Satya Nadella has finally had enough. After all the AI-driven hype, he is calling out the emperor’s new clothes, and is pulling back the curtain which reveals that – surprise, surprise – AI hasn’t delivered diddly squat on its promises.
The guy appeard on podcast of a guy called Dwarkesh Patel, and he was friggin’ brutally honest, and that is rare for a Big Tech CEO. Nadella confessed that we have all been duped by the AI hype and the race to Artificial General Intelligence.
Forget the fluff.
Forget the inflated milestones.
Nadella’s not buying the fairy tale anymore.
You think AGI is just around the corner? You think this is the future? The man himself, who has staked billions of dollars on this so-called AI revolution through Microsoft’s investment in OpenAI, and of course it’s own huge investments in tooling, cloud-services, and their own models, is here to tell you something that no one wants to hear. . .
AI isn’t doing jack shit.
That’s right.
The future isn’t a world where robots will take over, but one where we’re all just sitting around, waiting for productivity to trickle in, slowly.
Yeeeeeey!
Finally reprieve from the Doomsday Apocalypse !!
“Artificial General Intelligence? Please. That’s just nonsense,” said the man.
Ouch.
That’s a huge slap to the face of anyone who thought that we were going to get an all-knowing machine anytime soon. But he challenges us to look at the real business metrics instead, and that is productivity growth of companies working with AI.
Is AI making the economy grow?
Are we seeing the rapid advancements we were promised?
Apparently not.
Sure, the guys at Klarna were the first to adopt AI and reduced their customer service staff to nearly zilch by replacing them with ChatGPT powered bots. But their customers do not have the luxury of being picky. So this is an easy kill.
Now Microsoft and its buddies at OpenAI keep pushing out tools that make everyone all hot and horny, but Nadella is quick to point out that the proof isn’t in your new ChatGPT or in the next “breakthrough” in AI.
It-is-in-the-numbers.
If AI is really that big of a deal, they would have seen growth at their customers, real growth. But now, we are watching the slow churn of a tech that still requires a human hand to slap the thing into motion.
More rants after the commercial brake:
- Comment, or share the article; that will really help spread the word 🙌
- Connect with me on Linkedin 🙏
- Subscribe to TechTonic Shifts to get your daily dose of tech 📰
- Visit TechTonic Shifts blog, full of slop, I know you will like !
Forget AGI. How about some real economic growth, Sam?
Nadella’s biggest point was that the industry has been so obsessed with the idea of AGI, you know, the utopian dream where machines surpass human intelligence, that we have lost sight of the fact that technology should be doing one thing and one thing only, which is making money.
You know. . .
Dough.
Moolah.
Bucks.
Bread.
Greenbacks.
Everybody went quiet when he said, “The real benchmark is, seeing the world growing at 10 percent”.
Thank you Satya!
This isn’t about chasing after the mythical AGI. This is about the basics. Can this tech actually drive economic growth? Can it do what it says it will do, or are we all stuck in a multi-billion dollar hamster-wheel.
I think you know the answe.
It’s a simple equation, really.
If AI was the game-changer it was promised to be, we would be looking at an uptick in productivity. And, oh yeah, that whole “economy growing faster” thing would be nice too. But if you’re waiting for that magic moment, you might want to sit back and relax for a bit longer because nothing even close to that has materialized yet.
Nadella isn’t the only one calling out AI for its lack of progress.
The top AI systems on the market, including OpenAI’s flagship models, are still crawling at a snails pace. And they still need constant oversight. We are paying a premium for a system that is supposed to be revolutionary, but now we all have realized that it’s more like an ambitious parrot that can barely string a sentence together without crashing or lying.
But that’s the reality right now.
The AI revolution we were promised has yet to show up. And Nadella’s not holding his breath either. He may have sunk billions into this venture, but even he knows that a fancy neural network won’t magically cure the tech world’s problems.
The industry is clingint to Agentic AI
So here we all are. We are standing on the edge of the generative AI shoulder, watching the bubble pop, and sliding down into the trough of disillusionment*. Yeah, bubble popping, the hiss of air escaping ever so slightly, and we’re all staring at the overblown hype that’s about to deflate faster than a tech CEO’s ego after he realized that they just spent billions for nothing.
Generative AI promised us the moon, but instead, we got another overpriced email generator.
ChatGPT might churn out text that sounds vaguely human, but it’s not exactly revolutionizing productivity, is it? In fact, its only true achievement is making everyone feel vaguely uncomfortable about how much of our lives it’s going to take over, while it quietly does nothing to actually generate real-world value. The biggest mystery of the moment is why we still haven’t figured out that it’s not what it does, but how much it doesn’t do that matters. I am glad, I turned from being a sheepl fanboi into a gnarly critic, to be honest.
And just when it seemed like we were ready to take the scenic route to collective bankruptcy through the Generative AI bubble exploding, here comes Agentic AI.
Darnit.
Why can’t they just accept failure, go through the pain, and rebuild things with much less hype.
I tell you why.
Because the entire upper class of Big Tech has thousands of billions invested in this shit.
Read: The AI investment fest- corporate hunger games edition | LinkedIn
Now, Agentic AI is the next, and corporate-approved savior that has the audacity to promise what its predecessor failed to deliver, and that is actual money.
Yup.
That’s right.
GenAI was the playing around with cute little chatbots and endlessly regurgitating semi-coherent paragraphs, but Agentic AI decided that it would just straight-up cut to the chase and handle the one thing that actually keeps the lights on – that is generating revenue.
Forget the gimmicks.
Forget the fantastical claims.
Agentic AI says, “I’m not here to play, I’m here to cash in on your cluelessness”. Agentic AI might actually be able to answer the question that’s been plaguing every tech startup since the dot-com bubble burst: “How the hell do we turn this into money?”
But before we anoint this new savior, let us first take a good look at its flaws. First off, Agentic AI is plagued by the same hallucination issues that have long bedeviled its predecessors. A study revealed that AI chatbots can hallucinate up to 27% of the time. In time, the bots have gotten better at this, and there’s even technology on the market that can help reduce this by fact checking the answers, and a recent publication stated that for regular texts, the hallucinations were reduced to 1.3% for ChatGPT (and the worst was 3%), but still you don’t want this to happen in your financial transactions. One in a hundred. You don’t want that to happen in NO transaction at all.
Relying on such a system for financial transactions is like trusting a compulsive liar with your bank account. These hallucinations are critical failures that lead to disastrous decisions, especially in high-stakes environments.
Agentic AI is totally unprepared for enterprise deployment.
Many AI platforms that claim they have agentic capabilities lack essential enterprise features like scalability and failover mechanisms.
I wrote a rather lengthy article (as usual), which I haven’t shared with y’all on LinkedIn about what’s needed for Agentic AI to operate on an enterprise level. The article is called “Why AI Frameworks Fall Short and I’ve Got a Solution”. Give it a read if you dear to, though it’s still in draft.
Current tools such as LangChain and Semantic Kernel are more comparable to simple “toolkits than actual frameworks”. They are suitable for prototyping but crumble under the weight of enterprise-grade tasks. They are unable to handle the complexities and scale required by large organizations.
The modular architectures that is used in Agentic AI systems, like microservices, are two-edged razorblades. They offer a nice degree of scalability and agility, but they also introduce significant drawbacks. According to Booz Allen (no, that’s not a nickname), this approach has some “significant drawbacks“, including increased complexity and potential integration issues, because you cannot mix agents from different vendors.
And there is one thing that needs to be said, and that is the AI trust paradox. This looms large over Agentic AI. When these systems become more adept at mimicking human behavior, users find it increasingly difficult to discern accurate information from fabrications. This paradox undermines trust in AI systems, because they are very proficient in generating plausible but incorrect responses and that can lead to misguided decisions. And this erosion of trust is unacceptable in mission critical applications. .
- Gartner’s HypeCycle terminology for when we have reached the peak of inflated expectations and standing on the shoulder before we slide down to the so-called trough of disillusionment.
Microsoft pulls the plug on AI server investments
Nadella had thought carefully about his words before he announced them on the show. Because a few days later Microsoft announced that they are backing out of the server game.
You would think that the company that practically is the cloud would have a little bit more faith in its own infrastructure. But here we are. After years of pushing to dominate the cloud market, Microsoft is now hesitating on investing further in its server infrastructure. And that raises a few eyebrow-raising questions.
Is this a tactical move?
Is is a sign of internal panic?
Or is it just the latest chapter in the corporate circus of shifting priorities. It is starting to feel like I am watching a Big Tech giant slowly pull the rug out from under itself and at the same time trying to convince us it’s “all part of the plan”.
Let’s talk about the bigger picture here.
This isn’t only about Microsoft retreating from hardware. This is a deeper reflection of the tech industry’s obsessive rush to “scale up” at any cost. A few months ago, everyone was jumping over themselves to invest in servers, data centers, and the next big thing in digital infrastructure.
Heck! Sam the Scam even went around begging for money to fund his $500 billion Stargate project (read: (One project Stargate please. That’ll be $500 Billion, sir. Would you like a bag with that? | LinkedIn). Microsoft did not pledge to make any monetary investment in this hallucinogetic dream, and now they’re quietly stepping back from investment in AI server capacity. The most cynical among us, including myself, may wonder if the whole tech world is finally realizing that there’s only so much money to go around, and they’ve already maxed out the credit card. After all, when AI projects don’t exactly rake in the cash and your biggest gamble is still stuck in its “future potential” phase, then maybe it’s time to stop pretending that big servers will magically fix everything.
And then there’s the wooley mammoth in the room.
This shift in investment priorities isn’t happening in a vacuum of course. Microsoft has already popped down a whopping $12 billion on OpenAI, and at the same time, they are pulling back on the physical infrastructure that is needed to support those ventures. It’s almost like they’re trying to juggle too many balls, and realizing they can’t even catch the ones that they’ve already thrown in the air.
The company seems to be saying, “We’ll let other players, like SoftBank, handle the heavy lifting in terms of physical infrastructure.
This is a big move for an even bigger company, and it is a bloody red flag. It’s a signal that something even more uncomfortable is coming for investors, that Microsoft might be running out of ways to make the whole “AI-cloud revolution” actually profitable. It’s easy to throw billions at new tech projects, but what happens when the returns just aren’t adding up.
When you have poured billions into AI, and now find that it is not quite as revolutionary for the bottom-line as it was promised.
So what do you do?
You stop investing in the physical infrastructure that supports it and double down on the “magical” parts of AI that sound good in a conference presentation but haven’t proven they can keep the company afloat.
It looks like the reality sunk in, and the future of AI doesn’t look quite as bright as the tech bros had us believe.
So why the hell are they still throwing money at it
It would be easy to dismiss Nadella’s comments as just a little bit of corporate self-preservation, and especially when you consider how much Microsoft has invested in AI. But the thing is that Nadella’s company injected $12 billion into OpenAI.
That’s billion with a “B”.
And now them peeps at Microsoft are publicly distancing themselves from AGI, yet theyt are still betting big on AI. They’re betting on the future AND publicly admitting at the same time that the present situation is a has been.
Then there’s the money that Microsoft was supposed to invest in Sam’s little $500 billion Stargate project. Microsoft’s initial backing for the project had a lot of people wondering if they know something we don’t. Nadella said, “I’m good for my $80 billion”. He made it sound like it’s pocket change.
Yeah, sure, Satya.
I’m just gonna believe you on that one.
But recently Microsoft decided to step back from a direct investment in Stargate, and that had a couple of reasons. First of all, OpenAi wanted to collaborate with multiple partners on their infrastructure. And because Softbank is their largest investor in Stargate, OpenAI has made a deal with Softbank to use their servers. And that is a big blow in the face of Nadella. Another reason is that Microsoft wants to optimize its spanding and offload some of the AI infrastructure cost to other investors, and then there’s a more strategic reason why they are doing this and that is that they are shifting towards product-development instead of solely building training data centers.
Is Microsoft shifting strategy or covering their tracks
So, is Microsoft really backing out of AI infrastructure investments because the returns just aren’t stacking up, or is this another strategic move that is camouflaged by a dose of corporate parseltongue?
Nadella has played it safe by admitting that the AI revolution hasn’t exactly delivered on what it promised, but the truth may be a bit darker. After all, this is the same company that poured $12 billion into OpenAI, and yet, now they’re pulling back from the servers they themselves claim are necessary to support their ambitions.
Makes you wonder.
Perhaps it’s not about poor ROI.
Maybe Microsoft’s looking for a way to justify their shift in strategy. Face it, because backing out of a tech venture when you’ve already committed billions is hardly a graceful move. It’s easier to say, “AI isn’t generating returns, so let’s focus on the software side” than to admit that you’ve simply overestimated the technology.
And while the investment in AI might not be paying off as expected, let’s not fool ourselves into thinking Microsoft is walking away from the money pit entirely. They surely aren’t abandoning AI altogether, they’re doubling down on it but in a slightly different form.
Namely, agentic AI.
Their next big thing.
Because why not launch one AI product after another when you’ve already invested your company’s future into this tech? Microsoft is pulling back on servers, but the cash is still flowing into the AI ecosystem. Agentic AI promises to solve the very problem generative AI failed to address and that’s making money. But it is riddled with the same issues like hallucinations, limited enterprise readiness, scalability problems. . . . So why continue to invest in a concept that still hasn’t delivered on the hype? It’s almost like they’re trying to distract us with new toys while they work to figure out how to turn AI into something that actually justifies their massive spend.
The real question here is, does Microsoft know something we don’t?
Or is this just another smoke-and-mirrors move that is designed to cover up their faltering AI ambitions. One can’t help but wonder if this is a sign of a broader strategy shift. One that recognizes AI’s limitations and focuses more on software and services that integrate AI rather than build the AI itself.
After all, building an empire on cloud infrastructure was the golden goose once upon a time, but as the tech industry matures, it seems that the real money might be in taking the AI models developed by others and integrating them into your own products. Maybe the lesson here is that when it comes to Big Tech, you can never trust them at face value, especially when they’re busy to make their bets on AI pay off before the market catches on to just how hollow it all really is.
Signing off from the big lie that is Big Tech.
Marco
Well, that’s a wrap for today. Tomorrow, I’ll have a fresh episode of TechTonic Shifts for you. If you enjoy my writing and want to support my work, feel free to buy me a coffee ♨️
Think a friend would enjoy this too? Share the newsletter and let them join the conversation. Google appreciates your likes by making my articles available to more readers.
To keep you doomscrolling 👇
- The AI kill switch. A PR stunt or a real solution? | LinkedIn
- ‘Doomsday clock’: it is 89 seconds to midnight | LinkedIn
- AIs dirty little secret. The human cost of ‘automated’ systems | LinkedIn
- Open-Source AI. How ‘open’ became a four-letter word | LinkedIn
- One project Stargate please. That’ll be $500 Billion, sir. Would you like a bag with that? | LinkedIn
- The Paris AI Action summit. 500 billion just for “ethical AI” | LinkedIn
- People are building Tarpits to trap and trick AI scrapers | LinkedIn
- The first written warning about AI doom dates back to 1863 | LinkedIn
- How I quit chasing every AI trend (and finally got my sh** together) | LinkedIn
- The dark visitors lurking in your digital shadows | LinkedIn
- Understanding AI hallucinations | LinkedIn
- Sam’s glow-in-the-dark ambition | LinkedIn
- The $95 million apology for Siri’s secret recordings | LinkedIn
- Prediction: OpenAI will go public, and here comes the greedy shitshow | LinkedIn
- Devin the first “AI software engineer” is useless. | LinkedIn
- Self-replicating AI signals a dangerous new era | LinkedIn
- Bill says: only three jobs will survive | LinkedIn
- The AI forged in darkness | LinkedIn
Leave a Reply