If there is one thing Marc Benioff loves, it is AI, and this lust has driven him utterly delusional. Remember when we were all snorting with laughter at Musk’s pathetic AI humanoid robot? The one so inept at even pretending to be functional that the intern teleprompting it, like a crap 90s Disney animatronic, caused it to catastrophically fail and fall on its arse by simply logging out. At the same time we were all pointing and laughing, Benioff was singing its praises, saying it could transform how businesses operate, and shared a video of him interacting with Optimus to prove it. In this video, Optimus was not just a laggy, buggy mess, but we never see it completing the painfully simple task Benioff gave it of taking him to the kitchen, so he can get a Coke, heavily suggesting it never did. Let’s not forget, it is a humanoid robot, it is meant to go to the kitchen to get you a Coke, not guide you there! You lowered the bar to the floor, and it still failed. And that was, apparently, the best thing Benioff had to back up his inane claim. Genuinely, Benioff glazes AI so hard that it feels like a perverted, delusional fetish. Which is a problem, as Benioff is the CEO of Salesforce, one of the largest customer relations companies on the planet. So, he can follow through on his perverted delusion, and last September, he did just that by firing nearly half of Salesforce’s staff and replacing them with AI. But all this did was prove how inept, manic, depraved and hollow the AI revolution, and Benioff, are.
Let’s start at the beginning.
As a company, Salesforce relies heavily on its staff. Most are administrators, consultants, business analysts and developers, who work closely with the likes of Amazon, Walmart and Microsoft to build and operate bespoke customer relations systems for them. Salesforce is possibly one of the most consumer-focused companies on the planet, as its entire sales pitch is optimising and navigating the highly complex and social world of business customer relations.
Back in September, Benioff announced that Salesforce had an agentic AI capable of increasing productivity so much that it was handling customer conversations at scale and was doing 50% of the work at Salesforce. This meant that they had fewer cases to worry about, and no longer needed to “actively backfill support engineer roles.” So, Benioff indiscriminately fired 4,000 of their 9,000-strong staff, as this magical AI had replaced them.
Magic, the thing which doesn’t exist and instead refers to tricking people, is the right word here because, guess what? An AI absolutely cannot replace such a socially and technically complex job, and Salesforce is only just finding that out.
Recently, senior executives at Salesforce have admitted, both internally and publicly, that they massively overestimated AI’s capabilities. They have found that AI simply can’t cope with the complex nature of customer service and totally fails at nuanced issues, escalations, and long-tail customer problems. They even say that it has caused a marked decline in service quality and far more complaints.
But the problems go far deeper than that.
Both employees and executives have said that the company is wasting countless resources on firefighting to stabilise operations since the mass AI layoff. Employees have to spend so much time stepping in to correct the wildly wrong AI-generated responses that AI is wasting more time than it saves. In other words, this AI reduces productivity, not increases it.
But there is also a huge problem here with expertise and skill debt. On top of the firefighting to correct the AI, executives have also highlighted how they are also having to firefight to stabilise their systems from problems that were previously easily solved by staff who had the required experience and skill. However, these staff were fired in the AI layoffs.
Expertise, experience and skilled employees are really hard for a company to acquire. You see, much of the expertise, experience, and skills required are unique to the company and its operations. These operations will have quirks, common problems, and unique issues that even the most experienced outsider will really struggle with, but are effortless to someone with experience within the company. As such, these attributes are not only vital, but are nurtured and grown within a company, and cannot be hired in on a whim. What Salesforce has done is chuck all this experience out the window, and now they are suffering.
To say this is a damning indictment on Benioff, his capabilities as a CEO, and his cult-like push for AI is an understatement. It demonstrates that he is dangerously ignorant, as it was painfully evident that this would happen; more on that in a minute.
Salesforce is doing a U-turn, and is “reframing its AI strategy,” shifting away from trying to replace workers, and instead doing what executives call “rebalancing.” Basically, they are trying to re-hire these critical roles and expertise they have lost, take AI out of the driver’s seat, and augment workers rather than replace them. More on why that is still a terrible idea in a minute.
Now, those fully-paid-up members of the “AI revolution” cult will say that this is an example of incorrect AI deployment, and an issue isolated to Salesforce. And, while I do agree that Benioff did a truly horrific job here, that isn’t true at all. You see, this ignorant AI overconfidence and dramatic deployment, followed by a rapid walk back, is happening across the entire economy.
An Economist Survey from March 2025 found that 74.1% of companies with 100 to 249 employees, and 62.4% of companies with over 250 employees, said they did not use AI. Those figures had increased to 81.4% and 68.6% respectively. These findings are corroborated by recent analysis, which found that AI adoption rates in large companies fell from 14% at the beginning of 2025 to 12% by the end of 2025. There is also a recent analysis which found that the cancellation rate of corporate AI initiatives dramatically increased in 2025 to 42% from 17% in 2024. Then there is the now-infamous MIT research, which found 95% of corporate AI pilots failed to deliver any gains, and the METR report, which found that AI slowed down experienced workers.
Basically, every company that has tried to realise the promises of the AI productivity hype has been burned by the reality of its truly awful performance and is now desperately trying to row it all back.
Again, delusional members of the AI-cult will still praise CEOs like Benioff for being brave and taking risks. After all, the potential upside of AI is seemingly enormous! Not having to pay half your workforce is a colossal cost saver. To them, the risk is worth the reward.
Except, it absolutely wasn’t, as we have known AI does this for ages! In other words, the rewards never existed; we knew they were fictional. So, these are truly ignorant, needless risks that shoot yourself in the foot with a nuclear bomb levels of self-sabotage.
Take the Carnegie Mellon University paper that found AI agents fail basic tasks 70% of the time. The aforementioned METR report, which found that AI coding tools are so inaccurate that it takes more time for the developer to correct their output than the time the developer saves by using the AI. Coding is arguably the best application for AI, as the highly constrained nature and clear objectives should dramatically increase accuracy over more complex, open-ended tasks, such as customer service. A few months before Salesforce’s mass layoffs, Anthropic tried to see whether its Claude AI could run a vending machine, another highly constrained task that AI should find easy. It failed spectacularly and not only failed basic tasks like stock management, but also lost every penny. So, the fact that AI can’t increase productivity in these ideal constrained situations, means it likely can’t for most applications. In fact, Melbourne researchers have corroborated METR’s findings and found that AI’s rampant errors and inaccuracies means it can’t even augment the simplest of tasks, as the time it takes to find and correct its errors is more than the time saved by using AI. They did find that it can sometimes improve the productivity of the lowest-skilled jobs, but only as a tool to smooth out already flawed worker outputs. For example, someone transcribing a meeting in a second language they aren’t fluent in would benefit from AI help, but the results still won’t be good, and will still require correction.
Basically, we have known for years now that AI is too inaccurate to even augment the most constrained jobs, so it can’t improve productivity through augmentation, let alone automate jobs.
There is an argument that these AIs will get better, and these problems will be solved, but even that position is not supported by reality. Things like the efficient compute frontier and the Floridi conjecture mean that AI is about as good as it will ever be, even if trillions of dollars are poured into its development (read more here for an overview of this). So no, AI isn’t magically going to get infinitely better overnight; that is just tech bro bulls**t.
But, we also know that deploying AI into a company seriously damages the capability of its workforce for ages, too.
Take the studies from JYX and Carnegie Mellon, which found that workers of any skill level lose vital skills, critical thinking skills and expertise when they use AI. Like a muscle, these attributes require constant use to maintain; otherwise, they erode away. Even massively AI-biased research from think tanks explicitly pushing for AI adoption has shown that using AI to augment jobs creates cognitive offloading, leading to massively detrimental skill atrophy, which damages both the worker and the business. These studies all point out how this skill atrophy is a vicious circle. Not only does it make firefighting and problem-solving harder, as this heavily relies on skill and expertise, but as the worker’s skill and expertise reduce, their ability to spot and correct the constant errors the AI makes. So, they let more errors into the system, which creates more problems, which they now lack the skills to solve, or even spot, so more errors are created and so on.
If Benioff and his fellow AI layoff CEOs actually paid attention to the science and research, and didn’t just suckle at the teat of neo-fascist tech bros, they would have known from the get-go that AI cannot augment, let alone automate, workers. They would also know that AI is pretty much guaranteed to horrifically undermine their biggest asset, the wonderful human skills and expertise of their workforce, if they try to widely augment them with AI.
This is why all these AI layoffs are, actually, totally bollocks! And that isn’t just me saying that, but the boffins at Oxford Economics, who advise the UK government.
They published a study a few days ago that destroys the AI Layoff narrative. They found that “firms don’t appear to be replacing workers with AI on a significant scale and we doubt that unemployment rates will be pushed up heavily by AI over the next few years”. Basically, they found that most mass layoffs were being presented as AI-related so that CEOs could look better to investors. It reframes them from “times are hard, we need to let people go” to “look, we are being innovative!” while not actually replacing said workers with AI. It seems Benioff didn’t get that memo.
All in all, Oxford Economics estimated that only 4.5% of layoffs were caused by AI replacing workers in 2025, and they even suggest that figure is likely far higher than reality.
Oh, and just to tie everything in a nice, neat bow, they also found that AI can’t automate jobs, and AI augmentation is not making workers more efficient, and that this won’t change for a good while. Honestly, I think I need to write an article just on this study.
The “AI Revolution” is closer to a hoax than reality. It is a fable, a fiction, it only exists in the minds of its fanatical followers. And fanatical they are because they have to be.
The likes of Benioff have not only swallowed the “AI revolution” propaganda hook, line, and sinker but have also championed and projected it, whilst restructuring their entire world around these demonstrably false promises. The damage their actions are causing, and the drastic U-turns they are making, are a damning indictment of their capability and value as leaders.
But it goes deeper than that.
The big lesson from all of this is that a workforce’s skill, experience and wellbeing are possibly the most valuable assets any organisation has. Not only are these critical to their functionality, but it takes time and care to nurture and grow these vital attributes. This is the only reason why these AIs deal such huge damage and are so useless. They undermine this valuable and incredibly difficult-to-acquire asset.
The failure of the “AI revolution” shows that the wider corporate world and its leaders, like Benioff, are causing immense damage by drastically undervaluing their workforce by being wilfully ignorant of just how much everything depends on them and their well-being. It strongly suggests that this is the asset they should be investing in, not this AI bollocks, because this is what underlines the core value of what they do. It heavily suggests that if these leaders pay their workers properly, treat their experience as valid and unique, bring them into leadership decisions, invest in their wellbeing, nurture their growth and protect them through hard times, then they can create real, tangible, solid growth, as they can grow the very foundation of the organisation’s value.
There is a wider conversation here as to how the entire Western economy has transitioned into a death cult. But all I will say for now is, do you think the likes of Benioff have the capability, the humility and the empathy to recognise this lesson? Or do you think they will carry on in the deluded, egotistical ways?
Thanks for reading! Don’t forget to check out my YouTube channel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!
Sources: Tech Crunch, Arxiv, Futurism, METR, Fortune, Huff Post, Maarthandam, CNBC, Will Lockett, Will Lockett, Will Lockett, Will Lockett