Breaking: Trump asks the impossible of Anthropic
We recorded 4 claims from this post. 2 were checked against evidence. 2 can’t be checked. 3 more are held. Read the post at its source.
Claims from this post (4)
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: SupportedSaid
Observation
“It took longer than I expected, but well, here we are:”
Our reading: The politics and inadequacy of generative AI guardrails, which Marcus predicted in 2024 would become central, has now become a central issue, though later than he expected.
No horizon Evidence as of
Limitation: Prominence is qualitative.
Assessment: Can’t be checkedSaid
Observation
“Next-token predictors simply aren‘t built for safety.”
Our reading: Next-token predictors (LLMs) simply aren't built for safety.
No horizon
Scope: next-token predictors
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Either we curtail LLMs until we find a better technology, or we live with consequences.”
Our reading: Society must either restrict LLMs until a better technology is found or accept the consequences of their unsafe guardrails.
No horizon
Qualifier: until a better technology is found
Assessment: SupportedSaid
Observation
“Importantly, this is not an Anthropic problem, it’s a Generative AI problem.”
Our reading: The inability to block all jailbreaks is a problem with generative AI in general, not one specific to Anthropic.
No horizon Evidence as of
Limitation: The assessment was published in May, but the accessible source does not give individual experiment dates.
Held from this post (3)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: For LLMs, it is not possible to guarantee that guardrails cannot be circumvented by jailbreaks.