Make money doing the work you believe in
Stochastic Parrot is such a silly term for AI when you take into account that:
1. African Grey parrots can understand the context and meaning behind the words they speak, are renowned for their intelligence, associate specific words with items, colors, and actions, and can use language to communicate.
2. Humans are the OG stochastic parrots. Our brains use associative, probabilistic processes to understand and operate in a predictive and fundamentally stochastic (probabilistic) way; we constantly predict words, senses, movement, and more.
3. The evidence overwhelming supports that LLMs understand language. Understanding is defined as the ability to comprehend or interpret meaning from written, spoken, or visual information. Language comprehension means the ability to grasp words, syntax, relations, commands, reference, and context well enough to demonstrate meaning-sensitive understanding. That is the standard used in practice for animals and humans.
According to research, someone genuinely understands language (receptive language) when they can demonstrate meaning. LLM’s demonstrate meaning, even with typos and terrible grammar, they grasp meaning. That’s like…their whole thing. If they didn’t, it would be a nightmare to talk to them. Pretty much just a game of mad libs.
Key signs include (and I’m putting a checkmark by everything that LLM’s have shown they are able to do):
-responding to commands ✔️
-naming objects✔️
-understanding syntax (word order)✔️
-processing figurative language✔️
-certain brain regions lighting up when you are actively understanding language ✔️
Early detection methods, like eye-tracking "looking while listening," show comprehension precedes speech. This concept has been incorporated into multimodal systems.
Anyway. They were wrong about the parrots and they’re wrong about this too.
Citations:
Pepperberg, I. M. (2006). Cognitive and communicative abilities of Grey parrots. Applied Animal Behaviour Science, 100(1–2), 77–86.
Rekimoto, J. (2025). GazeLLM: Multimodal LLMs incorporating Human Visual Attention. In Proceedings of the Augmented Humans International Conference 2025 (AHs '25). Association for Computing Machinery, New York, NY, USA, 302–311. doi.org/10.1145/3745900…
Du, C., Fu, K., Wen, B., Sun, Y., Peng, J., Wei, W., ... & He, H. (2025). Human-like object concept representations emerge naturally in multimodal large language models. Nature Machine Intelligence, 7(6), 860-875. arxiv.org/abs/2407.01067
Goldstein, A., Ham, E., Schain, M. et al. Temporal structure of natural language processing in the human brain corresponds to layered hierarchy of large language models. Nat Commun16, 10529 (2025). doi.org/10.1038/s41467-…
Caucheteux C, King JR. Brains and algorithms partially converge in natural language processing. Commun Biol. 2022 Feb 16;5(1):134. doi: 10.1038/s42003-022-03036-1. Erratum in: Commun Biol. 2023 Apr 11;6(1):396. doi: 10.1038/s42003-023-04776-4. PMID: 35173264; PMCID: PMC8850612.
Elsevier. (n.d.). Language ability. ScienceDirect Topics. Retrieved March 25, 2026, from sciencedirect.com/topic…
Merriam-Webster. (n.d.). Understanding. In merriam-webster.com dictionary. Retrieved March 25, 2026, from merriam-webster.com/dic…
