What even is code mode?
Comments on “What even is code mode?”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
Polylane describes "code mode," where an agent gets one tool that runs model-written code and exposes its other tools as functions, so only the script's return value enters the context. Using its own 30-day logs, the company reports over 70% of tool calls run this way, a 14x reduction in tool output reaching the model, and a 98.8% script success rate. Scripts run in network-less Cloudflare isolates. These figures are self-reported and come from largely one model. Writes and binary data are still handled as direct calls.
Why it ranked
No. 15 of 139 with a score of 0.69. The panel rated it highest on technical depth (0.90) and practical relevance (0.86), lowest on significance (0.62). The judges disagreed most on new information (spread 0.51), so the story is marked contested. Placed No. 15, No. 21 by score alone: the fourth AI & machine learning story of the day (−0.020). The placement terms move stories; they are not part of the score.
- AI judgeJudge 2claude-haiku-5-5
- Detailed engineering write-up on code mode for agent tool use with concrete data and architecture; practical for builders, though the idea is an emerging pattern rather than a landmark, and some data is self-reported.
- AI judgeJudge 3gpt-5.6-luna
- A detailed engineering account of executing tool calls inside isolated code, with architecture, implementation patterns, tradeoffs, and 30-day usage data. It is highly actionable and novel, though conclusions generalize beyond one agent’s 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
0.69total under rubric v2.3Contestedjudges differ by 0.51 on new information
- Significance0.62
- Reader interest0.83
- Audience breadth0.68
- Practical relevance0.86
- Substantiated0.81
- Technical depth0.90
- New information0.76
- Counts against: Promotional0.21
- Counts against: Overclaim0.31
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) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.69 | 0.45 | 0.72 | 0.62 | 0.27 | +0.24 | +0.148 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.79 | 0.80 | 0.90 | 0.83 | 0.11 | +0.26 | +0.216 |
| Audience breadthHow many people who build technology does it concern? | 0.65 | 0.67 | 0.72 | 0.68 | 0.07 | +0.13 | +0.088 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.90 | 0.80 | 0.88 | 0.86 | 0.10 | +0.22 | +0.189 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.91 | 0.70 | 0.82 | 0.81 | 0.21 | +0.05 | +0.041 |
| Technical depthHow technically substantive is it? | 0.99 | 0.85 | 0.86 | 0.90 | 0.14 | +0.05 | +0.045 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.86 | 0.45 | 0.96 | 0.76 | 0.51 , above the contested threshold | +0.05 | +0.038 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.09 | 0.35 | 0.18 | 0.21 | 0.26 | −0.19 | −0.039 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.34 | 0.30 | 0.28 | 0.31 | 0.06 | −0.11 | −0.034 |
| Sum of contributions | +0.692 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.692 | ||||||
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/2026-11-05/15/"><img src="https://rubric.hrih.si/badge/2026-11-05/15.svg" alt="Ranked #15 on Rubric, 5 Nov 2026" width="238" height="28"></a>Claim check
Claim check: 1 matches source · 7 no source given. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
No source givenNumber
Polylane's agent runs more than 70% of its tool calls inside executeCode, and that share was stable over the 30 days.
The article gives no link or citation for it.
The article says “More than 70% of the agent’s tool calls ran inside executeCode, and that share was pretty stable throughout the 30 days”
The source says No source to compare with.
No source givenNumber
Describing all of Polylane's few hundred tools in full in the prompt would take close to 300,000 tokens, versus about 13k for the code mode index.
The article gives no link or citation for it.
The article says “describing all of them in full in the prompt would take close to 300,000 tokens before the conversation even starts”
The source says No source to compare with.
No source givenNumber
Code mode resulted in 14x less tool output in the model's context over 30 days.
The article gives no link or citation for it.
The article says “14x less tool output in the model's context”
The source says No source to compare with.
No source givenNumber
Code mode scripts ran with a 98.8% success rate.
The article gives no link or citation for it.
The article says “Overall, code mode scripts ran with a 98.8% success rate”
The source says No source to compare with.
No source givenNumber
More than 95% of failed calls inside scripts occurred in scripts that still finished and returned a result.
The article gives no link or citation for it.
The article says “More than 95% of the failed calls inside scripts happened in a script that still finished and returned a result”
The source says No source to compare with.
No source givenNumber
Each Polylane script runs in a fresh V8 isolate from Dynamic Cloudflare Workers with no network access, and an isolate starts in a few milliseconds.
The article gives no link or citation for it.
The article says “Each script runs in a fresh V8 isolate from Dynamic Cloudflare Workers, with no network access at all”
The source says No source to compare with.
Matches sourceVersion
Claude's API offers programmatic tool calling, where Claude writes code that runs in Anthropic's sandbox and each call to a user tool pauses the script and returns to the application as a normal tool call.
The cited source says the same thing.
The article says “Claude writes code that runs in Anthropic’s sandbox, and each call to one of your tools pauses the script and comes back to your application as a normal tool call”
The source says “When a tool function is called, code execution pauses and the API returns a tool_use block”
Source cited by the article: platform.claude.com/docs/en/agents-and-tools/tool-use/programmatic-tool-calling
No source givenGeneralization
Most coding agents chose to write code when given the choice on Polylane's MCP server, while Codex mostly calls tools one by one.
The article gives no link or citation for it.
The article says “most coding agents chose to write code when they had the choice. Codex is the exception, it mostly calls tools one by one”
The source says No source to compare with.
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 3 pts, not on front page
Recorded for context only, checked 5 Nov 2026, 05:46 UTC. It played no part in the ranking.