Hacker News Reader

> "everything is moving very fast"

Then slow down.

With this objective lack or control, sooner or later your LLM experiments in production will drive into a wall instead of hitting a little pothole like this diagram.

And at the same time, they have time to quickly brush it off with "looks like a vendor" even though people are still investigating. Yes, we can see it's moving really fast, probably "move fast break things" been infecting Microsoft, users are leaving Microsoft behind because everything is breaking then clueless VPs blame it on moving too fast?

Jokes on you, I’ll cash out by then and move to the next gig.

- Put on your seatbelts, man!

- I can't, moving too fast!

"Driving into a wall" is still a positive outcome. It's just as likely to drive into a crowd.

Serious loss of life is a plausible LLM outcome, particularly for Microsoft who does both operating systems (incidents can be much worse than the Crowdstrike bricking) and chatbot assistants that can offer lethal advice. Catastrophic property damage is hopefully more likely.

> This excuse is hollow to me. In an organization of this size, it takes multiple people screwing up for a failure to reach the public, or at least it should.

Completely with you on this, plus I would add following thoughts:

I don't think the size of the company should automatically be a proxy measure for a certain level of quality. Surely you can have slobs prevailing in a company of any size.

However - this kind of mistake should not be happening in a valuable company. Microsoft is currently still priced as a very valuable company, even with the significant corrections post Satyas crazy CapEx commitments from 2 weeks ago.

However it seems recently the mistakes, errors and "vendors without guidelines" pile up a bit too much for a supposedly 3-4T USD worth company, culminating in this weird random but very educational case. If anything, it's indicator that Microsoft may not really be as valuable as it is currently still perceived.

You’re incorrect on how the publishing process works. If a vendor wrote the document, it has a single repo owner (all those docs are in github) that would need to sign off on a PR. There isn’t multiple layers or really any friction to get content on learn.msft.

I suggested that if there is no review process, it is a systemic issue, and that if there is a review process that failed to catch something this egregious, it is a systemic issue. My supposition is that regardless of how the publishing process works, there is a systemic failure here, and I made no claims as to how it actually works, so I'm not sure where the "you're incorrect on how it works" is coming from.

You said it takes multiple people screwing up, implying that publishing content had multiple gates/reviewers.

It doesn’t.

But if there are no gates, doesn't that mean the people who should have put the gates in there screwed up?

There is no singular publishing org at MSFT. Each product publishes its own docs, generally following a style guide. But the doc process is up to the doc owner(s).

I think you're barking up the wrong tree here.

What?

This is how it works. There are too many people here like the op that make assumptions on what the process is/should be.

There is no such thing as "making an assumption" on what a process "should be". I am asserting what it should be. A multi-trillion dollar company should absolutely have a robust review process in place. If one single person can submit plagiarised and defective material onto an official platform that implicates the company as a whole in copyright infringement, management has failed, ergo multiple people have failed, ergo the failure is systemic.

It is extremely well-known that individual humans make mistakes. Therefore, any well-functioning system has guards in place to catch mistakes, such that it takes multiple people making mistakes for an individual mistake to cascade to system failure. A system that does not have these guards in place at all, and allows one individual's failure to immediately become a system failure, is a bad system, and management staff who implement bad systems are as responsible for their failure as the individual who made the mistake. Let us be grateful that you do not work in an engineering or aviation capacity, given the great lengths you are going to defend the "correctness" of a bad system.

My dog does this thing where she picks a stick and gets you to pull on it, and she will pull on her end, too. She gets very focused on it. Pulling on the stick is the most important thing to her in that moment, when in fact it's just a stick she chose to turn into this tug of war.

That's not entirely unlike what you're doing here. You latched onto a misunderstanding of OP's intent, and by making a thing out of it got people to pull back, and now you also keep tugging on your end.

Except she does it on purpose and enjoys it, while I think you did it inadvertently and you do not seem that happy. But then, you're not a dog, of course.

You could stop pulling on the stick. I do enjoy these doggy similes, though. :)

This is a perfect description. I've probably been the dog at some point.

p_ing, see my nearby comment about what we mean by "multiple". Does that comment make any false "assumptions"? Or, is it you who are mistaken, persistently failing to understand what your interlocutors are saying?

It can be hard to resist.

That seems to further make the case that it's a systemic problem.

The organization would have more guardrails in place if it prioritized "don't break things" over "move fast".

There have been no Gates at Microsoft for a long time.

I've seen better review processes in hobby projects

Neither deadlines nor cheap work for hire help any sort of review process, while an hobby project is normally done by someone who cares.

This is correct. It just takes one person to review it and you’re good to go.

There’s also a service that rates your grammar/clarity and you have to be above a certain score.

I'll quote the relevant part of the parent post:

> that is in itself a failure of the system

... and add some Beer flavor: POSIWID (the purpose of a system is what it does)

A postmortem for that but not Copilot in notepad.exe? Priorities…

I’d also love a post-mortem on their guide to pirating the entire Harry Potter series for AI use. (https://devblogs.microsoft.com/azure-sql/langchain-with-sqlv...)

I’ve lost trust in anything Microsoft publishes anymore.

An entire post mortem for a morged diagram is wild

post morgem

Right to morgue

It's post morgem time. [0]

[0] https://knowyourmeme.com/memes/its-morbin-time

*post morgem ti൬.

Morgem? I barely know 'em!

You could say the original author has enjoyed some post-morgem clarity.

Oldest trick in the book... Shoot the vendor.

> In either case -- no review process, or a failed review process -- the failure is definitionally systemic.

Ortho and grammar errors should have been corrected, but do you really expect a review process to identify that a diagram is a copy from another one some rando already published on the internet years ago?

plenty of people on the internet recognised it immediately, so sure, he may have been a rando when he created it, but not so much 15 years later..

Just that tiny image on his blog was enough for me to go "oh yeah, I used his diagram to explain this type of git workflow to colleagues a decade ago". Someone should have spotted that right away.

Did the one MSFT employee that “reviewed” it know of this image? If not, it doesn’t matter how many people “on the Internet” recognized this image.

I’ll never understand the implied projection.

(I don’t think this was reviewed closely if at all)

I would hope that the person who reviews their training on gitflow, knows something about gitflow. And if you know something about gitflow, it's not that strange to expect to recognise the most iconic gitflow diagram.

But even if you don't recognise the original, at least you should be able to tell that the generated copy is bullshit.

Again, I don't think this was reviewed. It was an assignment to a vendor 'write document and I'll hit publish'. There's a great chance the MSFT document _owner_ has no experience in the relevant area.

It’s not just a copy. It’s a caricature of a copy with a plenty of nonsense in it: typos and weird “text”, broken arrows, etc. Even a cursory look gives a feeling that something’s fishy.

"Legal reviewed it and did not flag any issues!"

This is the same Microsoft that promised to indemnify any of its customers sued over copyright lawsuits as a result of using its AIs. [0] So I'm sure legal reviewed it the same way, saying "Yep, our war chest is still ample".

[0]: https://www.reuters.com/technology/microsoft-defend-customer...

Weird text was already deemed acceptable by microsoft in their documentation as they machine translated most screenshots instead of recreating them in different locales, leading to the same problems as this image.

Shouldn't "where are we sourcing our content" be part of any publication review process?

The Large Laundering Machine, sorry, I mean the Large Language Model! Provenance? Where's that?

No. I'd expect that "continvouclous morging" gets caught.

That expectation is pure covfefe.

Here is the original: https://nvie.com/posts/a-successful-git-branching-model/

Here is the slop copy: https://web.archive.org/web/20251205141857/https://learn.mic...

The 'Time' axis points the wrong way, and is misspelled, using a non-existent letter - 'Tim' where the m has an extra hump.

It's pretty clear this wasn't reviewed at all.

Yes. This is expected at any serious company as intellectual property violations can have serious consequences.

I don't think the characterization of this being a diagram from "some rando" is accurate or fair.

The original content is highly influential... which should be self-evident by the fact it is being reproduced verbatim ten years later, and was immediately recognized.

> but do you really expect a review process to identify that a diagram is a copy from another one some rando already published on the internet years ago?

We aren't talking about just some random image from some random blog. The article we are talking about is about a specific topic, which when searched online one of the first is the article containing the original image (at least for google, bing seems to be really struggling to give me the article but under images it is again the first).

I would cut some slack if this were a really obscure topic almost noone talks about, but it's been a thing talked about in the programmer space for ages.

Yes? It's a famous diagram, at least in the world of Git workflows, so I would expect a reviewer of Microsoft's Git workflow documentation to be familiar with it.

(But the main issue is that the diagram is slop, not that it's a copy.)

I would personally expect review to evaluate for correctness. Such a review would have stopped this from being published. This diagram as published is literal nonsense.

Yeah, isn't this why we're told everything "moves so much slower at a bigco" than at a startup?

Seems like this is going to be the year of AI slop being released everywhere by Microsoft. Just wish they'd put as much effort into a post morten for this one as they're doing for a diagram on a blog post https://github.com/microsoft/onnxruntime/issues/27263#issuec...

LOL, calling Scott Hanselman a 'VP of something' is funny. Been listening to his stuff for years, even when I despised MS. Always seems genuinely nice. Probably one of the main reasons I these days have a more positive image of Microsoft.

Scott is definitely one of the good guys.

I don't this person, but immediately trying to foist blame for a really embarrassing screwup onto a "vendor" does not really sound like "good guy" behavior to me?

I thought the same thing. Scott is basically CEO of devrel at Microsoft.

Maybe we're of a different age to remember when Scott was super influential as a blogger / conference speaker, but even now he's not some random VP.

> everything is moving very fast right now

Now that's an interesting comment for him to include. The cynic in me could find / can think of lots of reasons from my YouTube feed as to why that might be so. What else is going on at Microsoft that could cause this sense of urgency?

My guess is there is some communication going out to every "manager", even the M1, that says this is your priority.

For example, I know of an unrelated mandate Microsoft has for its management. Anything security team analysis flags in code that you or your team owns must be fixed or somehow acceptably mitigated within the deadline specified. It doesn't matter if it is Newton soft json being "vulnerable" and the entire system is only built for use by msft employees. If you let this deadline slip, you have to explain yourself and might lose your bonus.

Ok so the remediation for the Newton soft case is easy enough that it is worth doing but the point is I have a conspiracy theory that internally msft has such a memo (yes, beyond what is publicly disclosed) going to all managers saying they must adopt copilot, whatever copilot means.

From the beginning, one of the advertising tricks they have used for AI is FOMO. I presume that is so they can sell you as much of it as they can before you realize its flaws.

Everybody's so worried about getting in on the ground floor of something that they don't even imagine it could be a massive flop.

Any excuse that tries to play down its own fault by pointing out other companies also have faults, is dishonest.

And that's exactly what happened here.

The VP blames a vendor of course, but didn't Microsoft recently announce they were going to vibe code everything? Because this image looks like it comes from the kind of company that thinks it can vibe code everything.

> This excuse is hollow to me. In an organization of this size, it takes multiple people screwing up for a failure to reach the public, or at least it should.

Only if this is considered a failure.

Native English speakers may not know, but for a very long time (since before automatic translation tools became adequate) pretty much all MSFT docs were machine translated to the user agent language by default. Initially they were as useless as they were hilarious - a true slop before the term was invented.

Microsoft seems to have thrown quality assurance overboard completely. Vibe generate everything, throw it at a wall, see what sticks. Tech bros are so afraid of regulation they even drop regulation inside their own companies. (just kidding)

It's not just throwing QA out, they are actively striving for lower quality because it saves money.

They're chasing that sweet cost reduction by making cheap steel without regard for what it'll be used for in the future.

Just a thought: the timeline of the vibe techs rolling out and the timeline of increasing product rot, sloppiness, and user-hostile “has anyone ever actually used this shit!?!” coming out of MS overlap.

Vibing won’t help out at all, and years from now we’re gonna have project math on why 10x-LLM-ing mediocre devs on a busted project that’s behind schedule isn’t the play (like how adding more devs to a late project generally makes it more late). But it takes years for those failures to aggregate and spread up the stack.

I believe the vibing is highlighting the missteps from the wave right before which has been cloud-first, cloud-integrated, cloud-upselling that cannibalized MS’s core products, multiplied by the massive MS layoff waves. MS used to have a lot of devs that made a lot of culture who are simply gone. The weakened offerings, breakdown of vision, and platform enshittification have been obvious for a while. And then ChatGPT came.

Stock price reflects how attractive stocks are for stock purchasers on the stock market, not how good something is. MS has been doing great things for their stock price.

LLMs make getting into emacs and Linux and OSS and OCaml easier than ever. SteamOS is maturing. Windows Subsytem for Linux is a mature bridge. It’s a bold time for MS to be betting on brand loyalty and product love, even if their shit worked.