Show HN: Gave Claude Code ADHD.. Now it thinks 3x better
A new method called ADHD has been introduced to enhance the ideation capabilities of large language models. This approach allows for parallel divergent thinking, producing a wider range of creative solutions rather than defaulting to common responses. In evaluations, ADHD significantly outperformed traditional methods in terms of novelty and breadth of ideas.
- ▪ADHD prevents premature convergence by generating multiple divergent branches before converging on the best ideas.
- ▪The method was tested against a single-shot baseline and won 5 out of 6 evaluations on open-ended engineering problems.
- ▪ADHD improves metrics such as novelty, breadth, and trap detection compared to existing methods.
Opening excerpt (first ~120 words) tap to expand
Preprint · v0.1 · 2026-05-25 ADHD: Parallel Divergent Ideation for Coding Agents Tree-of-thought with cognitive-frame branching, generator–critic separation, and pruning. Udit Akhouri Raj · github.com/UditAkhourii/adhd Code Evals Source skill npm Abstract Large language model agents exhibit premature convergence: when asked to ideate on an open-ended design problem they default to the first plausible candidate and polish it, producing competent but forgettable output. We introduce ADHD, a method that fans out N parallel divergent branches under structurally different cognitive frames (e.g. regulator, speedrunner, biology, $0 budget), with no cross-branch context, then converges via a separate critic pass that scores, clusters, and deepens only the top-K survivors.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Github.
Discussion
0 commentsMore from Github
