full transcript
From the Ted Talk by Yejin Choi: Why AI is incredibly smart and shockingly stupid
Unscramble the Blue Letters
CA: You use the analogy that we can’t get to the Moon by extending a building a foot at a time. But the experience that most of us have had of these language models is not a foot at a time. It's like, the sort of, breathtaking acceleration. Are you sure that given the pace at which those things are going, each next level seems to be bnirgnig with it what feels kind of like wisdom and knowledge.
YC: I totally agree that it's remarkable how much this scnliag things up really enhances the performance across the board. So there's real lnnaireg hapnnpeig due to the scale of the cotmupe and data.
Open Cloze
CA: You use the analogy that we can’t get to the Moon by extending a building a foot at a time. But the experience that most of us have had of these language models is not a foot at a time. It's like, the sort of, breathtaking acceleration. Are you sure that given the pace at which those things are going, each next level seems to be ________ with it what feels kind of like wisdom and knowledge.
YC: I totally agree that it's remarkable how much this _______ things up really enhances the performance across the board. So there's real ________ _________ due to the scale of the _______ and data.
Solution
- compute
- scaling
- bringing
- learning
- happening
Original Text
CA: You use the analogy that we can’t get to the Moon by extending a building a foot at a time. But the experience that most of us have had of these language models is not a foot at a time. It's like, the sort of, breathtaking acceleration. Are you sure that given the pace at which those things are going, each next level seems to be bringing with it what feels kind of like wisdom and knowledge.
YC: I totally agree that it's remarkable how much this scaling things up really enhances the performance across the board. So there's real learning happening due to the scale of the compute and data.
Frequently Occurring Word Combinations
ngrams of length 2
| collocation |
frequency |
| common sense |
13 |
| language models |
4 |
| dark matter |
4 |
| artificial intelligence |
3 |
| bar exam |
3 |
| ai today |
3 |
| ai models |
3 |
| deep neural |
3 |
| commonsense knowledge |
3 |
| trillion words |
2 |
| easily fixed |
2 |
| tech companies |
2 |
| ai sustainable |
2 |
| greatest ai |
2 |
| similar examples |
2 |
| normal matter |
2 |
| paper clips |
2 |
| world works |
2 |
| research topic |
2 |
| raw web |
2 |
| teach ai |
2 |
| large language |
2 |
| language model |
2 |
| commonsense models |
2 |
| neural networks |
2 |
ngrams of length 3
| collocation |
frequency |
| deep neural networks |
2 |
Important Words
- acceleration
- agree
- analogy
- board
- breathtaking
- bringing
- building
- compute
- data
- due
- enhances
- experience
- extending
- feels
- foot
- happening
- kind
- knowledge
- language
- learning
- level
- models
- moon
- pace
- performance
- real
- remarkable
- scale
- scaling
- sort
- time
- totally
- wisdom