“For all the efforts that OpenAI and other leaders of deep learning, such as Geoffrey Hinton and Yann LeCun, have put into running neurosymbolic AI, and me personally, down over the last decade, the cutting edge is finally, if quietly and without public acknowledgement, tilting towards neurosymbolic AI.”
Our reading, the sentence that was checked:
The cutting edge of AI is finally tilting towards neurosymbolic AI, quietly and without public acknowledgement.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI, paragraph 24.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-how-o3-and-grok-4-accidentally-vindicated-dcbd3b-a1
We recorded 47 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-07-13. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
Conditions
- quietly and without public acknowledgement (qualifier, stated by the speaker)
From the speaker
- the cutting edge
- finally
- quietly and without public acknowledgement
- tilting towards neurosymbolic AI
Our interpretation
- The ordinary technical reading is increasing integration of neural models with explicitly executed symbolic operations, including code interpreters. It does not require replacing the neural model.
- The essay makes o3 and Grok 4 its central examples. I read the acknowledgement qualifier as their developers not presenting this integration as a turn toward neurosymbolic AI, rather than concealing the existence of tools.
- The essay separately recognizes Google DeepMind's hybrid systems. I therefore do not read the qualifier as denying every public acknowledgement by every laboratory.
- The surrounding text repeatedly locates the development in 2025. The target is 2025 through publication, the latest public development window described by the essay.
- The allegation about past personal disparagement is background to the supplied normalized proposition and is not separately scored.
Counts as supported if: Public descriptions of the named frontier systems show increased integration of neural reasoning with executable symbolic tools in 2025, while describing that integration without adopting the neurosymbolic label.
Counts as contradicted if: The named systems do not incorporate the claimed symbolic operations, or their relevant announcements explicitly acknowledge their development as a turn toward neurosymbolic AI.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Under the essay's broad definition, its central technical observation holds: o3 and Grok 4 integrated reasoning with executable tools, including code interpreters. Their launch descriptions openly discussed tools without framing the change as neurosymbolic AI. This supports the contextual reading of the claim, not a universal assertion that hybrid methods went unacknowledged. DeepMind's explicit terminology is a material limit, and Marcus himself recognizes that exception.
Limitations
- The result depends on counting code-interpreter integration as neurosymbolic AI.
- The acknowledgement finding concerns the reviewed launch descriptions of the named systems.
- A literal claim that all frontier development was quiet or unacknowledged would not receive this outcome.
Evidence (3 items)
Supports the claim: OpenAI,
academic benchmarks and real-world tasks
Why it matters here: The surrounding technical description identifies Python execution and newly combined tool use within reasoning, and describes training the models to choose tools. It presents this as reinforcement learning and agentic tool use, without calling it neurosymbolic AI.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Supports the claim: xAI,
Grok 4 was trained with reinforcement learning to use tools.
Why it matters here: The section identifies a code interpreter and browsing as tools that augment the model's reasoning. The launch description calls this native tool use rather than neurosymbolic AI.
- Publisher
- xAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: arXiv; Google DeepMind researchers,
neuro-symbolic system
Why it matters here: The AlphaGeometry2 paper openly describes the neurosymbolic lineage of its system. This defeats a blanket claim that no frontier laboratory acknowledged such work, but the essay itself recognizes DeepMind as an exception.
- Publisher
- arXiv; Google DeepMind researchers
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, How o3 and Grok 4 Accidentally Vindicated Neurosymbolic AI
For all the efforts that OpenAI and other leaders of deep learning, such as Geoffrey Hinton and Yann LeCun, have put into running neurosymbolic AI, and me personally, down over the last decade, the cutting edge is finally, if quietly and without public acknowledgement, tilting towards neurosymbolic AI.
- Where
- paragraph 24
- Published
- 2025-07-13
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-how-o3-and-grok-4-accidentally-vindicated-dcbd3b-a1: Current. Outcome: Supported. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-how-o3-and-grok-4-accidentally-vindicated-dcbd3b-a1.