Do Stronger Type Systems Make Coding Agents Better?
Comments on “Do Stronger Type Systems Make Coding Agents Better?”
3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked
The author tested whether strict type annotations help coding agents (Claude Code and Codex) work faster, cheaper, or more correctly, using a 3,000-line Python billing codebase, six code versions, 41 tasks and several hundred runs. Types made almost no difference to correctness, speed, or token use.
Why it ranked
No. 43 of 155 with a score of 0.64. The panel rated it highest on new information (0.92) and substantiated (0.80), lowest on significance (0.50).
- AI judgeJudge 2claude-haiku-5-5
- Substantive, well-scoped experiment on typed vs untyped code for coding agents with many runs; findings are clear but scope is small, single-author, Claude-favoring, with some overclaim risk.
- AI judgeJudge 3gpt-5.6-luna
- Original comparative study with 766 runs, reproducible methodology, codebase transformations, logs, and limitations. Highly relevant to agent-assisted development, though small task sets and model-specific design limit generalization; conclusions occasionally overstate beyond the evidence.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.64total under rubric v2.2
- Significance0.50
- Reader interest0.75
- Audience breadth0.63
- Practical relevance0.69
- Substantiated0.80
- Technical depth0.75
- New information0.92
- Counts against: Promotional0.04
- Counts against: Overclaim0.30
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| Significance | 0.58 | 0.35 | 0.58 | 0.50 | 0.23 | +0.24 | +0.121 |
| Reader interest | 0.70 | 0.70 | 0.86 | 0.75 | 0.16 | +0.26 | +0.196 |
| Audience breadth | 0.77 | 0.50 | 0.63 | 0.63 | 0.27 | +0.13 | +0.082 |
| Practical relevance | 0.76 | 0.60 | 0.72 | 0.69 | 0.15 | +0.22 | +0.152 |
| Substantiated | 0.97 | 0.60 | 0.83 | 0.80 | 0.37 | +0.05 | +0.040 |
| Technical depth | 0.90 | 0.60 | 0.76 | 0.75 | 0.29 | +0.05 | +0.038 |
| New information | 0.89 | 0.90 | 0.98 | 0.92 | 0.09 | +0.05 | +0.046 |
| Promotional | 0.05 | 0.05 | 0.01 | 0.04 | 0.04 | −0.19 | −0.007 |
| Overclaim | 0.21 | 0.30 | 0.38 | 0.30 | 0.17 | −0.11 | −0.033 |
| Sum of contributions | +0.635 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.635 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-10-28/43/"><img src="https://rubric.hrih.si/badge/2026-10-28/43.svg" alt="Ranked #43 on Rubric, 28 Oct 2026" width="245" height="28"></a>Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 4 pts, not on front page
Recorded for context only, checked 28 Oct 2026, 06:47 UTC. It played no part in the ranking.