Research / Engineering / News / Blog
Latest
Research papers, engineering notes, company news, and practical Blog articles from Licklider, newest first.
Latest entries
September 15, 2026
Engineering / Implementation note
Making verification results more useful to research agents
nomue development improves difficult calculations, makes completed checks explicit, and preserves the meaning of historical results as versions change.
Product development update — integrated candidates and internal milestones; no new hosted release or public verifier support
September 15, 2026
Research / Publication
Same Test, Different p — preprint v1.0
An audit of R, SciPy, and Julia shows why matching statistical decisions can still hide differences in definitions, numerical precision, and valid probability values.
Preprint v1.0 — not peer reviewed; reproducibility data and code available
September 14, 2026
Engineering / Upstream report
statsmodels loses finite Welch results when an intermediate sum overflows
Four exactly represented observations per group make statsmodels overflow an intermediate sum, losing finite variances and Welch t-test results.
Fix merged into statsmodels main in PR #10255 — not yet released
September 13, 2026
Engineering / Implementation note
Separating format checks from verification results
An experimental Holm verifier separates format and declaration checks from integrity, context and arithmetic results, preserving explicit outcomes for checks that did not run.
Historical unissued candidate.4; candidate.5 now implements revised dependencies; formal adoption and support remain open
September 12, 2026
Engineering / Upstream report
Checking Welch results with exact rescaling
Exact inputs and independent references expose a changed Welch p-value despite a finite result and no warning in a SciPy boundary test.
NumPy float64 repair PR #26209 open; current head is mergeable with 56 / 56 checks passing; issue #26169 open; not merged or released
September 11, 2026
Engineering / Implementation note
When a verification call must discard its result
An experimental Holm verifier connects Record checks to shared execution budgets, operating-system limits and cleanup evidence before deciding whether a result can be returned.
Historical execution candidate preserved; candidate.5 successor has a merged development checkpoint; no additional supported capability
September 11, 2026
Engineering / Implementation note
Binding Holm corrections to the intended comparisons
An experiment checks exact Holm adjustments together with the expected declaration and supplied p-values, including changes that leave the displayed answer unchanged.
Original binding experiment preserved; successor includes scoped numerical review and deterministic sort repair; no additional supported capability
September 11, 2026
Engineering / Implementation note
Checking factorial probability evidence against the raw observations
The factorial candidate now joins Record checks, exact probability evidence, a complete report and controlled execution in an independently reviewed readiness package.
Independently reviewed, unissued Release 4 candidate; D01/D07 amendment discussion open; no Protocol support
September 11, 2026
Blog / Practical guide
What to decide before asking an agent to compare several groups
Specify the comparisons, research conditions, and outputs you need so an agent can propose an analysis that answers your question.
Illustrative planning example; not a tested agent workflow
September 10, 2026
Engineering / Technical method
Checking factorial statistics without trusting rounded intermediates
Exact arithmetic and probability bounds offer a path beyond scaling repairs, while a review shows why matching rounded answers does not certify an interval.
Reviewed research components now connected experimentally and archived; no additional supported capability