LLMs as Giant Lookup-Tables of Shallow Circuits

Early 2026 LLMs in scaffolds, from simple ones such as giving the model access to a scratchpad/" chain of thought " up to MCP servers, skills, and context compaction &c are quite capable. ( Obligatory meme link to the METR graph.) Yet: If someone had told me in 2019 that systems with such capability would exist in 2026, I would strongly predict that they would be almost uncontrollable optimizers, ruthlessly & tirelessly pursuing their goals and finding edge instantiations in everything. But they don't seem to be doing that.