news LinkedIn

#ai | Denis O. | 15 comments

The post argues that contemporary “agentic AI” is largely an elaborate software harness built around fundamentally unchanged large language models (LLMs). Tools such as memory, retrieval, planning, verification, context management, and repeated model calls can create sophisticated workflows, but they do not alter the underlying model or give it genuine agency, autonomy, or an internal continuous loop. In this view, the LLM supplies the neural component while the surrounding orchestration supplies a symbolic component, resembling cognitive architectures from the 1990s. The author contends that these systems lack stable autonomy because context can accumulate, assumptions can be compressed or lost, errors can propagate, and behavior can drift. Developers respond by adding more checkpoints, retrieval, context engineering, and external control, effectively building increasingly complex harnesses around the same limitations. The post also rejects claims that such systems possess consciousness or persistent selves. Memory supplied by software is not equivalent to an enduring identity; an externally maintained loop is not an inner recursive process; and generated statements such as “I am afraid” do not demonstrate subjective experience or suffering. The author believes observers are mistaking complex behavior and orchestration for genuine understanding or awareness. Comments broadly support or extend this critique. Several compare LLM systems to Frankenstein, embalmed horses, clockwork automata, or sophisticated writing machines. Others emphasize maintaining meaningful human control, treating LLMs as tools rather than unified intelligences, and recognizing that system components can malfunction without becoming autonomous. One commenter argues that genuine synthetic subjective cognition would require an analog, embodied, continuously changing substrate rather than primarily digital computation.

Build this article’s timeline

Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.

Log in to build

Timeline generation is available to signed-in community members.

Five Ws 🤖 5 AI prefills · all filled
Who
Stefan Đorđević, Elena Chang, Max Winter, and LinkedIn
🤖 AI prefill
What
The post argues that contemporary “agentic AI” is largely an elaborate software harness built around fundamentally unchanged large language models (LLMs).
🤖 AI prefill
When
1990s and 18th century
🤖 AI prefill
Where
San Francisco Bay Area
🤖 AI prefill
Why
Centers on LLM harness architecture, agentic AI, and externally imposed agency
🤖 AI prefill

Login to answer and earn XP for each gap you fill.

LLM Harness Architecture LLMs, memory, tools, retrieval, planning Limits Of Agentic Autonomy agency, autonomy, persistent recursion, inner loop, state Consciousness And Subjectivity consciousness, inner experience, awareness, persistent self, subjective experience AI Anthropomorphism Critique dead horse, automata, Frankenstein, clockwork, anthropomorphism Alternative Cognitive Substrates analog hardware, continuous-time reservoir, adaptive ecology, historical transduction, causal influence
LLM harness architecture98% agentic AI97% externally imposed agency96% neurosymbolic AI94% cognitive architectures92% persistent recursion90% stable autonomy89% context accumulation and drift87% exogenous orchestration86% persistent self85%
48
Manipulation Score: Medium How much this article uses emotional framing vs. straight reporting
Loaded Language 7225%
Absolutist 5815%
Threat/Panic 1825%
Us vs Them 2815%
Engagement Bait 2220%
Manipulation score look right?

Factual Claims (10)

medium technology
Current frontier AI is primarily an elaborate harness around fixed, frozen LLMs.
medium technology
LLM systems combine neural networks with externally supplied symbolic components, creating a form of neurosymbolic AI.
low technology
Modern AI agent systems use software for memory, retrieval, tool calls, planning, verification, context management, and repeated model invocation.
high technology
The agency in current agentic AI is externally imposed by surrounding software rather than originating inside the language model.
medium technology
LLMs do not contain persistent recursion or an continuously running inner loop.
high technology
Current AI systems do not produce stable autonomy.
medium technology
Context accumulation, compression, error propagation, and system drift can degrade the performance of orchestrated LLM systems.
high scientific
Memory and software loops do not by themselves produce consciousness, subjective experience, awareness, or a persistent self.
high scientific
An externally implemented loop does not become an inner loop merely because an LLM participates in it.
high technology
Adding compute, longer context, better memory, tools, and additional orchestration does not make the underlying LLM alive or autonomous.

Comments

Login to view and post comments.

🏠Portal 📰Links ❓Q&A 📅Events 💼Jobs