Machina Sapiens

Can machines think? This troubling question, posed by Alan Turing in 1950, has perhaps been answered: today we can converse with a computer without being able to distinguish it from a human being.

Machines can pass university exams and programme other computers. ChatGPT, Bard and other “language models” have proved proficient at performing tasks far beyond their creators’ initial expectations, and we still do not know why. Trained simply to predict missing words in a text, such models have gained an understanding of the world and language that makes them capable of reasoning, planning, solving problems, as well as conversing almost flawlessly. Is this the secret of knowledge, and is it now in the hands of our creations? Perhaps we are no longer alone. And as we try to figure out how to share these powers with the “aliens” who now work at our side, we can wonder what else they may learn tomorrow. Are we approaching a critical threshold beyond which machines will attain superhuman performance?

Following on from The Shortcut, Nello Cristianini has authored another brilliant book - written like a gripping thriller - explaining the ideas behind a technology destined to change the world. If our worst terror has always stemmed from fear of the unknown, the cure, since time immemorial, is knowledge.