In 1863, philosopher Samuel Butler speculated that we would be the architects of our own eclipse: “We are ourselves creating our own successors…in the course of ages we shall find ourselves the inferior race.” He may have been imagining mechanical looms and steam engines rather than neural networks, but the concern lingers.

Science fiction of the 20th century gave us visions of malevolent, megalomaniacal artificial intelligences overthrowing humanity. However, large language models have shown no intent toward power, only an aptitude for anticipating patterns.

The risk, as it turns out, has not been physical, but existential.

Instead of calamity, we got competence. Large language models can provide the skills of a junior researcher or software engineer in a fraction of the time. Posing even more of an existential threat, AI has brought about the automation of

© Psychology Today