Start free trial

Five minutes.
AI, made clearer.

Ask what was timed.

Read the briefing

Members read every story in the archive, re-verified. Sources and corrections stay open to everyone.

What the pen marks mean
  • rememberThe takeaway: what to remember
  • figureThe number that matters
  • evidenceThe evidence line
  • rulingThe ruling
  • claimThe claim that fails

Each mark is drawn once, as you reach it. Numbers are only circled when they come from the story’s own figures.

GitHub's famous number says Copilot users finished a task 55% faster. A Harvard paper tracking 718 firms finds AI agents raised lines of code 30%, while finished work showed no statistically significant change.

A developer at a desk beside a tall pile of paper pages, with a narrow doorway behind where a single reviewer sits.
AI illustration

GitHub's 2022 post says developers using Copilot finished a task 55% faster.

Members · 30 days free

See what the evidence actually shows.

Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

  1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
  2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
  3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing hello@bskiller.com. This voluntary refund does not limit your rights under the Australian Consumer Law.

By starting your trial, you agree to the Terms.

BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

Open the evidence4 source pages

The claim we checked

Developers who used GitHub Copilot completed the task 55% faster than the developers who didn't use GitHub Copilot.

These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

github.blog ↗
55% faster than the developers who didn’t use GitHub Copilot
github.blog ↗
timed how long it took them to write an HTTP server in JavaScript
arxiv.org ↗
completed the task 55.8% faster than the control group
fion.ac ↗
a novel proprietary dataset from an engineering analytics platform, covering 300 million work events
fion.ac ↗
across 718 firms. We use a staggered difference-in-differences design
fion.ac ↗
However, productivity gains do not fully pass through to changes in software output or employment.
fion.ac ↗
For AI agents, this incomplete pass-through reflects a bottleneck from code review
arstechnica.com ↗
Harvard University researchers Fiona Chen and James Stratton made use of aggregated analytics data from
arstechnica.com ↗
leads to a 30 percent increase in total lines of code generated
arstechnica.com ↗
did not change in a statistically significant way after AI tools were introduced
arstechnica.com ↗
the resolution rate for Issues and Epics (i.e. wholesale software features) tracked by tools like Jira
arstechnica.com ↗
balloons 49 percent on average after AI agents are introduced
arstechnica.com ↗
a 14 percent increase in the share of workers performing code reviews
arstechnica.com ↗
absorbed by downstream constraints in the production process
arstechnica.com ↗
significant updates and upgrades to their output even since this study
Open this check in the full collection →
The idea, illustrated01

Go inside the check.

    The full explanation appears when your reading access is confirmed.
    Next: Rebuilt Office, about half done

    A solo developer rebuilt Word, Excel and PowerPoint with Claude, the post says. The WordCraft README says it is about 60% of the way to replacing Word for real work, and the spreadsheet app says 50%.

    A shiny new office building facade propped up by wooden beams, with an empty lot visible behind it.
    AI illustration

    A solo developer rebuilt Word, Excel and PowerPoint as free open-source apps using Claude.

    Members · 30 days free

    See what the evidence actually shows.

    Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

    1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
    2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
    3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing hello@bskiller.com. This voluntary refund does not limit your rights under the Australian Consumer Law.

    By starting your trial, you agree to the Terms.

    BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

    Open the evidence5 source pages

    The claim we checked

    A solo developer rebuilt Word, Excel and PowerPoint as free open-source apps using Claude.

    These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

    github.com ↗
    an open-source, clean-room reimplementation of Microsoft Word, rebuilt in pure Rust
    digitaltoday.co.kr ↗
    All three apps were built with Anthropic's Claude.
    digitaltoday.co.kr ↗
    Thomas posted the apps on the ArtCraft GitHub so anyone can download them for free.
    github.com ↗
    we estimate it is about 60% of the way to replacing all of Word for real work
    github.com ↗
    Beta needs testing against real-world .docx files
    github.com ↗
    our AI image and video studio
    github.com ↗
    the core workflow works end to end, but depth, fidelity and file compatibility are still rough
    github.com ↗
    we estimate it is about 50% ready for real work
    github.com ↗
    is about 55% of the way to replacing PowerPoint for real work
    thurrott.com ↗
    ArtCraft claims that the current version of this app is 87 percent complete for ribbon-based commands and about 62 percent overall
    digitaltoday.co.kr ↗
    criticism followed that it was vibe-coding software and AI low-quality content, or AI slop
    digitaltoday.co.kr ↗
    most of the apps still fall far short of the level of the products they aim to replace
    Open this check in the full collection →
    The idea, illustrated02

    Go inside the check.

      The full explanation appears when your reading access is confirmed.
      Next: Skip button, not grit

      Ten minutes of AI erodes your ability to persist, Berkeley's headline says. The study measured skipped problems on a paid online task, three test problems each, and its largest trial found the skip gap was not significant.

      A small kitchen timer sits beside a tiny stack of three worksheets under a giant magnifying glass.
      AI illustration

      Berkeley News headlined that using AI for just 10 minutes erodes your ability to persist at hard things.

      Members · 30 days free

      See what the evidence actually shows.

      Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

      1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
      2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
      3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing hello@bskiller.com. This voluntary refund does not limit your rights under the Australian Consumer Law.

      By starting your trial, you agree to the Terms.

      BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

      Open the evidence4 source pages

      The claim we checked

      Using AI for just 10 minutes erodes your ability to persist at hard things

      These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

      news.berkeley.edu ↗
      Using AI for just 10 minutes erodes your ability to persist at hard things
      news.berkeley.edu ↗
      Whereas the AI users got answers wrong or gave up entirely when the AI assistance was withdrawn, the group without AI stuck with the task
      news.berkeley.edu ↗
      designed a series of randomized controlled trials and recruited 1,222 participants online
      news.berkeley.edu ↗
      You almost couldn’t ask for a clearer story
      arxiv.org ↗
      we find that although AI assistance improves performance in the short-term, people perform significantly worse without AI and are more likely to give up
      arxiv.org ↗
      Notably, these effects emerge after only brief interactions with AI (approximately 10 minutes).
      arxiv.org ↗
      participants were asked to solve 3 additional fraction problems
      arxiv.org ↗
      We recruited 667 US-based participants from the online research platform Prolific and paid them $3.40 for participation
      arxiv.org ↗
      but the result was not significant
      arxiv.org ↗
      Persistence costs are concentrated among those who obtain direct solutions
      arxiv.org ↗
      their payment didn’t depend on how many questions they solve correctly
      arxiv.org ↗
      our study took approximately 13 minutes to complete
      psychologytoday.com ↗
      Ten minutes of AI use erodes the persistence that 10,000 hours demands.
      psychologytoday.com ↗
      Not after months of dependency. After a single brief interaction.
      Open this check in the full collection →
      The idea, illustrated03

      Go inside the check.

        The full explanation appears when your reading access is confirmed.
        The finish line

        Edition complete

        You’re up to speed.

        That’s the 13 Oct 2026 briefing. Keep the useful bits. Leave the noise.

        Reading estimate: 541 words at 200 words per minute. Source quotes and the optional sections below add reading time.

        Get tomorrow's check in your inbox.

        One AI claim a night, checked against independent sources, receipts attached. Free.

        Free email updates. Unsubscribe any time.