<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/">
    <channel>
        <title>Gerra Insights</title>
        <link>https://www.gerra.com</link>
        <description>Stay informed with product updates, company news, and insights on robotics and automation.</description>
        <lastBuildDate>Sat, 15 Aug 2026 02:07:33 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <image>
            <title>Gerra Insights</title>
            <url>https://www.gerra.com/favicon.ico</url>
            <link>https://www.gerra.com</link>
        </image>
        <copyright>All rights reserved 2026</copyright>
        <item>
            <title><![CDATA[Partial Credit Is What Turns a Benchmark Into an Environment]]></title>
            <link>https://www.gerra.com/research/partial-credit-makes-an-environment</link>
            <guid>https://www.gerra.com/research/partial-credit-makes-an-environment</guid>
            <pubDate>Thu, 13 Aug 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[A pass/fail task is a measurement. A task that scores partial progress is a reward signal. Most published benchmarks are the first thing being sold as the second.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Automated De-Identification Passes Corpora That Humans Reject]]></title>
            <link>https://www.gerra.com/research/why-automated-deidentification-passes-what-humans-reject</link>
            <guid>https://www.gerra.com/research/why-automated-deidentification-passes-what-humans-reject</guid>
            <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[The scanner and the pipeline share assumptions, so the scanner confirms what the build already believes. The leaks that survive are the ones nobody wrote a rule for.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Plagiarist Gate: Proving a Benchmark Cannot Be Passed by Recall]]></title>
            <link>https://www.gerra.com/research/the-plagiarist-gate</link>
            <guid>https://www.gerra.com/research/the-plagiarist-gate</guid>
            <pubDate>Tue, 28 Jul 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[Contamination is usually argued. It can be demonstrated instead: run the real library through your own grader and require it to fail.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Read a Data Vendor's Data Room]]></title>
            <link>https://www.gerra.com/research/how-to-read-a-data-room</link>
            <guid>https://www.gerra.com/research/how-to-read-a-data-room</guid>
            <pubDate>Sun, 19 Jul 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[What a serious data room contains and how to inspect it: dated schemas, delivery-shape samples, checksums, reconcilable manifests, stated limits.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Audit a Codebase Dataset Before You License It]]></title>
            <link>https://www.gerra.com/research/how-to-audit-a-codebase-dataset</link>
            <guid>https://www.gerra.com/research/how-to-audit-a-codebase-dataset</guid>
            <pubDate>Tue, 14 Jul 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[A buyer checklist for repository completeness, Git integrity, provenance, deduplication, privacy, tests, and reproducible release metrics.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Git History Becomes a Code-Model Evaluation Dataset]]></title>
            <link>https://www.gerra.com/research/git-history-datasets-for-code-model-evaluation</link>
            <guid>https://www.gerra.com/research/git-history-datasets-for-code-model-evaluation</guid>
            <pubDate>Tue, 07 Jul 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[A practical method for turning authentic commits, patches, and tests into point-in-time coding-agent evaluations without leaking the answer.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Code Data Do Codex, Claude Code, and SWE Agents Need?]]></title>
            <link>https://www.gerra.com/research/code-data-for-codex-claude-code-and-swe-agents</link>
            <guid>https://www.gerra.com/research/code-data-for-codex-claude-code-and-swe-agents</guid>
            <pubDate>Tue, 30 Jun 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[A model-agnostic framework for choosing repository data for terminal coding agents: context, history, environments, tests, and honest evaluation boundaries.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Repository-Level Code Data for Coding Agents]]></title>
            <link>https://www.gerra.com/research/repository-level-code-data-for-coding-agents</link>
            <guid>https://www.gerra.com/research/repository-level-code-data-for-coding-agents</guid>
            <pubDate>Tue, 23 Jun 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[Why complete repositories, cross-file context, tests, and change history are more useful for coding-agent training than disconnected source files.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agents Fail at Office Work Because the Data Never Existed]]></title>
            <link>https://www.gerra.com/research/organizational-workflow-data-for-agents</link>
            <guid>https://www.gerra.com/research/organizational-workflow-data-for-agents</guid>
            <pubDate>Tue, 09 Jun 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[Models trained on the public web have never seen how organizations run. What consented, cross-tool company data adds to agent training and evals.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Reading Federal Software Spending Before the Street Does]]></title>
            <link>https://www.gerra.com/research/federal-contract-data-for-software-revenue</link>
            <guid>https://www.gerra.com/research/federal-contract-data-for-software-revenue</guid>
            <pubDate>Tue, 02 Jun 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[US federal software procurement is public record. A method note on turning USAspending and FPDS-NG awards into a ticker-level revenue signal.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Prediction Markets Know About Your Portfolio]]></title>
            <link>https://www.gerra.com/research/prediction-market-data-for-equity-research</link>
            <guid>https://www.gerra.com/research/prediction-market-data-for-equity-research</guid>
            <pubDate>Tue, 26 May 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[How live event probabilities map to exposed securities with signed direction and sensitivity, and why point-in-time probability history matters.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Evaluate Retail Sentiment Data Before You Backtest It]]></title>
            <link>https://www.gerra.com/research/how-to-evaluate-social-sentiment-data</link>
            <guid>https://www.gerra.com/research/how-to-evaluate-social-sentiment-data</guid>
            <pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[A buyer checklist for point-in-time integrity, deleted-message survivorship, human vs NLP labels, spam filtering, and leakage-safe backtests.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Identical GPUs Rent at Wildly Different Prices]]></title>
            <link>https://www.gerra.com/research/why-gpu-rental-prices-vary</link>
            <guid>https://www.gerra.com/research/why-gpu-rental-prices-vary</guid>
            <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[The same H100 rents for $1.20 to $22.02 per hour. What a 38-provider panel shows about GPU price dispersion, and why the equity signal is a null.]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Sim2Real Gap Costs Robotics Billions Annually]]></title>
            <link>https://www.gerra.com/research/the-sim2real-gap-costs-robotics-billions-annually</link>
            <guid>https://www.gerra.com/research/the-sim2real-gap-costs-robotics-billions-annually</guid>
            <pubDate>Thu, 19 Jun 2025 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[Simulation-to-reality transfer failures create multi-billion dollar losses across the robotics industry, with success rates plummeting from 90% in simulation to under 20% on real hardware, extending development cycles by 6-12 months.]]></content:encoded>
            <enclosure url="https://www.gerra.com/research/sim2-real-gap/what-is-the-sim2real-gap.png" length="0" type="image/png"/>
        </item>
        <item>
            <title><![CDATA[The Global Humanoid Robot Supply Chain: Architecture, Economics, and Strategic Implications]]></title>
            <link>https://www.gerra.com/research/humanoid-robot-supply-chain</link>
            <guid>https://www.gerra.com/research/humanoid-robot-supply-chain</guid>
            <pubDate>Mon, 16 Jun 2025 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[Chinese manufacturing controls 63% of key humanoid robot supply chains while component costs plummet 40% annually. An in-depth analysis of the supply chain dynamics reshaping the $66 billion humanoid robotics market.]]></content:encoded>
            <enclosure url="https://www.gerra.com/research/humanoid-robot-supply-chain/assembly-line.png" length="0" type="image/png"/>
        </item>
        <item>
            <title><![CDATA[The Complete Robot Activation Playbook]]></title>
            <link>https://www.gerra.com/research/robot-activation-playbook</link>
            <guid>https://www.gerra.com/research/robot-activation-playbook</guid>
            <pubDate>Sun, 08 Jun 2025 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[How top agencies 10X event ROI without doubling budgets - the comprehensive guide to deploying humanoid robots for experiential marketing campaigns.]]></content:encoded>
            <enclosure url="https://www.gerra.com/screenshots/tron1-atlanta.png" length="0" type="image/png"/>
        </item>
        <item>
            <title><![CDATA[The QDD Actuator Renaissance]]></title>
            <link>https://www.gerra.com/research/qdd-actuator-renaissance</link>
            <guid>https://www.gerra.com/research/qdd-actuator-renaissance</guid>
            <pubDate>Sun, 01 Jun 2025 00:00:00 GMT</pubDate>
            <content:encoded><![CDATA[How a 2016 MIT breakthrough spawned the era of $3k humanoids – and why China now dominates the torque supply chain]]></content:encoded>
            <enclosure url="https://www.gerra.com/research/cost-curve.png" length="0" type="image/png"/>
        </item>
    </channel>
</rss>