The Cycle of AI

Progress is made in AI, people lose their shit thinking Jarvis-level AGI is just around the corner, the singularity gets canceled, then nothing is AI, until more progress is made. Rinse & repeat.

Want to get notified about new posts? Join the mailing list and follow on X/Twitter.

This cycle repeats every couple decades:

Progress is made in AI, people lose their shit thinking Jarvis-level AGI is just around the corner, the singularity gets canceled, then nothing is AI, until more progress is made. Rinse & repeat.

~1960: If it searches through a maze of actions to reach a goal, it’s AI!

~1970: That’s not AI, that’s just reasoning as search.

~1980: If it encodes domain knowledge as if-then rules and chains them together to make conclusions, it’s AI!

~1990: That’s not AI, that’s just expert systems.

~2000: If it uses statistics to learn patterns from data, that’s AI!

~2010: That’s not AI, that’s just machine learning. Unless it’s a neural net – in that case, it’s AI!

~2020: That’s not AI, that’s just deep learning. Unless it’s a transformer – in that case, it’s AI!

Truth is, all of this is AI. You can call it all AI because that’s what it is. Ways to make computers make intelligent decisions.

It’s just not AGI. But that doesn’t mean it’s not AI.

Want to get notified about new posts? Join the mailing list and follow on X/Twitter.