67 comments

[ 2.5 ms ] story [ 82.4 ms ] thread
AI agent technology likely isn’t ready for the kind of high-stakes autonomous business work Microsoft is promising.

It's unbelievable to me that tech leaders lack the insight to recognize this.

So how to explain the current AI mania being widely promoted?

I think the best fit explanation is simple con artistry. They know the product is fundamentally flawed and won't perform as being promised. But the money to be made selling the fantasy is simply too good to ignore.

In other words --- pure greed. Over the longer term, this is a weakness, not a strength.

People think that because AI cannot replace a senior dev, it's a worthless con.

Meanwhile, pretty much every single person in my life is using LLMs almost daily.

Guys, these things are not going away, and people will pay more money to use them in future.

Even my mom asks ChatGPT to make a baking applet with a picture she uploads of the recipe, that creates a simple checklist for adding ingredients (she forgets ingredients pretty often). She loves it.

This is where LLMs shine for regular people. She doesn't need it to create a 500k LOC turn-key baking tracking SaaS AWS back-end 5 million recipes on tap kitchen assistant app.

She just needs a bespoke one off check list.

I mean, see Windows Vista. It was eventually patched up to the point where it was semi-usable (and then quietly killed off), but on introduction it was a complete mess. But... something had to be shipped, and this was something, so it was shipped.

(Vista wasn't the only one; Windows ME never even made it to semi-usable, and no-one even remembers that Windows 8 _existed_.)

Microsoft has _never_, as far as I know, been a company to be particularly concerned about product quality. The copilot stuff may be unusually bad, but it's not that aberrant for MS.

>So how to explain the current AI mania being widely promoted?

CEOs have been sold on the ludicrous idea that "AI" will replace 60-80% of their total employee headcount over the next 2-3 years. This is also priced into current equity valuations.

At this point, the people in charge have signed off on so much AI spending that they need it to succeed, otherwise they are the ones responsible for massive losses.
(comment deleted)
If you click through to the article shared yesterday[0]:

> Microsoft denies report of lowering targets for AI software sales growth

This Ars Technica article cites the same reporting as that Reuters piece but doesn't (yet) include anything about MSFT's rebuttal.

[0]: https://news.ycombinator.com/item?id=46135388

Have we finally reached peak AI already? In that event we will see the falling down phase next.
Yea, we're getting their, had some people reach out to me who only do so once a hype bubble is well formed
Not until we put it on the blockchain
Top signal. Phase transition is imminent.
Is "The Information" credible? It's the sole source.
But is it sold enough to regular Windows Home users? If MS brings an ultimatum: "you need to buy AI services to use Windows", they might get a bunch more clueless subscribers. In the same way as there's no ability to set up Windows without internet connection and MS account they could make it mandatory to subscribe to Copilot.
Despite having an unlimited warchest I'm not expecting Microsoft to come out as a winner from this AI race whilst having the necessary resources. The easy investment was to throw billions at OpenAI to gain access to their tech, but that puts them in a weird position of not investing heavily in cultivating their own AI talent and being in control of their own destiny by having their own horse in the race with their own SOTA models.

Apple's having a similar issue, unlimited wealth that's outsourcing to external SOTA model providers.

As someone who appreciates machine learning, the main dissonance I have with interacting with Microsoft's implementation of AI feels like "don't worry, we will do the thinking for you".

This appears everywhere, with every tool trying to autocomplete every sentence and action, creating a very clunky ecosystem where I am constantly pressing 'escape' and 'backspace' to undo some action that is trying to rewrite what I am doing to something I don't want or didn't intend.

It is wasting time and none of the things I want are optimized, their tools feel like they are helping people write "good morning team, today we are going to do a Business, but first we must discuss the dinner reservations" emails.

Dear MS please use AI to autocomplete my billing address correctly when I fill out web forms, thanks
(comment deleted)
Perhaps this is a feature and not a bug for MS. Every time you hit escape or accept, you're giving them more training samples. The more training data they can get you to give them, the better. So they WANT to be throwing out possibly irrelevant suggestions at every opportunity.
As an old school interface/interaction designer, I see this as a direct consequence of how the discipline of software design has evolved in the last decade or two.

We’ve went from conceiving of software as tools - constructs that enhance and amplify their user’s skills and capabilities - to magic boxes that should aim to do everything with just one button (and maybe even that is one action too many).

This shift in thinking is visible in how junior designers and product managers are trained and incentivized to think about their work. “Anticipating the user’s intent”, “providing a magical experience”, “making the simplest, most beautiful and intuitive product” - all things that are so routine parlance now that they sound trite, but that would make any software designer from the 80s/90s catatonic because of how orthogonal they are to good tool design.

To caricature a bit, the industry went from being run by people designing heavy machinery to people designing Disneyland rides. Disneyland rides are great and have their place, but you probably don’t want your tractor to be designed like one.

As much as I love JetBrains (IntelliJ and friends), I have the same feeling this year. The ratio that I undo an accidental tab/whatever far exceeds the accepted ones. I'm not anti-LLM -- they are great for many things, but I am tired of undoing shitting suggestions. Literally, many of them produce a syntax error. Please don't read this post as dumping on JetBrains. I still love their products.

    > It is wasting time and none of the things I want are optimized, their tools feel like they are helping people write "good morning team, today we are going to do a Business, but first we must discuss the dinner reservations" emails.
No trolling: This is genius-level sarcasm. You do realise that most "business" emails are essentially this, right? Oh, right, you knew that already!
AI is people looking at EV hype and saying - I'll 100x it.

It has all the same components, just on much higher scale:

1. Billionaire con-man convincing large part of market and industry (Altman in AI vs Musk in EV) that new tech will take over in few years.

2. Insane valuations not supported by an actual ROI.

3. Very interesting and amazing underlying technology.

4. Governments jumping on the hype and enabling it.

What can you even do in the ms enterprise ecosystem with their copilot integration?

Is it just for chatting? Is it a glorified RAG?

Can you tell copilot co to create a presentation? Make a visualisation in a spreadsheet?

It's basically clippy without the funny animations.
Gamma.app is what people use for that now. And it definitely doesn't feel like PowerPoint with AI slapped on.
The difference between poison and medicine is the amount. AI is great and very useful, but they want the AI to replace you instead of supporting your needs.

"AI everywhere" is worse than "AI nowhere". What we need is "AI somewhere".

I fully agree.

It's great for prototypes reactors, sometimes debugging (I'm kinda annoyed at how googling seems less and less efficient nowadays), and also reviewing.

But you gotta be critical about it. It's like a junior assistant eager to show work - you shouldnt take it and deliver it as is, it needs to be filtered by a competent person with taste!

Even Devblogs and anything related to Java,.NET, C++ and Python out of Redmond seems to be all around AI and anything else are now low priority tickets on their roadmaps.

No wonder there is this exhaustion.

I went to Ignite a few weeks ago, and the theme of the event and most talks was "look at how we're leveraging AI in this product to add value".

Separately, the theme from talking to Every. Single. Person on the buy-side was gigantic eye roll yes I cant wait for AI to solve all my problems.

Companies I support are being directed from their presidents to use ai, literally a solution in search of a problem.

Super interesting how this arc has played out for Microsoft. They went from having this massive advantage in being an early OpenAI partner with early access to their models to largely losing the consumer AI space: Copilot is almost never mentioned in the same breath as Claude and ChatGPT. Though I guess their huge stake in OpenAI will still pay out massively from a valuation perspective.
They made Copilot the term for AI and smeared it everywhere to the point that it has no meaning and therefore no usage when talking about AI.
Hearing similar stories play out elsewhere too with targets being missed left and right.

There’s definitely something there with AI but a giant chasm between reality and the sales expectations on what’s needed to make the current financial engineering on AI make any sense.

I wonder if it’s because Microsoft is hyper focused on a bunch of crap people don’t want or need?
Anyone who has had the pleasure of being forced to migrate to their new Fabric product can tell you why sales are low. It's terrible not just because it's a rushed buggy pile of garbage they want people to Alpha test on users but because of the "AI First" design they are forcing into it. They hide so much of what's happening in the background it is hard to feel like you can trust any of it. Like agentic "thinking" models with zero way to look into what it did to get to the conclusion.
Good. Go make your OS useful and stop alienating your enterprise customers.
Hopefully this is the beginning of the trough of disillusionment, and the steady return of rationalism.
People are wondering how we got here when these AI's make so many mistakes.

But the one thing they're really good at is marketing.

That's why it's all over linkedin etc, marketing people see how great it is and think it must be great at everything else too.

I think MSFT really needs some validated user stories. How many users want to, "Improve my writing," "Create an image," "Understand what is changed" (e.g. recent edits), or "Visualize my data."?

Those are the four use cases featured by the Microsoft 365 Copilot App (https://m365.cloud.microsoft/).

Conversely, I bet there are a lot of people who want AI to improve things they are already doing repeatedly. For example, I click the same button in Epic every day because Epic can't remove a tab. Maybe Copilot could learn that I do this and just...do it for me? Like, Copilot could watch my daily habits and offer automation for recurring things.