"AI could write 90% of the code within a short timeframe, dramatically increasing the productivity of remaining developers" Anthropic CEO Dario Amodei is quoted saying [1] .
"Not so fast" says Niall Kishtainy in his contribution published on the IMF website [2]: "Artificial intelligence may rival steam, electricity, and computing—but history suggests its full economic impact will take time" .
He prominently refers to the infamous "Solow Paradox", which Robert Solow summarized in 1987 as follows [3]: "You can see the computer age everywhere but in the productivity statistics."
And indeed, despite massive investments in information technology (IT), it took a quarter of a century after the observed decline in productivity growth during the 1970s and 1980s before the impact on productivity became visible in the late 1990s – when the internet began its triumphant march.
Clearly, widespread adoption in the broader economy, accompanied by necessary organizational and personnel changes, takes time.
So, we have to wait quite a while again – or do we?
There is an alternative to technological debt and slow, costly software development.
To bridge the gap, we should look to Southeast Asia: Vietnam, the rising tiger. Here, the strong focus on education is paying off, a dynamic software industry has emerged and is growing rapidly, costs are still moderate, one can benefit from China's speed without being entangled in geopolitical tensions, and – most importantly – one can expect innovative contributions, not just obedient implementation.
And once AI has actually permeated all business processes, including areas of software development that are not purely coding, Vietnam will still be the right choice for you. Vietnamese young talents learn fast and are eager to apply various flavours of AI, leading to a situation where this south-east Asian country already now has become one of the regions making use of AI the most.
So, if you need support, try Vietnam.
Feel free to ask us.
[1] Greenblatt, R. (2025, October 22). Is 90% of code at Anthropic being written by AIs? LessWrong.
- https://www.lesswrong.com/posts/prSnGGAgfWtZexYLp/is-90-of-code-at anthropic-being-written-by-ais
- A community-authored discussion on LessWrong examining claims around AI-generated code at Anthropic, including an analysis of the assertion that 90 % of code is written by AI and commentary on how such metrics can be interpreted. It engages with public statements by industry leaders about AI productivity and reflects on productivity measurement nuances in software engineering.
[2] Kishtainy, N. (2025, December). A new industrial revolution? Finance & Development(IMF Magazine). International Monetary Fund.
- https://www.imf.org/en/publications/fandd/issues/2025/12/a-new-industrial revolution-niall-kishtainy
- This IMF Finance & Development article situates artificial intelligence within the context of historical industrial revolutions, exploring the long-term economic impacts of AI on productivity, growth, and structural change. It provides a high-level analytical perspective on how transformative technologies historically diffuse into economies and why the full effects take time to manifest.
[3] Freund, L. (2018, October 2). The productivity paradox – a survey. Dezernat Zukunft.
- https://dezernatzukunft.org/en/the-productivity-paradox-a-survey-2/
- A thorough online survey of the productivity paradox (often linked to Solow’s observation that technological advances do not always immediately translate into measurable productivity gains), outlining competing hypotheses—such as mismeasurement, structural changes, and sectoral reallocation—that help explain apparent disconnects between innovation and macro productivity statistics.
No comments:
Post a Comment