The bad news: AI is going pretty much as I expected
We recorded 7 claims from this post. 5 were checked against evidence. 2 can’t be checked. 3 more are held. Read the post at its source.
Claims from this post (7)
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
“The notion on which this Substack was founded in May 2022 — that LLM scaling would not bring us to AGI — has gone mainstream.”
What we tested: The view that LLM scaling will not bring AGI has become mainstream.
What happened: The publicly reported survey showed broad skepticism about scaling existing approaches to AGI. This supports mainstream acceptance of skepticism, rather than certainty that scaling cannot succeed. (The Decoder, )
Assessment: UnresolvedSaid
“that LLM scaling would not bring us to AGI”
What we tested: Scaling LLMs will not produce AGI.
Assessment: UnresolvedSaid
“Perhaps I also no longer need to warn people that the idea that coding would disappear soon is bullshit.”
What we tested: Human coding will not disappear soon due to AI.
Assessment: SupportedSaid
“The concerns about potential copyright violations that Reid Southen and I expressed here in and at greater length in the IEEE Spectrum in January 2024 continue almost two years later to be a problem.”
What we tested: Copyright violation concerns with generative AI remain a problem as of September 2025.
What happened: The report documents Warner's new lawsuit and examples of allegedly infringing generated characters. This establishes a concrete continuing output-related dispute, without establishing the allegations as judicial findings. (Ars Technica, )
Assessment: SupportedSaid
“Meanwhile, more darkly, the Jurassic Park moment of mass-produced AI-generated misinformation that I projected in December 2022 has arrived.”
What we tested: Mass-produced AI-generated misinformation has arrived at scale.
What happened: The accounts published 9,784 videos from March 2023 through June 2024, attracting over 380 million views. The investigation documents specific false narratives and describes using TrueMedia.org to assess AI-generated audio; it does not establish that every counted video was false. (NewsGuard, )
Assessment: Can’t be checkedSaid
“If there is any true exponential nowadays, I fear it is this: the rapid doubling of AI-generated misinformation.”
What we tested: Marcus fears that, if there is any true exponential nowadays, it is the rapid doubling of AI-generated misinformation.
Assessment: Can’t be checkedSaid
“Let us all hope that we can learn something from the mistakes of the last few years, and find a new path — to an AI that better serves humanity.”
What we tested: Marcus hopes that we all can learn something from the mistakes of the last few years and find a new path to an AI that better serves humanity.
Held from this post (3)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The criteria are ambiguous. Our reading, not assessed: As far as Marcus knows, the problem of generative AI producing wildly derivative outputs has not gone away.
- The two reviews read the claim differently
- The two reviews read the claim differently