|
|
|
|
|
|
by dnautics
3198 days ago
|
|
|
i think it's very unlikely. Keep in mind the genesis of the current round of AI algorithms have largely been enabled by 1) the easy access to parallel computing (e.g. GPU, clustering solutions), and 2) large amounts of ingestable data. The truly spectacular AGI will emerge when you can do "one shot" learning. With the exception of zero-shot translation (which is a phenomenal result) we haven't seen very much along those lines. |
|
|