A Developer on AI: Technically Promising, Societally Bleak — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- The author expresses surprise that neural networks and LLMs work so well despite generating tokens one after another with no internal planning or reasoning algorithm — yet planning and reasoning still emerge from the process.
- Yudkowsky is cited arguing that a superintelligent AI will inevitably destroy humanity; the author notes nobody seriously tries to sandbox AIs because they are too useful with access.
- AI companies' crawlers and agents are compared to a locust plague that descends on open wikis and forums, making it increasingly difficult to sustain open web communities.
- Artists are described as suffering from AI-slop competition that makes it harder to earn a living from creative work that always required sacrifice.
- AI companies are criticized for environmental destruction through greedy resource consumption, framed within broader anger at political systems failing to rein in the power of capital and billionaires.
- Software development was radically transformed in 2026, and the author feels happy to live through this age of discovery, able to generate code, images, and music they otherwise could not have created.
- Mistral served as reviewer for the piece, though the author typed every word themselves; the conclusion weighs the good against the bad as positive on a technological level but bleak on a societal level.
Why it matters: The essay crystallizes a tension many in tech now articulate: AI tools have made individual creative work — code, images, music — far more accessible, while AI companies' unchecked crawlers, resource consumption, and political influence threaten open web communities and democratic oversight. The author's split verdict — positive technologically, bleak societally — mirrors a debate the industry has not yet resolved.
Ask SkimNews
