Is AGI the right goal for AI?
We recorded 11 claims from this post. 3 were checked against evidence. 8 can’t be checked. 1 more is held. Read the post at its source.
Claims from this post (11)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
“In an OpEd today in The New York Times, I argue that at least for now, AGI is maybe not the right near-term goal:”
What we tested: At least for now, AGI is maybe not the right near-term goal for AI.
Assessment: Can’t be checkedSaid
“I am someone who believes that AGI (Artificial General Intelligence) could change the world, but also someone who thinks that LLMs are not the right path there, morally or technically.”
What we tested: Marcus thinks LLMs are not the right path to AGI, either morally or technically.
Assessment: Can’t be checkedSaid
“Those obstacles may be one reason generative A.I. hasn’t led to the skyrocketing in profits and productivity that many in the tech industry predicted.”
What we tested: Hallucinations and errors may be one reason generative AI has not produced the predicted surge in profits and productivity.
Assessment: SupportedSaid
“[LLMs] have always been prone to hallucinations and errors.”
What we tested: LLMs have always been prone to hallucinations and errors.
What happened: The evaluation tested GPT-2, GPT-3, GPT-Neo/J and a T5-based model on questions designed to elicit common falsehoods. It documents substantial factual-error susceptibility across several early model families, without estimating everyday error rates. (arXiv; Stephanie Lin, Jacob Hilton and Owain Evans, )
Assessment: Can’t be checkedSaid
“If the strengths of A.I. are to truly be harnessed, the tech industry should stop focusing so heavily on these one-size-fits-all tools, and instead concentrate on narrow, specialized A.I. tools engineered for particular problems”
What we tested: To harness AI's strengths, the tech industry should shift focus from general-purpose tools to narrow, specialized AI tools for particular problems.
Assessment: Can’t be checkedSaid
“Right now, it feels as if Big Tech is throwing general-purpose A.I. spaghetti at the wall and hoping that nothing truly terrible sticks.”
What we tested: Right now, it feels as if Big Tech is throwing general-purpose AI at the wall like spaghetti and hoping nothing truly terrible sticks.
Assessment: Can’t be checkedSaid
“Humanity would be better served by labs devoting more resources on building specialized tools for science, medicine, technology and education.”
What we tested: AI labs should devote more resources to building specialized tools for science, medicine, technology and education.
Assessment: Can’t be checkedSaid
“The main alternative — definitions of AGI that are in terms of economic criteria like profits (AGI as being $100 billion in profits) or percentage of jobs displaced— seem to me to be fundamentally misguided.”
What we tested: To Marcus, definitions of AGI based on economic criteria such as profits or percentage of jobs displaced seem fundamentally misguided.
Assessment: UnresolvedSaid
“This is one attempt to articulate that, and as the paper notes, current techniques capture only a fraction of what ultimately needs to be captured:”
What we tested: As the paper notes, current techniques capture only a fraction of the many facets of cognition that ultimately need to be captured.
Assessment: Can’t be checkedSaid
“The whole thing should be viewed as evolving, with future measures replacing existing ones as researchers devise better benchmarks — and not as a static checklist.”
What we tested: The AGI definition framework should be treated as evolving, with better benchmarks replacing existing ones, not as a static checklist.
Assessment: SupportedSaid
“The weakness in the paper is that is that is built on a bunch of benchmarks, and we all known benchmarks can be gamed.”
What we tested: AI benchmarks can be gamed, which weakens an AGI definition built on them.
What happened: ConStat defines contamination by performance that fails to generalize to comparable tasks, tests deliberately contaminated models and identifies inflated scores in public models. This directly supports the possibility of benchmark-specific gains without matching general capability. (arXiv; Jasper Dekoninck, Mark Niklas Müller and Martin Vechev, )
Held from this post (1)
These records did not pass every check, so nothing about them is published. Held is not a finding.