Microsoft CEO: Time To Move 'Beyond the Arguments of Slop vs Sophistication' (snscratchpad.com) 98
The tech industry needs to move "beyond the arguments of slop vs sophistication" and develop a new "theory of the mind" that accounts for humans now equipped with "cognitive amplifier tools," Microsoft CEO Satya Nadella wrote in a year-end reflection blog. The post frames 2026 as yet another "pivotal year for AI" -- but one that "feels different in a few notable ways." Nadella claims the industry has moved past the initial discovery phase and is now "beginning to distinguish between 'spectacle' and 'substance.'" He argues for evolving beyond Steve Jobs' famous "bicycles for the mind" framing, positioning AI instead as "scaffolding" for human potential rather than a substitute.
"We will evolve from models to systems when it comes to deploying AI for real world impact," Nadella writes, adding that these systems must consider their societal impact on people and the planet. "For AI to have societal permission it must have real world eval impact."
"We will evolve from models to systems when it comes to deploying AI for real world impact," Nadella writes, adding that these systems must consider their societal impact on people and the planet. "For AI to have societal permission it must have real world eval impact."
wow (Score:1, Informative)
What the hell does that even mean. LLMs our a joke and all the companies making them no that. All that is left is for the music to stop. Like Open AI and there what ever guy. They know its over. All of them are saying it. Curor AI actaully said it software make unsable shit.+
Re: (Score:1, Funny)
You're right. Although LLMs can write and spell correctly and you can't, you're intelligent and they are not.
Suggesting a better headline (Score:4, Insightful)
/. headline - "Microsoft CEO: Time To Move 'Beyond the Arguments of Slop vs Sophistication'"
Suggesting a better headline - "Microsoft CEO tired of justifying AI and glossing over the negatives of AI"
Re:wow (Score:5, Informative)
The proper way to use this piece is if you are an employee at Microsoft, now you know what you need to do to get a promotion. Find a way to measure the positive impact of AI, and the Nadela will be happy. No need to worry about what's real, only about what Nadela will like. It's actually clear and informative for people who read it like that.
Meanwhile, back in the real world, it's nonsense. It's just noise in the blogosphere, because the CEO of Microsoft is no more important than any other boke with a bog.
Re: wow (Score:4, Informative)
The proper way is to leave Microsoft because all the article contains is corporate marketing dravel. It might convince clueless investors that only sees a chance to make money by selling vaporware.
Most AI won't help you much and it can get summaries wrong so I don't trust AI. When you can get an AI that understands satire and irony you might get somewhere, but it would be mission impossible.
Re: (Score:2)
Well, I assume there are few people working at Microsoft with personal integrity or honor, so I expect there will be quite a few promotions now.
Re: (Score:2)
If two managers do well and deserve a promotion, one of them has to stab the other in the back.
The Reflection Engine (Score:2)
so much negativity, my experience is the opposite but then again I'm not a coder who obsesses over saving 0.2 ms or inventing my own "for" loop to prove how brilliant I am. :)
https://www.scry.llc/2025/06/0... [scry.llc]
"Many people use ChatGPT (AI) thinking they want answers â" but sometimes they get a conversational mirror which amplifies intent and direction. Like a thought forge: ideas enter half-formed and exit sharper, cooler, or occasionally twisted into something new.
Or perhaps today it's a semantic mirror
Re: (Score:2)
I have a code base that is over a million + lines of code and does 250k medical tranactions a day(EDI). It does another 500k HL7 messages and millions of db crud operations. Cursor AI gets confused at around 800 lines of code. It can not even build a production app server from start to finish. Lama 4 in face book actauly does a good job but really the same limitations.
The big problem with LLM is they will never be able to exceed what is in there training data. They do not learn they ar statistical model of
Re: (Score:3)
Indeed. And LLMs cannot scale up, because that requires actual insight and model-building. These things are restricted to smart humans. Admittedly, smart humans are a minority and most humans would not ever reach these 800 lines. But in actual software projects, 800 lines is tiny. Maybe you can do, say, the software for a heater or a simple electronic scale in that.
Re: (Score:2)
> which amplifies intent and direction
Which amplifies what the idiot machine has mis-identified as your intent and direction. Or the weights pre-prompted into the AI.
> Like a thought forge: ideas enter half-formed and
exit as melted sludge, or occasionally twisted. Like all those people pushed into psychosis for either murder or suicide.
> so much negativity ... inventing my own "for" loop to prove how brilliant I am
You really are an idiot. Nobody is going to believe you're brilliant.
Re: (Score:2)
Guido van Rossum said, without realizing that he was saying it... that nobody should have to make code that isn't aesthetically pleasing and that can't be made to look like a series of poetic stanzas.
I agree with him.
That's why I don't use Python.
Re: (Score:2)
"... Or perhaps today it's a semantic mirror which reflects back insights you didn't know you had"
Hahahaha, no. That one is called "delusion". To be fair, sycophantic tactics are often aiming at creating exactly this impression. This approach works well on weaker minds and the commercial LLMs are all using it.
Re: (Score:2)
"conversational mirror" this is Adam Becker's argument. I presume it is contained in his book: More Everything Forever: AI Overlords, Space Empires, and Silicon Valley's Crusade to Control the Fate of Humanity.
I've not read the book but saw him on CSPAN. One comment stuck with me. The AI Big-Wankers (my words, not his) think they are the smart people on the planet because they are rich and everyone else isn't. They suffer from Monkey Butt Syndrome (my paraphrase of a phrase I heard long ago): If you are on
Re: The Reflection Engine (Score:2)
Another argument of Becker's is that the AI Big Wankers want to take every aspect of your life, digitize it, monetize it, and sell it back to you at inflated prices
Who are the buyers if most people lose their jobs?
Re: (Score:2)
No idea. Fairly sure he had copilot write it.
Everyone that isn't a member of the ruling class (Score:5, Informative)
It's not a disruptive technology like the Internet it's a destructive technology. There are no benefits to it for anyone except the billionaires that are going to own it. It does free them from the chains of capitalism. With it there are no longer dependent on consumers or employees. They can truly have their cake and eat it too.
The problem is there is no place for us in that world. We don't even get to be peasants.
If you're reading this you are not one of the engineers that is going to be allowed to have food and shelter. You are one of the serfs. If you are one of the engineers you would be too busy to be wasting your time on this website
And everybody should rightly be despised (Score:1, Interesting)
Because very very very soon we are all except the ruling class lose access to electricity no kidding mark my words. It is absolutely crazy that we are all very very soon going to lose access to electricity because it's all going to go to grift and robots replacing all our jobs. Billionaires got private armies to protect them and they're working on robots & automation to render the consumers they depend on for their wealth moot. Do a bit of googling and you will find a study showing that 70% of the middl
Re: (Score:2, Informative)
He isn't using AI at all. He is simply reposting verbatim rsilvergun comments without the links. Everything was really said by rsilvergun.
Of course, rsilvergun is such a nutbar that it sounds just like AI crap although...
Proof:
without links:
https://slashdot.org/comments.... [slashdot.org]
with links:
https://slashdot.org/comments.... [slashdot.org]
Re: (Score:1, Informative)
rsilvergun is craving so much for attention, I wouldn't be surprised at all if he was the one behind it all...
I am a nut bar (Score:2)
It's being run through a primitive llm probably running on a graphics card on the computer of the numb nut who has nothing better to do besides stalk me.
The only thing sadder than being me is copying me.
Re: (Score:2)
There are no benefits to it for anyone except the billionaires that are going to own it.
This is actually incredibly wrong and naive. The LLM technology DOES have utility and I personally get incredible use out of them as coding assistants in particular. The problem with the technology is that the people financially invested in it have become high on their own supply of hype. They really do want it to be a do everything job killer and they are leveraging trillions to do so. Unfortunately for all of us, this will never work because the LLM technology is not capable of it and never will be.
Bu
Re: (Score:3, Insightful)
The ridiculous amounts of water and electricity (Score:2)
The only thing that makes it worthwhile is that billionaires are sick and tired of having to be dependent on consumers and employees. So they will spend any amount of electricity and water to break that dependency.
Re: (Score:2)
Well first off, LLMs don't have to be run in multiple mega server farms. Secondly, as I said, they are very effective when purpose-built. Thirdly, you should focus your ire on the billionaires and their corporations abuses. There is no way in hell LLMs will ever lead to any sort of replacement for humans. But these assholes won't be satisfied until they collapse the economy trying.
Re: (Score:2)
The LLM technology DOES have utility and I personally get incredible use out of them as coding assistants in particular.
That claim alone already nicely shows you are incompetent. Because nobody gets that level of use out of coding assistants. Sure, there are some people that struggle at "Hello, world" and these may mistakenly think what they get is "incredible". But these people do not count. The other option is that you are in delusion or lying.
Nice reference: https://mikelovesrobots.substa... [substack.com]
Re: (Score:2)
FTFY. I'm sorry that you're too stupid to figure out how to make this technology work for you.
Re: (Score:1)
And of course my html formatting was wrong....let's try again.
Because nobody gets that level of use out of coding assistants.
I'm sorry that you're too stupid to figure out how to make this technology work for you.
Re: (Score:2)
You wish. That is deep delusion.
Re: (Score:2)
No what's delusional is trying to tell a stranger on the internet that their actual experiences are delusional. Like, you see that right? You're at conspiracy level shit here because nothing I say will convince you otherwise. You'll listen to some rando's blog post about his "study" that assumes the grandiose dreams about the actual productivity gains from using Ai. But you will ignore the guy simply saying that "yes I find them useful".
My argument: If so many developers are so extraordinarily productive using these tools, where is the flood of shovelware?
This quote right here proves to me that this guy is just blogging f
Re: Everyone that isn't a member of the ruling cla (Score:2)
The problem I have is they are being shoved stem our throats whether we want them or not. In many cases they work out to an investor jerk-off feature than real value.
The marketing item âoeuses AIâ comes off as more of a bio hazard symbol to me than anything useful. I interpret it as a lot of hand waving instead of real value. If AI is solving something useful then sell me that point, otherwise itâ(TM)s going to signal to me something that is cloud dependent and potentially near future e-waste
Re:Everyone that isn't a member of the ruling clas (Score:5, Insightful)
Nadela wrote:
We need to get beyond the arguments of slop vs sophistication and develop a new equilibrium in terms of our “theory of the mind” that accounts for humans being equipped with these new cognitive amplifier tools as we relate to each other. This is the product design question we need to debate and answer.
Interpretation:
"I'm tired of hearing about slop, stop talking about it. Now build me something I can sell, or I'll throw chairs."
Re: (Score:2)
Interpretation:
"I'm tired of hearing about slop, stop talking about it. Now build me something I can sell, or I'll throw chairs."
Exactly. Also "I know I wasted mountains of money on this stupidly, get me something that helps hiding how incompetent I am!"
Re: (Score:3)
I must be a member of the ruling class then, because I *love* AI and what it does for me! And like all new automation technologies, I believe AI will have many positive and negative effects, but overall will benefit mankind.
In my 35-year career, most of my work has been automating business processes. And yet, companies can't find enough people to manage their business processes. We've automated away 95% of all farm labor, and yet farmers can't find enough workers. We've automated away most factory work, and
Re: (Score:2)
"Ruling class" is a buzzword on the left, just as "woke" is a buzzword on the right.
And just as "woke" isn't a real thing, the "ruling class" isn't a real thing. Both are straw men representing whatever is despised by the left or the right, respectively.
Slop merchant promotes slop (Score:5, Insightful)
Re: (Score:3, Interesting)
Also, gambler deeply invested in a scam tries to claim the scam is not a scam. I do not think MS will make any profit from LLM-type AI though. It is just far too bad. The only ones that will make moderate profits are makers of small, customized, restricted-purpose LLMs and these will be small companies and maybe some universities.
Re: (Score:2)
Further, those of us in the research side of things will have to spend more time sifting the wheat from the chaff when it comes to software development directions that are incidental part of research projects.
Luckily, the lack of a measurable outcome is its own punishment.
New Theory! (Score:2)
Let me be the first to create a new Theory of Human Minds Enhanced by Slop Insertion Tools(TM).
It goes like this:
Law 1: Human (average intelligence) + Slop Insertion Tool = Enhanced Human (below average intelligence).
Law 2: Enhanced Human + Bridge = Money Machine.
Re: (Score:3)
Law 2: Enhanced Human + Bridge = Money Machine.
How much are you interested in paying for this "Bridge"? Because have I got a deal for you...
Re: (Score:2)
Re: (Score:2)
Exactly. I mean, even average intelligence people can see how bad the LLM-type Artificial Idiots are. Need to bring the standards down to actually sell this crap.
Must be wrong (Score:3)
I had sort of felt that way, but if Microsoft is pushing that line it must be wrong. They've got a history.
Re: (Score:2)
Indeed. I remember the ads about Linux being a "cancer" and these days MS has its own Linux distro and Azure is apparently running more Linux loads than Windows ones. And Win11 is looking more and more outdated in comparison. Since I generally do not pay any attention to MS statements, this is the only one I remember as it made the general news here.
How about (Score:2)
AI shills (Score:4, Insightful)
Truely a legend in his own mind. We are already plagued by too much artificial intelligence. What we need is the real thing, not something that will creatively cow back the mistakes of others, while adding a bit of its own fancy and marketing it as truth...
Re: (Score:2)
We are already plagued by too much artificial intelligence. What we need is the real thing, not something that will creatively cow back the mistakes of others, while adding a bit of its own fancy and marketing it as truth...
Indeed. But there does not seem to be a way to get more of it, education is not doing it. Now we could make sure all decision makers are actually high intelligence people with intact personal morals and no maliciousness. And we would probably just about have enough of these people. But since that would impact profits and some zealot agendas, it is not going to happen. Instead we put dumb and dumber, with no morals or integrity and high level of maliciousness in power, with predictable results.
Translation (Score:1)
Re: (Score:2)
"After setting fire to immense amounts of money on fire
You gotta be super thorough when you're trying to burn money, I guess.
"The Narrative is not Going my Way- (Score:3)
accept the slop (Score:4, Insightful)
"Accept the slop and stop talking about it. Its the same thing you get from the Intern, just like the Intern our systems are learning and this year will be just great! So stop shaming us for being a joke."
Nah, I'll call it how I see it.
Re: (Score:2)
Indeed. The "Great New Thing" is simply another step towards full enshittification.
Agreed (Score:4, Insightful)
It's time to move on to the realization that it's all slop
Re: (Score:2)
You know, except for the occasional instance of "better search" (which is really not enough to even remotely justify the effort), it actually seems to be.
Re: (Score:2)
Sometimes slop is a better search.
Re: (Score:2)
Sometimes slop is a better search.
Yes, when the search has been deliberately compromised to promote advertising.
Crappy lies ... (Score:2)
... from the head of a thoroughly crappy company. Yes, it is slop. Yes, your investments in it are sunk cost you will not recover. Yes, the whole thing will crash in the (somewhat?) near future. Yes, the same thing happened to all AI hypes so far and there have been a couple.
Stop trying to confuse the issue. There was nothing "pivotal" about AI in 2025 and the improvements were marginal or not even there.
Loudness = Power!!! (Score:1)
https://youtu.be/IiZeOgxpCmI [youtu.be]
- Evil Kneval
Typo? (Score:2)
For AI to have societal permission it must have real world [evil] impact.
Yeah, right (Score:3)
Sounds like exactly the kind of thing I'd exepct Microslop to say as they force "AI" into every aspect of Windows.
CEO slop argument (Score:4, Insightful)
Most people don't want AI (Score:3)
Yet MSFT et al. seem compelled to push it on us. Well, The Generative AI Industry is Fraudulent, Immoral and Dangerous [skoll.ca]
Re: (Score:2)
I found myself hoping (Score:2)
that the summary had in some way mis-identified or mis-represented the post. So, to check the text verbatim:
> What matters is not the power of any given model, but how people choose to apply it to achieve their goals. We need to get beyond the arguments of slop vs sophistication and develop a new equilibrium in terms of our “theory of the mind” that accounts for humans being equipped with these new cognitive amplifier tools
godfuckingdamnit Nadella you piece of trash:
- anybody building a tool
Re: (Score:2)
To rephrase him;
What matters is not the power of any given gun, but how people choose to apply it to achieve their goals. We need to get beyond the arguments of bad vs good and develop a new equilibrium in terms of our "theory of life" that accounts for humans being equipped with these new power amplifier tools.
big assumption (Score:4, Insightful)
The software you wrote, that you don't fully understand well enough to debug when it misbehaves. Why do you assume it amplifies cognition? Seems like it amplifies stupidity, or more generously it amplifies errors. Common errors found in forum posts on the Internet over the last 50 years. Maybe don't train an LLM on people posting confidently the most unhinged or ridiculous things. All we did was teach AI how to speak with certainty on topics that it has the barest knowledge of.
Translation (Score:2)
In other words, "MS has invested so much into the AI boondoggle that everyone had better jump on board before I lose my bonuses and stock options."
Microsoft CEO: says our AI does not work! (Score:3)
Crazy Train (Score:1)
Get off the crazy train while you can folks. Linux and iOS are only a switch away. Hell, at this point I wouldn't snicker if you moved to a Chromebook. Good for you for escaping.
Cognitive Amplifier Tools? (Score:2)
So Microsoft's next attempt to replace Rover will be a C.A.T.?
a new theory of the mind? (Score:2)
'...develop a new "theory of the mind" that accounts for humans now equipped with "cognitive amplifier tools,"...'
Why would a theory of the mind include anything external to the body? How can this CEO be so stupid?
Yes, shut the f up and keep ... (Score:2)
... walking, Do not talk. That's what they really mean. No criticism. No reflection.
sigh just more slop (Score:2)
Microsoft theory of the mind :o (Score:2)
Misspelled (Score:2)
Satya,
You misspelled "evil".
T
Do you guys not have 365? (Score:2)
He says you have to use Office like that and he lucked into the cloud transition making Billions, how dare you users disagree? Are you Billionaires? Consume the sophistication from Nadella's trough.
Summed up (Score:2)
Slop
Translation (Score:2)
"Just accept mediocrity and don't resist, it'll be less painful, because we sure as hell won't stop trying to make money off of AI."
Fuck off Satya... (Score:2)
I'm getting so tired of the pseudo-profound bullshit churned out by whatever this guy imagines is his intellect. His verbal meanderings remind me of New Age 'insights'. I bet he and Peter Thiel could have some amazing-to-themselves conversations. I wish they'd have those conversations with each other - in private - and STFU when it comes to public blathering.
AI Spooge (Score:2)
Spooge is better than Slop because it represents tech bros cumming in their pants at the thought of how much money AI will make them while everyone else suffers.
Re: (Score:2)
"Sunken cost fallacy" (Score:2)
A year ago I found that my pension fund used my money to invest in Microslop. That made me so angry, because it was obvious to pretty much everyone in the industry already then that the bubble is going to burst -- and then Microslop is going to be hit hard.
That Nadella is still under the delusion that his venture will succeed is mind-boggling.
Corporate slop, lets coin it now! (Score:2)
For reference, I neither like or dislike AI tools for making content. At the end of the day, it just comes down to how lazy you are. The recent Coke/McDonalds ad's were just lazy. The tools are fine and they HAVE to have conventional video editing tools so they "could" of fixed many of the issues in post. If you ever played with LLM workflow you can control how content is generated and then generate a shit ton of sets to get what you want. Even if you get the odd 6 fingered Santa you COULD go in and to