|
|
|
|
|
|
by visarga
336 days ago
|
|
|
> The ease of training AI to solve a task is proportional to how verifiable the task is. All tasks that are possible to solve and easy to verify will be solved by AI. I've been saying this for years. Especially related to predictions like "AGI in N years".. no, it will come a different speeds per domain. All proportional to the scale of verification. Verifiable domains are usually math and code. Games also fit the bill. But for real world tasks there is the 1B people using AI, we verify, the signal is there implicit in the chat logs. |
|
|