@Filapek @FNArena Apologies. Here's the image:

@rao2z @rao2z You might be interested to read this paper that distinguishes between artificial intelligence and artificial expertise:

@douwekiela Agreed. You might be interested to read a paper that distinguishes between artificial intelligence and artificial expertise:

@SGRodriques @tom_doerr @SGRodriques Based on my paper, 50 to 70% of academics admit to using AI in their work when surveyed anonymously, but only 3.5% acknowledge it formally in Elsevier articles (which require them to make acknowledgement):

Anonymous surveys suggest 50-70% of researchers use AI tools; but across 568 Elsevier journal articles, I found only ≈ 3.5% acknowledged AI use. I discuss possible causes and consequences of the AI attribution gap in my recent paper:

Takeaways: (1) Someone you think is 'pretty good' on three key traits is likely extraordinary—value them more; (2) If AI/robotic IQ ever hits 130, programming elevated C and ES could be the easy step toward thrice-exceptionality.

RT @ACPIDCongress: Abstract Submissions are open for this years' ACPID conference! Visit to submit your abstract!…

@JeremyNguyenPhD

New paper! Based on the performance of 591 LLMs on 12 tasks, we found: 1) A positive manifold (mean r = .73) 2) An Artificial General Ability (AGA) factor explaining 66% of test variance. 3) Parameters correlate positively with AGA (r ≈ .6)

RT @ACPIDCongress: Applying for an #academicjob ? Avoid some common mistakes & learn ways to make your application packet stand out! Works…

RT @UWApsych: Will #AI mean the death of higher education, or is it a brave new dawn for learning? Join us at #UWA on July 30th for an inn…

@JeremyNguyenPhD I subscribe to GPT, but leaning toward Perplexity lately. I can't say which model it's using. The free mode is definitely useful for R. However, I use the 'Pro' mode (limited use for non-subscribers) when it's a tougher problem (makes a difference).

@mikeknoop @fchollet This is academic paper, which contends there is more evidence for artificial achievement/expertise than artificial intelligence, shares your point of view:

RT @iqmobile: “Our results indicate that amplitudes of the ERP components measured in #executivefunction tasks almost exclusively represent…

Longitudinal study in school-aged children suggests there is more evidence for IQ influencing Need for Cognition, rather than the other way around. Notably, Gc and Gq demonstrated the largest effects (not Gf).

@nlw In your AIBreakdown podcast, you emphasized the need for clear definitions of AI and AGI. Here's a link to a recently published paper that offers such definitions, along with a discussion on Artificial Expertise (what we probably have currently):

@apdemetriou Your impression would be consistent with a recently published paper in Intelligence on the current nature of artificial systems/models like GPT, which may be better described as 'artificial achievement' or 'artificial expertise':