Transformers May Face a Choice: The Junk Pile or Pizza Hut
November 4, 2025
Another short essay from a real and still-alive dinobaby. If you see an image, we used AI. The dinobaby is not an artist like Grandma Moses.
I read a marketing collateral-type of write up in Venture Beat. The puffy delight carries this title “The Beginning of the End of the Transformer Era? Neuro-Symbolic AI Startup AUI Announces New Funding at $750M Valuation.” The transformer is a Googley thing. Obviously with many users of Google’s Googley AI, Google perceives itself as the Big Dog in smart software. Sorry, Sam AI-Man, Google really, really believes it is the leader; otherwise, why would Apple turn to Google for help with its AI challenges? Ah, you don’t know? Too bad, Sam, I feel for you.
Thanks, MidJourney. Good enough.
This write up makes clear that someone has $750 million reasons to fund a different approach to smart software. Contrarian brilliance or dumb move? I don’t know. The write up says:
AUI is the company behind Apollo-1, a new foundation model built for task-oriented dialog, which it describes as the "economic half" of conversational AI — distinct from the open-ended dialog handled by LLMs like ChatGPT and Gemini. The firm argues that existing LLMs lack the determinism, policy enforcement, and operational certainty required by enterprises, especially in regulated sectors.
But there’s more:
Apollo-1’s core innovation is its neuro-symbolic architecture, which separates linguistic fluency from task reasoning. Instead of using the most common technology underpinning most LLMs and conversational AI systems today — the vaunted transformer architecture described in the seminal 2017 Google paper "Attention Is All You Need" — AUI’s system integrates two layers:
Neural modules, powered by LLMs, handle perception: encoding user inputs and generating natural language responses.
A symbolic reasoning engine, developed over several years, interprets structured task elements such as intents, entities, and parameters. This symbolic state engine determines the appropriate next actions using deterministic logic.
This hybrid architecture allows Apollo-1 to maintain state continuity, enforce organizational policies, and reliably trigger tool or API calls — capabilities that transformer-only agents lack.
What’s important is that interest in an alternative to the Googley approach is growing. The idea is that maybe — just maybe — Google’s transformer is burning cash and not getting much smarter with each billion dollar camp fire. Consequently individuals with a different approach warrant a closer look.
The marketing oriented write up ends this way:’
While LLMs have advanced general-purpose dialog and creativity, they remain probabilistic — a barrier to enterprise deployment in finance, healthcare, and customer service. Apollo-1 targets this gap by offering a system where policy adherence and deterministic task completion are first-class design goals.
Researchers around the world are working overtime to find a way to deliver smart software without the Mad Magazine economics of power, CPUs, and litigation associated with the Googley approach. When a practical breakthrough takes place, outfits mired in Googley methods may be working at a job their mothers did not envision for her progeny.
Stephen E Arnold, November 4, 2025
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
