Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

They’re so desperate to make it work.

> “A source that has seen materials related to sales has confirmed that, as of August 2025, Microsoft has around eight million active licensed users of Microsoft 365 Copilot, amounting to a 1.81% conversion rate across the 440 million Microsoft 365 subscribers.”

Microsoft 365 Copilot's commercial failure - https://news.ycombinator.com/item?id=45476045 - October 2025



Our org has legal/security requirements around AI. MS is not making themselves available to address these concerns or giving our admins the information that is being asked for.

There are also ever changing, inscrutable licensing schemes, and too many similarly named products. It's impossible to know what to buy or how much it is going to cost.

This is even before you get to the quality of the product itself. I mean, part of the reason people aren't using it because it is hard to buy and use.


>inscrutable licensing schemes, and too many similarly named products.

This has always been Microsoft's MO. They used to have a team you could call to figure out which product you actually needed.


My company has both enterprise ChatGPT and Copilot. We originally had CoPilot; we learned it can search for files on OneDrive and Emails and that is the extent of its usefulness. Because everyone was doing work on their personal ChatGPT, the company had to buy the enterprise version of ChatGPT due to risk/compliance issues. Although CoPilot supposedly runs using the same engine, it feels severely limited compared to ChatGPT.


I seem to remember a similar arc with Cortana. Didn't it briefly appear everywhere in MS365?


At this point its a sunk cost fallacy


Yeah Im just waiting for this to finally play out.

If you understand product design, economics, corporate finance and various other disciplines well - its obvious what is going on.

MSFT wants to juice their numbers for the next earnings call to keep the mania going. Zzzzzzzzzzz

Ive spoken to people who work in the finance sector - portfolio management and tax audit - they laugh hysterically at how bad the tools are (copilot in particular) and resent how much they are being pushed down on them.


I think it's more nuanced than that. Microsoft is desperate for growth in a growth challenged macro, and Nadella has guided valuations and stock targets that are simply unobtainable without this growth. That's one part of the AI slop shovel. The second part is institutions that are desperate to cut labor costs or find other efficiency gains to maintain historical financial performance targets during ZIRP that are simply no longer obtainable. They are willing to push their workers through the sausage machine of "AI" to try to make it happen.

And so the performance art must continue, at least until the AI investment music stops. Your options are get rich if in a position to (due to irrational exuberance and unsophisticated capital investment), or play along until the facade falls to keep your job. "It is what it is."


> Nadella has guided valuations and stock targets that are simply unobtainable without this growth

I read a couple of articles and he seems to have been completely suckered in by the damn thing

from https://archive.is/oWbZB#selection-2387.572-2401.668

> Copilot consumes Nadella’s life outside the office as well. He likes podcasts, but instead of listening to them, he loads transcripts into the Copilot app on his iPhone so he can chat with the voice assistant about the content of an episode in the car on his commute to Redmond. At the office, he relies on Copilot to deliver summaries of messages he receives in Outlook and Teams and toggles among at least 10 custom agents from Copilot Studio. He views them as his AI chiefs of staff, delegating meeting prep, research and other tasks to the bots. “I’m an email typist,” Nadella jokes of his job, noting that Copilot is thankfully very good at triaging his messages.

this is no different to ending up in an alternate reality populated by your AI "girlfriends"

and will produce similar results


Sounds like a dogfooding sales pitch to me. Nadella has learned from embarrassing prior incidents where Windows Phone developers were found to be using the iPhones or Androids as their personal devices, or when pictures of Microsoft employee offices showed Macs littering the workspace.


But if real, indicates, that he does take it serious. Eat your own dogfood and all that.

I do not see it as AI girlfriend. I see it as pushing the limits of the technology for productivity.

" Copilot is thankfully very good at triaging his messages"

But .. I would never do that, without also having a human check what important bits don't get through.


At some point a mistake will be made that will be very embarrassing for someone that matters


People have "satired" similar hypotheticals about Sam Altman [0], and we know it's happened with people like Blake lemoine.

It makes me curious how much, if any, of the current LLM hype is simply coming from people who are dealing with the ELIZA effect.

[0] https://medium.com/where-thought-bends/the-7-trillion-delusi...


How'd ELIZA do on the International Math Olympiad?


It's quite fortunate that the ELIZA effect isn't defined by IMO scores, isn't it?


How is that point relevant? The ELIZA effect involves humans incorrectly attributing a simple computer program with intelligence. Meanwhile, a complex ML model that solves difficult problems it hasn't seen before is intelligent, whether you or I like it or not.

That part is no longer up for debate. The question is how useful and scalable this particular form of intelligence will turn out to be.


Yes but the sunk cost is also driving that - the investment has to show growth in cashflows, to the extent it is beating the hurdle rate to add value.


Ahh, Microsoft's AI capital investment sunk cost fallacy, certainly, I agree with that. The stock will be punished when the investment does not show cashflow returns for sure.

MSFT: Microsoft CEO Offloads $75 Million in Stock Amid AI Boom - https://finance.yahoo.com/news/msft-microsoft-ceo-offloads-7... - September 5th, 2025


> institutions that are desperate to cut labor costs...

I wonder if those who resist copilot aren't driven by the nascent feeling they are training their replacement.

> ... to maintain historical financial performance targets during ZIRP that are simply no longer obtainable.

If all else fails, they can ZIRP back to fantasy land and avoid being the losers by shifting it to the public once again.


Microsoft really wants that AI company valuation.


8 million sold licenses doesn't sound like a failure to me. Why do people assume that the market for Office programs and LLM tools is the same or even comparable?


> Why do people assume that the market for Office programs and LLM tools is the same or even comparable?

People ? No, Microsoft




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: