Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Microsoft has a problem: lack of demand for its AI products

windowscentral.com

Post reply on HN
Print view View original

351–360 of 385 posts

Re: Microsoft has a problem: lack of demand for its AI products

#351

Post by pjmlp »

ThrowawayB7 wrote: post #265
sylens wrote: post #116

I think the biggest revelation of the last 3 years or so is that Microsoft does not have either the will or the talent (or both) to effectively execute anymore. Everything it currently stands on is a legacy product with roots in the Ballmer or Gates eras. They owe their Azure footprint and "success" today to Ballmer. Their inability to produce anything useful with Copilot is the largest example of this, but there are…

We've been hearing empty punditry like that for the last several decades all while they've been a poster child for "number go up": https://www.macrotrends.net/stocks/charts/MSFT/microsoft/rev... These predictions about the decline of Microsoft are like the Year of Linux on the Desktop; neither is going to happen anytime soon. Y'all can start predicting doom when there's a multi-year trend of declining revenue for MSF…

Just as Proton running Windows binaries is "Linux Gaming", WSL is the "Year of Linux on the Desktop".

Re: Microsoft has a problem: lack of demand for its AI products

#352

Post by vladvasiliu »

danudey wrote: post #247
vladvasiliu wrote: post #229

Earlier quoted context omitted.

On new-outlook-on-Windows there’s an option to disable the copilot button. Wanna take a guess what that does?

Honestly, I can't tell if you're being sarcastic and it actually does disable the button, or if you're being jaded and cynical and it just hides it for six hours then it comes back. Either option seems equally likely at this point.

The button stays in the bar, only greyed out. Its dropdown is still functional, allowing you to turn it back on.

Re: Microsoft has a problem: lack of demand for its AI products

#353

Post by raxxorraxor »

burningChrome wrote: post #170
throw310822 wrote: post #75

Earlier quoted context omitted.

The other day I've clicked on one of Outlook calendar's copilot prefilled questions: "who are the main attendees of this meeting". It started a long winding speech that went nowhere, so I typed in "but WHO are the attendees" and finally it admitted "I don't know, I can't see that".

I've fooled around with some vibe coding on several LLM's like Claude, Gemini, and ChatGPT with some pretty decent results. Since I have a full Copilot license at my corporate day gig, I figured I would try using Copilot for a basic static site. Nothing too hard, and something that's been handled easily with the other LLM's. The prompt was pretty basic just to get something to start working with. "Build a four page t…

I guess this is code people working at MS would commit to a useless file in order to simulate work done with AI support.

There are cheap 1b open coding models that do a far better job.

Re: Microsoft has a problem: lack of demand for its AI products

#354

Post by saubeidl »

cherrycherry98 wrote: post #336
saubeidl wrote: post #190

Earlier quoted context omitted.

As somebody doesn't consider himself a capitalist, wouldn't it be fair to say it is "the most efficient" in precisely one thing: capital reproducing itself? And if so, why is that necessarily a good thing? Why should that be our goal as society as opposed to things like minimizing child mortality, increasing literacy rates, making sure we don't have a ton of our fellow humans living on the street in misery etc etc -…

If you care about minimizing child mortality, increasing literacy, pulling people up out of poverty, you should be a capitalist, as it's empirically the best way to meet those goals. This seems to be a hard thing for many to understand or accept because it is largely a second order effect, the capitalist primarily concerned with their own personal gain but winds up improving the lives of others as a side effect. This…

Then why are all of these metrics worse in the US than even in notoriously poor communist Cuba?

Re: Microsoft has a problem: lack of demand for its AI products

#355

Post by KurSix »

Microsoft's problem isn't bad AI, but a fundamentally wrong UX pattern. They're trying to shoehorn Copilot into linear workflows - Word, Excel, Outlook - but chat is a terrible interface for editing documents or analyzing tables. Google with Gemini is winning not because their model is better (though it is good), but because they integrate AI better as a tool, not as a conversational partner. This difference in mental model defines utility

Re: Microsoft has a problem: lack of demand for its AI products

#356

Post by KurSix »

ZeroConcerns wrote: post #97
throw310822 wrote: post #75

Earlier quoted context omitted.

The other day I've clicked on one of Outlook calendar's copilot prefilled questions: "who are the main attendees of this meeting". It started a long winding speech that went nowhere, so I typed in "but WHO are the attendees" and finally it admitted "I don't know, I can't see that".

Absolutely! There are so many scenarios where they could actually add some value, and they're fulfilling, like, exactly none of those? Even in Visual Studio Enterprise, their flagship developer product, the GPT integration mostly just destroys code regardless of model output. I truly cannot fathom how any of that made it past even a cursory review. Or how that situation would last for over 6 months, but, yet, here we…

Compare this with Cursor or Supermaven. There AI acts as an extension of the cursor, understanding the code's AST. VS Copilot often feels like smart copy-paste that doesn't understand where it is Microsoft, owning VS Code and LSP, should have done this better than anyone, but they lost to startups on UX

Re: Microsoft has a problem: lack of demand for its AI products

#357

Post by KurSix »

Sanzig wrote: post #137
ZeroConcerns wrote: post #97

Earlier quoted context omitted.

Absolutely! There are so many scenarios where they could actually add some value, and they're fulfilling, like, exactly none of those? Even in Visual Studio Enterprise, their flagship developer product, the GPT integration mostly just destroys code regardless of model output. I truly cannot fathom how any of that made it past even a cursory review. Or how that situation would last for over 6 months, but, yet, here we…

I have Copilot at work, it feels so useless sometimes. As an example, I had a report which I needed to make some batch edits to. I figured why not let the robot take a crack at it, so I clicked the Copilot button and spent a couple minutes describing what I needed changed. Copilot tells me it can't edit my current document, but it can create a new one. I figured okay, Microsoft doesn't want to set it loose on the ori…

The chatbot doesn't see the document structure; it sees tokens. When it generates a response, it's not editing; it's rewriting from scratch. Without a mechanism capable of precise diff/patch operations (like Cursor attempts with code), an LLM will always be a destructive tool for complex documents. Microsoft marketed this as an editor, but in reality, it's a summary generator

Re: Microsoft has a problem: lack of demand for its AI products

#358

Post by KurSix »

artrockalter wrote: post #221
throw310822 wrote: post #75

Earlier quoted context omitted.

The other day I've clicked on one of Outlook calendar's copilot prefilled questions: "who are the main attendees of this meeting". It started a long winding speech that went nowhere, so I typed in "but WHO are the attendees" and finally it admitted "I don't know, I can't see that".

It's so easy to ship completely broken AI features because you can't really unit test them and unit tests have been the main standard for whether code is working for a long time now. The most successful AI companies (OpenAI, Anthropic, Cursor) are all dogfooding their products as far as I can tell, and I don't really see any other reliable way to make sure the AI feature you ship actually works.

Traditional Microsoft devs are used to deterministic tests: assert result == expected, whereas AI requires probabilistic evals and quality monitoring in prod. I think Microsoft simply lacks the LLM Ops culture right now to build a quality evaluation pipeline before release; they are testing everything on users

Re: Microsoft has a problem: lack of demand for its AI products

#359

Post by KurSix »

lll-o-lll wrote: post #239
kayhantolga wrote: post #96

As a .NET developer who actually likes some Microsoft products, I can say this: the Copilot series is the worst thing they've shipped since Internet Explorer—and honestly, it might overtake it. The sad part is they had a huge head start before competitors gained access to powerful models, yet this is what we got. If you haven’t seen how bad it is, here’s one example: Copilot Terminal. In theory, it should help you wi…

Works good in Teams. Summarises meetings, collects action items. It’s pretty great actually. I don’t know how many forests we are burning to have a digital secretary, but surely the environment can take one more for the team?

This is arguably the only use case where a chat interface (or just text) actually works, because it's an information compression task. But building a $50 billion business strategy just on the fact that AI can recap a meeting is risky. If this is the only feature that works well, then the business ROI is questionable

Re: Microsoft has a problem: lack of demand for its AI products

#360

Post by theshrike79 »

ZeroConcerns wrote: post #62

Well, the major problem Microsoft is facing is that its AI products are not only shoddier than average , which is nothing new for them in many categories, but that this time the competition can actually easily leapfrog them. Like, I have a 'Copilot' button prominently displayed in my New Outlook on MacOS (the only platform where the app-with-that-designation is sort-of usable), and it's a dropdown menu, and it has...…

There are SO MANY Copilots

There's the Copilot cli, then we have the Copilot built in to VSCode, then we have the Copilot that's in Github, then we have th M365 Copilot and the Windows Copilot and maybe a few more.

All work very differently and have different skillsets.

Post reply on HN