Efficient Thinking

One question, from every angle: how much useful capability can a fixed intelligence get per unit of computation, and what moves that number? Each paper holds the model fixed, measures capability with an external verifier, and changes one thing: search, the evaluator, the judge, taste, feedback, labels, the internal ceiling, experience, symmetry, representation.

The result to try to break. A frozen 7B model, given a read-only linear head fitted on its own verified history, solves more problems at the same inference compute — and, at equal quality, spends 2.3× less. The mechanism transfers to a second search structure with the head re-fitted, and a second generation of experience beats the first, then a third does not. The headline reproduces from one command in about three hours on an Apple-silicon Mac.

git clone https://github.com/LouayAlsakka/efficient-thinking && cd efficient-thinking
bash experience/reproduce_matched_compute.sh      # base at the head's compute vs the head, n = 300, paired

It refuses to print a table if any gate fails and exits non-zero if the published interval is not reproduced. An independent run on a second machine by someone who had not touched the work read +8.0 [+3.0, +13.3] against the published +9.3 [+3.7, +15.0].

Figure 1: the quality–compute frontier, the paired effect everywhere it was measured, and the same-decision baselines

Papers

#TitleStateRead
IMeasuring What Capability Costs — search substitutes for size on a frozen chess evaluatormeasuredhtml · pdf · md
IIWhere Search Pays and Where It Can't — the evaluator's ceiling, in games and in languagemeasuredhtml · pdf · md
IIIEfficient Judging — when an LLM judge is worth its computemeasured, draftpdf · md
IVSearch Where Taste Is the Evaluatormachine arms measured; rater arms runningproposal
VThe Exchange Rate of Feedbackconceptconcept
VIThe Label Ceiling — why self-play plateausmeasured, unwrittenconcept · ledger
VIIThe Elicitation Gap — what a fixed system can get from insideconcept; probe runningconcept
VIIIExperience Priors — the constraint and the carrierclosed, reproducedhtml · pdf · md
VIII-bAccumulation — does experience move the frontier againmeasured; paper in progressrecord
IXSymmetry — invariance as thought compressionconceptconcept
XMachine-Native Representationconceptconcept · plan

Every paper opens with a state line — MEASURED, REGISTERED, CONCEPT or IDEA — so that a concept is never read as a finding. Withdrawn claims stay in the record, dated. Readings are written before runs. Result artefacts name their instruments by hash. The series plan and the one-question rule are in series-completion-plan.md.

Review

There is no venue, no referee and no gate. Review is public and runs through the repository's issues. The four places the author thinks Paper VIII is weakest, so you know where to aim: an alternative explanation for the matched-compute frontier shift; leakage in how the experience was constructed; a flaw in the paired statistics; a reason the second search structure does not establish transfer of the mechanism. A fatal flaw is worth more than praise, and the author will run any experiment you name.

Author

Louay Alsakka. An electrical engineer from the semiconductor industry; this is a first research project, done outside any institution with AI assistance, and published directly. The chess engine of Paper I is where the two interests meet.

Repository: github.com/LouayAlsakka/efficient-thinking