Thinking will become a hobby

blog dds: 2026-10-08 — Thinking will become a hobby

Logo row

Diomidis Spinellis blog

Title row

Thinking will become a hobby

Search button

Search

Modal search window

Modal content

#### Search the blog

See https://cse.google.com/cse/

Close

Content / Contents row

Left content

These lines are very difficult for me to write because I feel like digging my own grave. AI systems are improving by leaps and bounds. From the early 2022 fumbling release of ChatGPT, which couldn’t correctly count the number of a’s in my name (it claimed two), we reached OpenAI’s release of math papers where an internal AI model solved many hundreds of previously open math problems. Aside from philosophical arguments, practical claims regarding the superiority of human thought, on grounds such as hallucinations or bias, increasingly appear to me like rearguard defence. For the bulk of today’s cognitive work, the writing’s on the wall: AI will take over.

The three factors driving AI’s dominance in the modern work environment are better performance, lower cost, and higher availability. The performance of AI models is rising spectacularly on all fronts: perception, knowledge processing and creation, reasoning, judgement, and autonomous action. In parallel, human proficiency in key information-processing skills, low on average among adults , is even declining among students . Consequently, AI systems are performing better than humans in ever-wider range of tasks.

AI’s cost is also falling due to huge investments and scientific progress. Successive generations of AI models offer superior performance at lower cost, with older models retained only for backward compatibility. As a result, it is making financial sense to replace or augment human labour with AI systems across an increasingly broad range of activities.

Finally, and most obviously, AI systems can operate around the clock, without the need for rest or breaks. They can also help alleviate labour shortages arising from demographic trends or skills gaps. Meanwhile, trillions of dollars in projected data centre investment through 2030 are set to bring vast additional AI processing capacity online.

As happened in past technological revolutions, there will undoubtedly be fierce resistance to the replacement of cognitive work with AI. This time it will be fiercer because, unlike previous revolutions, which transformed manual labour, AI challenges knowledge workers, whose specialised expertise, control over information, and decision-making authority afford them greater economic and organisational power. Governments will try to clip AI’s wings through regulation and taxation. Such actions didn’t quell past revolutions so there’s little reason to believe they’ll work now.

So what’s the fate of human thought? My take is that cognitive work will diminish. Thought will increasingly become a leisure pursuit, as hunting did following animal domestication, calligraphy after typography, woodworking after industrialisation, and horse riding after motorisation. Machines transformed much of physical activity from necessity into choice; similarly, AI will decouple human thinking from its economic purpose.

I’ll end with my thoughts on how to prepare for and live through this transition. My take is that building this AI-driven world, riding through the bumps along the road, and preparing to live a fulfilling and satisfying life in an environment where AI dominates cognitive work will be hugely demanding, stressful, and intellectually challenging. Such is the nature of revolutions. Consequently, we must keep pace with and capitalise on AI’s advances; be flexible, adaptable, and resilient; and build a purposeful and rewarding life beyond work. My challenge ahead is to move from making a living by thinking, to making thinking part of a life worth living.

COMMENTS

Bluesky post

Post

Mastodon post

Toot!

Twitter button

Tweet

Facebook button

Share

Right content

Navigation

blog contents

dds blog

dds home

« A controversial proposal for hallucinated references

panel body

panel

Tagged as

AI

Opinion

Become a Unix command line wizard

Debug like a master

Compute with style

Syndication

This blog is also available as an RSS feed:

Recent posts

Thinking will become a hobby

(2026-10-08)

A controversial proposal for hallucinated references

(2026-08-17)

Blockchain technology in 500 simple words

(2026-08-03)

Barings, Boeing, and Bots

(2026-07-15)

Why agentic AI needs better experts

(2026-07-08)

Documenting AI-generated code commits

(2026-07-02)

Why reviewing AI-generated code is devilishly hard

(2026-05-23)

Empirical software research in the age of AI

(2026-04-13)

Vibe coding toward the incident horizon

(2026-03-02)

An initial analysis of the discovered Unix V4 tape

(2025-12-23)

HR

Modification time

Last modified: Thursday, October 8, 2026 12:42 am

License BEGIN

Creative Commons License logo

Unless otherwise expressly stated, all original material on this page created by Diomidis Spinellis is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License .

License END

container