.I. and Our Economic Future" discusses "AI will bring higher growth, but it takes time."
"The 2028 Global Intelligence Crisis" discusses "if time is compressed, growth has not yet spread, the crisis may arrive first."
So, this is not a matter of right or wrong, but a matter of timing. What truly needs reevaluation is not just AI stocks, but the entire set of income distribution, asset pricing, credit models, and social security systems built on the assumption that "human cognitive labor remains long-term scarce."