Lesson 35 · 3 min
Is AGI close? Both sides
Smart, informed people disagree about how soon AGI will come, and it helps to know the main arguments on each side.
People who think AGI is near point to fast progress. In a few years, chatbots went from clumsy to passing hard exams, writing working code and running multi-step tasks. Each new generation beats benchmarks that once looked far off.
People who are more doubtful point to stubborn gaps. Models still make basic mistakes, struggle to learn from experience after training, and can do well on a test without being reliable in real work. Benchmarks can also leak into training data.
Some argue progress will keep going with more data and computing power. Others think new ideas are needed, and that cost, energy and data limits could slow things down.
Nobody knows the date. A sensible stance is to learn the tools now, watch real-world results more than predictions, and be wary of anyone who sounds very certain either way.
Do it now
Hear both sides
- 1. Ask a chatbot for the strongest case on each side using the prompt below.
- 2. Note one argument from each side that you find convincing.
- 3. Write one sentence on what would change your mind.
Give me the three strongest arguments that AGI is coming soon and the three strongest arguments that it is still far away. Be fair to both sides and say where the evidence is weak.
Quick check
Keep going
Next →The knowledge cutoff