Unverified50% confidenceFactExact time
Current large language models are trained on massive amounts of text and are essentially performing statistical association learning at the symbol level, without a body, senses, or the ability to act in the physical world.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/16/2026
First Seen
Valid until: 1/14/2027
Sources
The Bottleneck of Intelligence Is Not Intelligence, But Continuous Contact with Reality
twitterfchollet10/14/2026
Related Entities
Related Claims
Unverified大语言模型在二维空间推理上存在系统性短板，因训练数据以文本为主、缺乏空间关系监督信号79% similarUnverifiedThe researchers acknowledge that the model still falls far short of classical language models trained on trillions of tokens on text tasks.79% similarUnverifiedTraditional keyword matching or rule engines can only process literal signals, whereas large language models have acquired a holistic understanding of semantics and sentiment through pre-training on massive amounts of human text.77% similarUnverifiedCurrent large language models are essentially statistics-based text generation systems with no persistent goals, desires, or self-awareness.76% similarVerifiedLarge language models have two inherent weaknesses: their knowledge is cut off at training time and they cannot obtain the latest information, and they can only generate text and cannot directly perform actions.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1044400API
curl https://kongchang.com/api/v1/knowledge/claims/1044400MCP
get_claim(id=1044400)