Skip to content
RCreddit.com·

what are people's opinions of symbolic AI?

AI summary

The current AI discourse, heavily influenced by ML/LLM, prompts a re-evaluation of symbolic AI. Revisiting older ideas, such as automated proof assistants, can illuminate present advances and offer more provocative conceptual formulations than modern tools like Lean or Agda. This perspective suggests that LLMs might not be creating entirely new intellectual programs but rather making previously marginal or impractical ones newly viable, sparking curiosity about others' views on this interpretation of the current AI landscape.

Time & source
Published
Sep 7, 2026, 05:40
Source type
Dev community
Tier
Community
Source status
Healthy
Tier is a per-source editorial setting, not a per-item score.

Times shown in UTC

More details
First seenSep 7, 2026, 14:00Time zoneUTC · UTC+0
Article

I've been thinking about how much current AI discourse is shaped by an ML/LLM-centric view of the field.

The current state of the art is genuinely impressive, but some of what looks radically new can also be understood as older ideas becoming newly practical. Mathematics seems like a good example: LLMs may make automated theorem proving and formal methods applicable to a much broader range of mathematical work.

But there is a much older intellectual lineage here—from Principia Mathematica to the Logic Theorist and the broader symbolic-AI tradition. A lot of the foundational questions were posed surprisingly early, then partly displaced by approaches that proved more practical or scalable.

I find revisiting those older ideas useful because they can illuminate what is actually interesting about present advances. In some cases, I even think the earlier conceptual formulations are more provocative than the modern tools themselves. The general idea of an automated proof assistant, for example, raises questions that aren't exhausted by the particulars of Lean, Agda, or whatever system an LLM happens to use successfully.

So even if LLMs become extremely good at operating these tools, that doesn't automatically surface the most distinctive or interesting questions about what they're doing. Those questions may require interpretation—and some of the best interpretive frameworks may already exist in older traditions of logic, automated reasoning, symbolic AI, and philosophy of mathematics.

I'm curious whether others see the current moment this way: less as LLMs creating an entirely new intellectual program, and more as making previously marginal or impractical programs newly viable.

Source·reddit.com·Full text via RSS