The type of LLMs used today will not be able to make the leap the interviewer asked about, with the main reason given being the way LLMs are trained, by predicting masked words in text.
Unresolvable The public record could not settle it.
Prediction: Statements about events outside the person's control, checked once they fell due: did they come true?
Reasoning
Open-ended prediction with no horizon, and the material does not define what 'this' refers to beyond 'a giant leap'.
Cite this claim
Yann LeCun, “The type of LLMs used today will not be able to make the leap the interviewer asked about, with the main reason given being the way LLMs are trained, by predicting masked words in text.” (said 8 March 2024). Verdict: unresolvable. Reliability API, claim c_0257d109b0a8, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_0257d109b0a8
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More prediction claims by Yann LeCun
-
Mostly true
“Increasingly, the search engines that we’re going to use are not going to be search engines, they’re going to be dialogue systems”
-
Mostly true
“one of the major challenges of computer vision, machine learning, and AI in general in the next decade will be to devise methods that can automatically learn good features hierarchies from unlabeled…”
-
Mostly true
“We foresee an increase in the usage of such networks for a wide variety of applications, which are likely to violate the assumptions of current DL hardware.”
-
Mostly true
“As DL systems are increasingly used for high-level cognitive tasks, such memory modules will become commonplace and very large, requiring hardware support.”
-
Mostly true
“The next frontier for Deep Learning is natural language understanding.”
Natural language understanding will be the next frontier for Deep Learning.
Verdicts are made automatically; how verdicts are made.