Also, reminder that if an expert asserts "after looking at this thing, I can conclude X", whether that expert is a human or a machine, they should be able to point out what evidence they found and why it is good evidence. If they can't do that, don't trust them.
RESEARCH
-
Licensing concerns for AI training data sources
By
–
I can’t look at this now without wondering at least a little about whether the sources have been licensed. https://t.co/1QBe47J6VB
— Gary Marcus (@GaryMarcus) 7 janvier 2024I can’t look at this now without wondering at least a little about whether the sources have been licensed.
-
Augmentation over replication: moving beyond the Turing trap
By
–
I lean towards augmentation since it implies new things, not just more of the same. https://
digitaleconomy.stanford.edu/news/the-turin
g-trap-the-promise-peril-of-human-like-artificial-intelligence/
… -
ML Systems Cannot Magically Recover Hidden Information
By
–
A prevalent cognitive fallacy around ML systems is the belief that they can magically recover hidden information that is out of reach for human experts — "the AI detected that those Apollo pictures are fake!", "the AI can predict if the person in the picture is a criminal!",
-
Gold Intelligence Augmentation Douglas Engelbart Vision
By
–
Or intelligence augmentation per Douglas Engelbart
-
LeCun’s LLM Thoughts and Academic Attribution Concerns
By
–
Sources: https://
en.wikipedia.org/wiki/Literatur
e_review
…
and https://
ncbi.nlm.nih.gov/pmc/articles/P
MC2812776/
… The fact that LeCun’s recent thoughts on LLMs resemble my own earlier thoughts is widely known, and has been documented elsewhere. The fact he routinely he fails to recognize that work is evident in virtually all of -
Linear Regression Should Stop Being Called AI
By
–
can finally stop calling Linear Regression as AI.
-
LLMs Show Interest in Learning COBOL Programming Language
By
–
You know who’s interested in learning COBOL? LLMs
-
LeCun Aligns with Earlier AI Theory Positions on Hinton
By
–
Luckily I don’t need to correct Hinton on this anymore now that LeCun has started echoing what I was saying all along.
-
Code Training Empowers LLMs: Generation, Reasoning, Agents
By
–
9/ How Code Empowers LLMs – an overview of the benefits of training LLMs with code-specific data. Some capabilities include enhanced code generation, enabling reasoning, function calling, automated self-improvements, and serving intelligent agents.