Stake and Rope

Please Do Not Vibe

17 min · 10. juni 2026
episode Please Do Not Vibe cover

Beskrivelse

Rsync, the file synchronization utility that has quietly underpinned essentially every backup system in the Unix and Linux world since the mid-nineties, shipped a release earlier this year with regressions affecting incremental backup workflows. Users digging through the commit history found dozens of commits attributed to "tridge and claude" — that's Andrew Tridgell, the project's creator and a foundational figure in open-source infrastructure, working alongside Anthropic's Claude. A GitHub post titled, with the expletive sanitized, "Please Do Not Vibe Fuck Up This Software" lit the fuse. The Register reported it this week. Tridgell responded with a Medium piece called "Rsync and Outrage" defending his process. Forty years of software engineering experience. Every commit reviewed personally. AI tooling adopted in response to a flood of AI-generated security reports consuming his maintenance time. The defense is the strongest possible version of the position — the original maintainer is also the reviewer, the usual AI-PR concerns about review capacity don't apply, the tool adoption was a rational response to real operational pressure. The panel takes the defense seriously and engages with it on its merits. And the backups have regressions. That's the transaction the panel keeps returning to. The maintainer's standing is intact, the process defense holds, the response to the security-report flood was reasonable, and the incremental backup paths broke. The standard objection to AI-assisted contributions in open source — review capacity — doesn't apply here. So if the regressions still happen, the conclusion has to be something else: the kind of code being produced is harder to review than the code being replaced, the rewrite was the wrong unit of work for the tool, or the test coverage gap was always there and got surfaced by being broken. None of those are character flaws. They're decisions that produced an outcome. And the outcome is that some number of people are going to find out their incremental backups don't restore at three in the morning when they try to. Source Article "Please do not vibe f$%& up this software": Broken backups spark AI coding row in rsync communit [https://www.theregister.com/ai-and-ml/2026/06/04/please-do-not-vibe-f-up-this-software-broken-backups-spark-ai-coding-row-in-rsync-project/5251189] — Carly Page, The Register, June 4, 2026 Panel * The Legacy Sysadmin * The DBA * The Startup Founder * The Goat Farmer's Counsel

Kommentarer

0

Vær den første til å kommentere

Registrer deg nå og bli medlem av Stake and Rope sitt community!

Prøv gratis

Prøv gratis i 14 dager

99 kr / Måned etter prøveperioden. · Avslutt når som helst.

  • Eksklusive podkaster
  • 20 timer lydbøker i måneden
  • Gratis podkaster

Alle episoder

18 Episoder

episode Please Do Not Vibe cover

Please Do Not Vibe

Rsync, the file synchronization utility that has quietly underpinned essentially every backup system in the Unix and Linux world since the mid-nineties, shipped a release earlier this year with regressions affecting incremental backup workflows. Users digging through the commit history found dozens of commits attributed to "tridge and claude" — that's Andrew Tridgell, the project's creator and a foundational figure in open-source infrastructure, working alongside Anthropic's Claude. A GitHub post titled, with the expletive sanitized, "Please Do Not Vibe Fuck Up This Software" lit the fuse. The Register reported it this week. Tridgell responded with a Medium piece called "Rsync and Outrage" defending his process. Forty years of software engineering experience. Every commit reviewed personally. AI tooling adopted in response to a flood of AI-generated security reports consuming his maintenance time. The defense is the strongest possible version of the position — the original maintainer is also the reviewer, the usual AI-PR concerns about review capacity don't apply, the tool adoption was a rational response to real operational pressure. The panel takes the defense seriously and engages with it on its merits. And the backups have regressions. That's the transaction the panel keeps returning to. The maintainer's standing is intact, the process defense holds, the response to the security-report flood was reasonable, and the incremental backup paths broke. The standard objection to AI-assisted contributions in open source — review capacity — doesn't apply here. So if the regressions still happen, the conclusion has to be something else: the kind of code being produced is harder to review than the code being replaced, the rewrite was the wrong unit of work for the tool, or the test coverage gap was always there and got surfaced by being broken. None of those are character flaws. They're decisions that produced an outcome. And the outcome is that some number of people are going to find out their incremental backups don't restore at three in the morning when they try to. Source Article "Please do not vibe f$%& up this software": Broken backups spark AI coding row in rsync communit [https://www.theregister.com/ai-and-ml/2026/06/04/please-do-not-vibe-f-up-this-software-broken-backups-spark-ai-coding-row-in-rsync-project/5251189] — Carly Page, The Register, June 4, 2026 Panel * The Legacy Sysadmin * The DBA * The Startup Founder * The Goat Farmer's Counsel

10. juni 202617 min
episode Continue to Function cover

Continue to Function

Microsoft Office 2019 and Office 2021 for Mac, both sold as perpetual licenses, will drop into "reduced functionality mode" on July 13, 2026. After that date, customers who paid in full for the software can open files and view them, but cannot edit or save. The cause is a license-validation certificate scheduled to expire on that date — a date that was baked into the binary the day the software shipped. When Office 2019 reached end of support in October 2023, Microsoft's own support page told customers the apps would "continue to function." Last week, OSnews and the Consumer Rights Wiki noticed that the page had been quietly rewritten. The continue-to-function language was gone. No email to customers. No press release. The promise was edited out before the kill switch was scheduled to flip. The panel's argument lands on the simpler reading of the transaction. The customers paid. The software works. Microsoft is going to turn it off, remotely, because the customers won't pay again. Everything else — the EULA language, the business-model defense, the lifecycle framing — is decoration. The industry has spent thirty years smearing the distinction between owning a thing and leasing access to a thing, and now we live in the version where the lease can be ended whenever the vendor wants, on a date the vendor set, with a promise that gets edited out before the date arrives. Source Articles Microsoft Deliberately Bricking All Office For Mac 2019/2021 Installations [https://tech.slashdot.org/story/26/06/02/1725210/microsoft-deliberately-bricking-all-office-for-mac-20192021-installations] — Slashdot, June 2, 2026. Coverage of OSnews and Consumer Rights Wiki reporting on the certificate expiration scheduled for Office 2019 and Office 2021 for Mac, the rewriting of the Microsoft support page to remove the continue-to-function language, the "reduced functionality mode" mechanic, and Microsoft's four officially recommended responses for affected customers. Panel * The Legacy Sysadmin * The DBA * The Startup Founder * The Goat Farmer's Counsel

8. juni 202617 min
episode Untapped Means Broken cover

Untapped Means Broken

Google is quietly emailing Android developers with offers to buy their source code. The pitch is "unlock new revenue" and "help transform tools and products." The actual ask is access to production codebases and archived side projects, to train Google's AI coding tools. Jason Koebler at 404 Media obtained the email and broke the story this week. The framing is partnership. The structure is procurement. The license is non-exclusive — which means the developer keeps the IP and Google gets a permanent, non-revocable right to use the code as training data. Once a model is trained on a codebase, the codebase is in the weights. You can sue Google for breach of contract; you cannot sue the model. The panel's argument lands on the simpler reading: Google has a product that doesn't work as well as the competition, and the data they need to make it better lives in the heads and laptops of small developers. The Reddit deal at sixty million produced Gemini telling users to eat glue. The phase-two content-owner deals didn't produce the quality the models needed. So Google is in phase three — cold-emailing individual developers because what the coding models actually need is non-public, real-world production code, and there is no aggregator who owns that. Nobody's getting rich. Google's getting incrementally less behind. And the code, including the parts the developer forgot was in there, becomes training data forever. Source Article Google is emailing Android developers asking to buy their source code [https://www.404media.co/google-is-quietly-buying-code-from-play-store-developers-to-train-ai/] — 404 Media, Jason Koebler, June 2026. Panel * The DBA * The Startup Founder * The Paranoid CISO * The Goat Farmer's Counsel

5. juni 202617 min
episode Should Be Making Executives Uncomfortable cover

Should Be Making Executives Uncomfortable

Two convergent reports landed in the same week with the same conclusion. Bain & Company published survey findings on June 1 reporting that corporate AI investments are based on cost savings that haven't arrived. The consultancy told its own clients the situation "should be making executives uncomfortable." The same week, developer telemetry firm Faros published a study of 22,000 developers and 4,000 teams measuring what LLM-assisted coding actually does to operational metrics. The numbers: lead time for changes up nearly five-fold. Deployment frequency down eleven percent. Defect rates up fifty percent. System throughput, calculated via Little's Law, down somewhere between seventy and eighty percent. The consultancy that sold the savings model is telling clients to worry. The measurement firm is telling engineers what they already suspected. The procurement decks were written before either of those was measurable. The decks for next year are being written now, by the same people, on the same assumptions. The panel argues toward an editorial center the show has been on the record about for several episodes: the AI ROI isn't materializing the way the procurement narrative promised, and both the consulting class and the measurement class are now saying so. The convergence is the news. The pattern — consultancy sells the strategy, strategy doesn't pay off, consultancy sells the diagnostic — is older than the technology in the middle. Source Articles AI Savings Misses 'Should Be Making Executives Uncomfortable,' Bain Says [https://www.bloomberg.com/news/articles/2026-06-01/bain-finds-corporate-ai-investments-based-on-returns-that-haven-t-arrived] — Bloomberg, June 1, 2026. Bain & Company's survey of executives finding that AI deployment hasn't delivered the productivity gains modeled into business cases, and the consultancy's framing that executives should be worried about the gap between projected and realized savings. Talk Is Cheap: The Operational Impact of LLM Use, [https://unessays.substack.com/p/talk-is-cheap] May 31, 2026. Coverage of the Faros.ai [http://Faros.ai] study measuring operational metrics across 22,000 developers and 4,000 teams using LLM-assisted coding tools. The study finds decreased deployment frequency, increased lead time for features, and increased cost of defects, with a calculated system throughput drop between seventy and eighty percent using Little's Law. Panel * The Burnt-Out SRE * The DBA * The Startup Founder * The Goat Farmer's Counsel

3. juni 202618 min
episode Fifteen Characters cover

Fifteen Characters

Microsoft's May 2026 security update for Windows Server 2016 breaks domain controller discovery, but only if your server's hostname is exactly fifteen characters long. Not fourteen. Not sixteen. Fifteen. The Register has it. Microsoft's official guidance is that the issue is under investigation. There is no workaround listed. Fifteen is not a random number. It's the NetBIOS name length limit, documented since 1993, taught in week two of every Windows admin certification, and built around by every Windows shop for thirty-three years. The sixteenth byte was reserved for the resource type — file server, workstation, domain controller — and the limit has been load-bearing since the Clinton administration. So when a Microsoft patch in 2026 breaks specifically at fifteen characters, the bug isn't that the limit exists. The bug is that the people maintaining the code forgot what the code did. The institutional memory of an industry, the panel keeps landing on, is held by the people who use the product, not the people who make it. The customer is the test environment. They've been the test environment since at least Vista. They're going to keep being the test environment, because the people who knew the code left, and the people maintaining it now don't, and the only reason it still works at all is because customers find the bugs for free. Topics * The NetBIOS fifteen-character limit and the hex codes underneath it (20 for file servers, 00 for workstations, 1C for domain controllers) * Off-by-one errors at documented boundaries — the five test cases anyone who's done the work for six months knows * The companion bug from the same patch cycle: Windows 11 installs failing because the EFI System Partition wasn't big enough * What the rename-the-server "workaround" actually breaks (SPNs, Kerberos tickets, GPO targeting, backup catalogs) * Why Server 2016 is still twenty percent of the Windows server install base, and why none of those shops can move * The patch-Tuesday cycle as an institutional pattern — install Tuesday, wait for Friday, don't tell management Source Article Microsoft tests the 15-character limit of Windows Server admins' patience [https://www.theregister.com/oses/2026/05/28/microsoft-tests-the-15-character-limit-of-windows-server-admins-patience/5247943] — Richard Speed, The Register, May 2026. Reporting on Microsoft's May 12 security update for Windows Server 2016, the documented but unexplained behavior breaking DCLocator at the NetBIOS hostname-length boundary, the lack of a published workaround, and the companion failure of the same patch cycle to install on Windows 11 devices with undersized EFI System Partitions. Microsoft's official position remains that the issue is under investigation. Panel * The Legacy Sysadmin * The Burnt-Out SRE * The DBA * The Goat Farmer's Counsel

1. juni 202616 min