Reminds me of the old joke "90% of the code is 90% of the work. The last 10% of the code is the other 90% of the work."
I have spent almost my entire adult life (since 1986) shipping products. One of the very first things that I learned, was that "shipping" > "designing".
There's so much work in delivering products that will carry your brand, and then must be supported.
I liken it to having children. Conceiving them is fun. Delivering them is painful. Raising them, is a lifetime of work.
In my experience, the same type of thing applies to products that we ship (and charge money for).
Fully agree. Shipping a complete product with a functioning user acquisition funnel is much harder. It's like; you have to build the whole product first with lots of features and then you have to try to create a highly condensed overview of all those features to expose them all on the landing page.
If you can't make the visitor understand your entire complex product in 10 seconds, then you've lost them.
Your product has to be complex because that's where the software market is at. All of the low-hanging fruits have been taken by the time you identify them. Sure, someone will find a way to make money using new low-hanging fruits that arise due to technological changes but it's not going to be you. You probably don't have the business connections to make that work.
> There's so much work in delivering products that will carry your brand, and then must be supported.
People think otherwise with AI partly because Anthropic kept telling us that they didn't have to write code or review code any more for most of their work. Their agent swarms just comb through their github, slack and wikis to figure out what to do next, and another swarm of agents just review, test, merge, deploy, A/B test, and revert the code. Boris alone merged nearly 300 PRs in the past week (or two?). So the top research labs seem have broken the productivity seal.
And then they talk about this recursively self-improving AI that is so powerful, so autonomous that they advocate that every company should be prepared to "pause" the effort. And their Fable/Mythos has this specific restriction as mentioned in their model card[1] that they are going to reject requests to tune and train models because, well you guess it, the models are too powerful to be used by mere mortals.
[1] We’ve implemented new interventions that limit Claude’s effectiveness for requests targeting frontier LLM development (for example, on building pretraining pipelines, distributed training infrastructure, or ML accelerator design). Using Claude to develop competing models already violates our Terms of Service, but enforcing this restriction through our safeguards avoids accelerating the actors most willing to violate these terms. Unlike our interventions for cybersecurity, biology and chemistry, and distillation attempts, these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT).
I like this analogy; raising children well like delivering products well pays dividends. They’re less likely to cause problems and if they do, they tend to be smaller in scope.
I skimmed the article, guilty, but I think what I got from it is that CEOs will CEO? No disrespect meant, I’ve seen your name here often and thoroughly enjoy the folklore that you share, but I don’t understand what context you reacted to. Cheers.
The context that they think that shipping is simple. Shipping (what you need all those annoying peons for) is really terribly difficult, and has a lot of moving parts that designers often fail to take into account, until the deployment people lock them into a restroom stall, and refuse to let them out, unless they listen.
That's common with newer engineers (and now, non-engineers). I believe that Mr. Dunning, and Mr. Kruger had something to say about it.
I also spent most of my career at hardware-oriented companies, and shipping hardware is orders of magnitude more difficult than shipping software.
You can actually get high-quality code out of them -- at least with Claude; not had a great experience with Gemini -- but for complex tasks requires riding them very, very hard and really understanding where things can go wrong and poking at them repeatedly. Iterate, iterate, iterate.
That describes my last week. What made it most annoying, was the need to release through TestFlight, because the memory issues would not appear, when tethered. Also, I was checking in constantly, because I had to revert and reset the context, several times.
Yup. In another post, I was grumping about having to accept a truly obese bunch of code from an LLM. This particular issue is the only one I could imagine even considering accepting that much pasta, but I sort of have to, because the LLM was able to quickly solve a problem that would have taken me a couple more weeks to address.
I remember a .sig that went something along the lines of:
I hate code, and want as little as possible in my software.
now it's closer to 95% of work can be done by AI and requires 5% mental effort, but 5% of the work requires 95% of the mental effort to finish because of all the unoptimial decisions AI has taken. I find that AI works best in small micro-service type architecture where each component has a clear goal and doesn't have interconnected parts within the same application that can break. But you do run into an issue where changes in microservice a need changes in microservice b and updating it is not ideal since it usually cascades thru the entire system or requires stacks of legacy support.
IME it’s possible to have good clear APIs, limited scopes/goals, etc in a normal (macro?) service. But it requires a level of discipline and process many teams are unwilling to engage in.
There are a lot of bad CEOs, though. It's a lot like a politician -- it's quite difficult to become a CEO, and the skills to make it to that position don't always intersect nicely with the skills necessary to actually do the job well.
Most of them got into a prestigious school on legacy, paid for by wealthy parents. Many were above average IQ, but by no means geniuses. They had access to computers earlier than others, due to said affluence. They seem unable or unwilling to comprehend they're overwhelmingly on average, "nepo babies" to steal a term from the world of entertainment.
CEOs do get there with lots of politics in almost all cases. It’s all about who’s ass you kiss and who’s ass you don’t and if you’re lucky with timing things might just fall into place.
I think it’s exceedingly rare that a CEO is actually competent at their job. In most cases it’s the labor class propping the company up, and in some cases the workers are doing so against the wishes of the CEO. Not that executives want to ruin the company, they’re just incompetent and therefore make terrible decisions constantly.
I'm very sympathetic to cooperatives, have traveled/know the Mondragon people (largest coop federation), etc.
However, I think there's a reason why coops seem to succeed at smaller scales, but there are essentially no large innovative coops.
There are a few large boring coops, and some small innovative ones, but seemingly something is making the CEO/investor board model the one large innovative companies are all using.
I suspect that it's both (1) access to capital is far harder for coops, and (2) that workplace democracy and hardcore mission focus aren't fully compatible. That is, "you cannot serve two masters" without losing focus on one of them.
If a company accumulates capital, it becomes vulnerable to the principal agent problem, and coops are way more vulnerable here than centralized companies.
If a company doesn't accumulate capital, it doesn't scale in complexity. It can grow by having more people do more of the same things, but it can't move into markets that demand anything complex.
The problem is that knowing the right people to get investment does seem to have utility coops struggle to get, I think? maybe CEOs are basically like producers on movies who are just there to network for you.
In software I can imagine a worker-owned consultancy, but not a product company. It would imply staying in one place working on one product for your whole life, which doesn't sound inspiring
A company need not be a single product, and working in a worker-owned cooperative need not be a lifetime commitment to a single firm (though cooperatives ideally will have less turnover than firms owned by capital separated from labor.)
Acting like centrally planned dictatorships is a good form of collaboration is just so off base. There's no reason to think that introducing democracy into the work place wouldn't immediately benefit both workers + customers.
If this sounds crazy the C suite + board already vote on who gets hired into the executive team, vote for the direction of the company, and vote for their compensation packages (hint, they never decrease them).
Why shouldn't workers be legally enabled to do the same? What is the justification to this? I'm curious to hear it because the only way people can justify the current system is declaring that some people are actually more deserving of prestige, money, and benefits while others deserve to suffer.
With income inequality increasing, healthcare outcomes worsening, and children literally becoming stupider isn't it time to question the current system and ask ourselves if this is the society we truly want?
There is academic research on this too if you're curious but it's mostly in English, Spanish, and Portuguese.
But yes, there isn't much "evidence" because this system hasn't been tried en masse; however if you look at our current neoliberal hellscape, it's pretty hard to imagine it doing worse. Also neoliberalism wasn't really "tried" either, it was thrusted upon us by a group of individuals that wanted it.
One thing to keep in mind is that society can change quite quickly if you want it to. I'm sure the children that died working in factories during the 1800s never imagined such a society where children are valued, cared for, educated, and protected but it did happen.
It has happen before and it can happen again. It only happened because people were willing to fight for it.
The rules are allowed to be changed at anytime if we deem so, a better world is possible.
Worker owned cooperatives have a variety of ways of doing this. Voting directly, electing people, etc. The main difference is that the cooperative typically doesn't buy the myth that the person making the high level decision needs to be paid 1000x the workers.
Maybe, but not necessarily for this reason. Even in a worker-owned coop, someone sets the overall direction. And how is that person going to be selected? It's still going to be largely politics.
on the contrary, it seems to be one of the few jobs that seems to require absolutely no qualifications to have.
What you need to do to be CEO is.... convince someone to lend you money in the hope that you'll get it back to them.
I've worked under some absolutely awful people who wouldn't pass an interview anywhere, but somehow they're CEOs, because they can smarm there way into more money consistently.
>convince someone to lend you money in the hope that you'll get it back to them
And it should be noted that many of these people lending money are in a similar situation of not being required to have any qualifications. Sure, some of them have worked their way up through sound investment after sound investment, but many of them were either born into their position or simply got lucky at some point along the way. Just think of all the money investors threw away pursuing crypto and NFTs for example. Many of those investments were transparently stupid from day 1.
Often, they are good at taking things, keeping things, misdirecting and setting boundaries (especially communication boundaries). They are good at keeping their positions.
This is a broad range of skills and to actually be a CEO, you need to really hone these skills and be among the very top. To be good at those, just enough to qualify for a modest CEO role at a small start-up, you generally don't have the time to be good at anything else.
Saying that you don't need any skills is mischaracterizing it. You don't need any value-creating skills, yes, but you need significant value-capturing skills.
I can imagine a world were all companies become empty of workers and only executives remain and they would just have meetings with each other while they starve and would explain it away as a new diet they're on. There would be no petrol and they would be forced to walk to work and would say that it's their new fitness routine... And they would all believe each other.
It's not difficult at all to become one, and the work involved in being a CEO is not particularly difficult in comparison to senior technical work at all. The only thing that is harder about being a CEO is the responsibility. I'm sure being the CEO of Microsoft or whatever is plenty difficult and demanding in many ways, but most CEOs are not that, and speaking just from experience most CEOs and CTOs are clueless morons.
With that said, I've been programming for 25 years and I've only been a CEO for 3, so take what I said with a pinch of salt.
I do think people overestimate titles like this a lot, though, and it really comes down to what the company actually does and what is demanding for that company at that position/role. The CTO of a some-bullshit-as-a-service company may as well be straight out of college, because they're likely doing something trivial that literally anyone (including LLMs) could put together. The CTO of a well-used and reliable streaming service that handles a meaningful part of the world's Internet traffic is obviously solving a more interesting and demanding problem, and their decisions are going to be more important.
He's the face of bad CEOs because people like to make up things about how bad he is. The Social Network, a major 2010 biopic about the early days of Facebook, famously cut his college sweetheart and now-wife out of the story in favor of a fabricated character arc involving an ex-girlfriend who does not exist.
Poor Marky Z. always getting a bad rap. Let s/he who hasn't made zillons getting billions of people hopelessly addicted to social media while facilitating a genocide or two and generally destroying the world order as we know it throw the first stone.
From your tone, it sounds like you may just be intending to warn me that it's cringe not to agree with any criticism of Zuckerberg? If that's so, I have to respectfully disagree; I think this is a bad attitude that leads towards being poorly informed about the world.
If you're interested in discussing the specific claims you're making, I really don't think that billions of people are hopelessly addicted to social media, and I would love to hear your basis for claiming this. My understanding (from e.g. https://www.nature.com/articles/s41598-025-27053-2) is that genuine compulsive use of social media is quite rare, and most people who describe themselves as "addicted" are just regular users who enjoy it but kinda feel like it's a waste of time.
> Meta Settles Lawsuit That Claimed Social Media Addiction Screwed Up Schools
> On Thursday, Meta settled a lawsuit brought by a Kentucky school district that claimed the tech giant’s social media platforms have created a mental health crisis at its schools.
> The case was considered the first of its kind and a bellwether (a case that is representative of a large pool of lawsuits and will be a test for future litigation). The plaintiffs argue that social media platforms have had a major negative impact on the mental health of school-age children, which in turn has caused a burden on the education system, as American schools were forced to redirect resources to counter this problem.
> The settlement comes shortly after Meta lost a key bellwether social media addiction trial. Back in March, a judge in Los Angeles ruled that Meta was liable for the adverse mental health effects a now 20-year-old suffered after getting addicted to Instagram from an early age. The representatives of the young woman argued successfully that it was Meta’s deliberate design choices, like the infinite scroll and face-altering filters on stories, that had exacerbated her addiction and subsequent mental health issues like self-harm and depression.
> The military junta in Myanmar was facilitated by Facebook to post hate speech that sought to foment sexual violence and promote genocide against the Rohingya. "Myanmar would have been a better place if Facebook had not arrived" Wynn-Williams writes.
> Wynn-Williams argued that Facebook failed to moderate hate speech against the Rohingya in Myanmar, including the use of the racial slur kalar. She noted that the company only had two Burmese language moderators, both based in Dublin, for the entire country, and claimed that one of the two moderators gave a pass to hate speech while removing pro-human rights content. She further claimed that she raised concerns that the moderator was "in cahoots with the" junta, only to have her concerns dismissed by the content team. Additionally, she claimed that her efforts to have Facebook's Community Standards rules translated into the Burmese language were resisted by the company communications team, who told her that "Myanmar isn’t a priority country" in the region.
> In the 2010s, personal data belonging to millions of Facebook users was collected by British consulting firm Cambridge Analytica for political advertising without informed consent.
> The data was collected through an app called "This Is Your Digital Life", developed by data scientist Aleksandr Kogan and his company Global Science Research in 2013. The app consisted of a series of questions to build psychological profiles on users, and collected the personal data of the users' Facebook friends via Facebook's Open Graph platform.[2] The app harvested the data of up to 87 million Facebook profiles. Cambridge Analytica used the data to analytically assist the 2016 presidential campaigns of Ted Cruz and Donald Trump.
> Other advertising agencies have been implementing various forms of psychological targeting for years and Facebook had patented a similar technology in 2012.
> Facebook manipulated the emotions of hundreds of thousands of its users, and found that they would pass on happy or sad emotions, it has said. The experiment, for which researchers did not gain specific consent, has provoked criticism from users with privacy and ethical concerns.
> For one week in 2012, Facebook skewed nearly 700,000 users’ news feeds to either be happier or sadder than normal. The experiment found that after the experiment was over users tended to post positive or negative comments according to the skew that was given to their news feed.
> The research has provoked distress because of the manipulation involved.
> Studies of real world networks show that what the researchers call ‘emotional contagion’ can be transferred through networks. But researchers say that the study is the first evidence that the effect can happen without direct interaction or nonverbal clues.
> Anyone who used the English version of Facebook automatically qualified for the experiment, the results of which were published earlier this month. Researchers analysed the words used in posts to automatically decide whether they were likely to be positive or negative, and shifted them up or down according to which group users fell into.
It's on my mind because they're releasing a sequel in October, and Aaron Sorkin has not (as far as I've seen) acknowledged that much of the original was not true nor promised to be more accurate this time around. I'm pretty confident that it's going to be about the same mix as last time, and I'm going to have to go around saying "actually the scene where Zuckerberg did suchandsuch terrible thing wasn't real", and people will respond by insinuating that it's lame for me to care.
CEOs understands that AI offers potential productivity increases. Using that productivity boost to cut staff is an unimaginative approach. Bolder approaches include using that boost to exceed the expectations of current customers, or to increase sales without proportional increase in staff, etc.
it's kind of like saying like you could replace a king with an AI. the position is a relationship to power more than it is a function with a productive output.
a bad king and a bad CEO could be replaced with a spinning top with no loss in productivity (and maybe some gain).
I still remember when I used ChatGPT the first time to write an email. I thought to myself “Oh. This sounds like 99% of the corporate communication from above”. We were joking that corporate had BossGPT for years and just didn’t tell anybody.
It also made me realize that most the so called “creatives” in marketing and PR also just repeat variations of the same few templates. Not much real creativity there.
My charitable word would be average. AI means average. And an average CEO isn’t that bad. It’d be a deliberate trade in exchange for less spending on salaries. (just one very large salary)
Who takes responsibility when the AI does something unethical or illegal? Do we put the computer in jail? Or do we just look the other way like we do with human CEOs?
I think Grok on x.com is a reasonable case study. There’s the stuff we heard about and then there’s a separate category of things that weren’t newsworthy, and only the newsworthy things prompted changes to be made.
Weirdly enough, I'd take the robot. At least we can pretend the robot doesn't know any better. The human is actively choosing to be a dick and profiting off it.
The flesh sack is choosing to be an insufferable twat, and the robot either doesn't have any choice or has a decent statistical justification for what it does.
>> Yes, the tools are powerful, but a CEO who thinks they replace the work of employees is simply a bad CEO.
This is a broad generalization of employees. There will be some "routine tasks" that can be done by AI, now that is a lot more powerful.
There won't be as many employees needed for routine work - for example L1 and L2 support work. For example, many companies had ML engineers building models for various models. Companies can get that off the shelf from AI companies. They don't need a big team of model builders now.
If L2 support work is the example then I doubt we’re near replacement.
L2 issues are already involved in some way often revealing some kind of system failure, requiring context and exploration to understand, and judgement (and perhaps even system overrides) to fix.
I could see “automated L2 is the new L1” improvements, but without a big capability jump and/or a resource bonfire, I don’t think even frontier models would effectively replace good L2 staff.
They might magnify good L2 staff so fewer are needed (and maybe even help L1 staff become L2).
I often use AI chatbot to generate step-by-step instruction for setting or repairing up some bit of tech. It's incredibly empowering, and saved me a a lot of money that would have been spent on buying replacement tech.
You know who can't do that? People who call L1 support.
Ideally contractors that benefit you personally (eg: your buddy who now owes you one), but definitely contractors that let you outsource the responsibility.
Even better if you get some management consultant to suggest the idea and/or do the subcontracting.
Definitely buys you a few quarters of bonus and some time to land your next gig.
The message isn't subtle, and isn't meant to be: "we don't care how, but we expect you to stick your nose into AI tools and find some way to fit them into your workflow".
Which indicates: the management believes there are productivity gains from AI use, but adoption lags due to inertia and reluctance to change existing workflows.
> Which indicates: the management believes there are productivity gains from AI use, but adoption lags due to inertia and reluctance to change existing workflows.
Methinks adoption lags due to management's inability to align incentives such that productivity gains are rewarded.
more of "we whined and cried and screamed that we needed new budget in order to buy these tools or we would literally die. now we have them, they don't work as well as we hoped, they aren't leading to productivity gains, and they're actively alienating our workforce and users alike. we're so screwed we literally have no idea how to do reverse this."
If the excavators only worked half the time, sometimes goes out of control bonking someone in the head and costs a billion dollars, then yeah they're pretty bad CEOs
A story from today - I needed a small utility to remap my logitech buttons under windows without installing their horrendous GHUB. Logitech Onboard Memory Manager still required ghub to be put into onboard mode.
The solution - linux has utility called piper. I downloaded the repo and just told codex - figure out what piper is doing and create me a small utility to do it under windows. So the jolly critter started experimenting with hex commands, then pulled some other repo on which piper depended figured out how to enable said onboard mode and 10-sh minutes later I had small python script that did what i needed to do.
That would have taken probably half a day of work for a human.
There are many stupid CEO and organizations which are not committed to quality. And a lot of employees that are too set in their ways. But the instinct that underinvesting in AI is more dangerous than overinvesting is right. Doomed if you do, doomed if you don't
"To err is human, but to really foul things up requires a computer" - this is from the 60, but right now is turned into overdrive.
This is a great article, and I agree with most of it.
The problem is that the wrong eyes are seeing it.
We need these kinds of articles to be published in places that executives read, and tailored to their audience.
AI recovery is going to be a big wave of consulting over the next several years, maybe very publicly or maybe quietly, but it's going to be a thing. That doesn't mean "all AI is bad" or any other such nonsense, it means that there are a lot of companies out there right now that are doing it wrong and will need help.
The executives that get ahead of this are going to be the winners.
Similar to Tesla naming it's driver assistant "auto-pilot" in 2015 and your average Joe thought he would be able to sleep while the car would drive him to work.
The CEO just hear AI and think of AGI. They expect Skynet.
Aircraft autopilot is over 100 years old, and Testla autopilot is an automotive version that fits the definition by analogy to what aircaft autopilot does and what the human metaphor means (doing simple tasks aithout higher-level cognition or handling of surprising stimuli). Autopilot is not end-to-end driverless.
> the definition by analogy to what aircaft autopilot does
It's a bad analogy. Autopilot just maintains the aircraft in some state, then there's the flight director which maintains the flight path, and you can connect or disconnect the two at will. When connected the director can change the autopilot state.
To use the flight director you must fully specify your flight. The weight, the fuel, the weather, expected winds, takeoff and landing runway length, runway conditions, expected brake demand, as well as every single waypoint you're going to cross and the expected state at that crossing.
> what the human metaphor means
We learned after high levels of cockpit automation that maintaining situational awareness was still required. Pilots are freed of some stress during high workload portions of the flight, provided they planned correctly in advance, and that zero changes to their flight plan (not likely) occur.
As a result pilots are mostly told and trained to hand fly the plane during take offs and landings if the weather allows for it. You should only use high levels of automation if the situation demands it.
The point isn’t whether the aviation term "autopilot" has a technically defensible meaning, but that "Autopilot" predictably made normal people think "the car drives itself", just like "AI" now makes non-technical CEOs think "AGI employee replacement machine".
You are arguing the dictionary while I’m arguing the predictable, if not very well calculated and purposeful, misunderstanding.
CEOs are probably the most replaceable position, period, by human or AI. Everyone just gives them information. They don't know any information themselves.
Problem is, a CEO can fire employees, find out it was a dumb decision, then leave with a million dollar severance package. So they don't really care when they're wrong.
The primary product of AI is labor displacement and consequently wage supression. This is what OpenAI and Anthropic are really selling. It didn't start with AI but AI is accelerating it.
This is what layoffs have been about since the pandemic. People in fear of losing their jobs do extra unpaid work and aren't asking for raises. The theoretical potential of AI gives companies the excuse to fire more people. The investment itself is directly used as a reason of why they need to cut back on labor.
Any sufficiently sized business can only feed the insatiable hunger for ever-increasing profits ultimately by cutting costs and raising prices. And what do we have now? High inflation and a decline in real wages. CEOs are just following this playbook.
And the result is that society is bouldering towards collapse. We're seeing the first hints of this with the youth unemployment crisis [1][2][3].
Also, who is going to buy anything when nobody has any money?
> Also, who is going to buy anything when nobody has any money?
This assumes that a mass consumer economy is necessary, when it isn't. Mass consumption is relatively new, for most of history economies functioned with just a small consuming elite and large underclass that consumed very little. We are already approaching that again in the states given that the top 10% of earners are already responsible for nearly half of all consumer spending.
There's a floor even in a mostly automated economy where some services are resistant to automation simply because the human element is the product. Luxury hospitality, personal care, etc. That billionaire is going to want a human masseuse, not a robot.
A highly automated economy could stabilize like this with a small elite population consuming luxury goods & services, served by a low-wage economic underclass human workforce.
Its certainly not a pleasant society, but its also not unsustainable given enough oppression or pacification (bread and circuses anyone?)
I don't have any sources regarding someone preferring humans for certain services over robots, just intuition there, and the fact that consuming human labor and time is itself a status signal of wealth, and the current growth of personal services.
As for connecting the dots, look at Brazil, one of the world's most unequal economies having a small consuming elite and a much, much larger low-wage service underclass. The gulf states as well. Granted, their circumstances don't map cleanly to a post-automation Western economy, but it does demonstrate that a largely bifurcated consumption based economy can exist and can be mildly stable.
Whether the US falls into that direction too will depending on politics. A bunch of mid-career knowledge workers aren't going to willingly to flip burgers in a service economy without some serious surveillance and oppression. But when thinking about it in those terms, the recent push for mass surveillance laws and tech along with the increasingly dangerous rhetoric around protests and "domestic terrorism" start to make sense.
and I don't know what worries me more - a burst in this bubble (and maybe some other tech stocks), or a failure of these valuations to be burst somehow, and even more concentration of capital and power around those corporations.
Most CEOs are not special. They are not especially smart, or skilled, or technical. The role self selects for sociopathy. That's not a quality that has any kind of linear relationship with intelligence. Quite the contrary.
A common misconception about AI is that it is intended to fully replace humans, which is incorrect. The purpose of AI is to reduce the need for human labor, and it has already been doing so. For example—though this is not an exact figure—a task that previously required 15 people might now only need 10. In no instance has the human element been completely replaced; rather, the reliance on manual labor has simply been reduced.
It's not exactly a misconception, when companies are pitching AI as a full and complete replacement for human employees. People are just reading the billboards on the side of the road.
I always found advertisements for AI to be so strange, why would you advertise your AI to the public as a danger for humanity that will also put everyone out of work? Such advertising would only appeal to sociopaths, but of course that's because it's intended to appeal to CEOs.
A "unit of work" that required X people to complete in Y time can now be done by X/Z people in Y time, where Z is whatever efficiency you are able to get out of applying AI tooling to your business.
For some companies, Z might be less than 1 though. ;-)
So you still need skilled people, just not the same amount as before, because you have different tools available to you.
This has happened before with other advancements in industrial/technological automation. It's not a new concept.
That sounds like 5 humans got replaced by AI. I don't think most people worry about whether all humans will be replaced, simply whether or not they will be replaced, or people they care about.
Or all 15 people are still employed doing the work of 22.5. Or even more people have been hired now that each person can generate 50% more value than they previously could. Or people are reallocated from the AI assisted task to another. Or some combination thereof.
Which is very short sighted. You or anyone close to you might not be replaced but it should be clear that you don't want to live in a society with 20% unemployment.
Reminds me of the old joke "90% of the code is 90% of the work. The last 10% of the code is the other 90% of the work."
I have spent almost my entire adult life (since 1986) shipping products. One of the very first things that I learned, was that "shipping" > "designing".
There's so much work in delivering products that will carry your brand, and then must be supported.
I liken it to having children. Conceiving them is fun. Delivering them is painful. Raising them, is a lifetime of work.
In my experience, the same type of thing applies to products that we ship (and charge money for).
Fully agree. Shipping a complete product with a functioning user acquisition funnel is much harder. It's like; you have to build the whole product first with lots of features and then you have to try to create a highly condensed overview of all those features to expose them all on the landing page.
If you can't make the visitor understand your entire complex product in 10 seconds, then you've lost them.
Your product has to be complex because that's where the software market is at. All of the low-hanging fruits have been taken by the time you identify them. Sure, someone will find a way to make money using new low-hanging fruits that arise due to technological changes but it's not going to be you. You probably don't have the business connections to make that work.
> There's so much work in delivering products that will carry your brand, and then must be supported.
People think otherwise with AI partly because Anthropic kept telling us that they didn't have to write code or review code any more for most of their work. Their agent swarms just comb through their github, slack and wikis to figure out what to do next, and another swarm of agents just review, test, merge, deploy, A/B test, and revert the code. Boris alone merged nearly 300 PRs in the past week (or two?). So the top research labs seem have broken the productivity seal.
And then they talk about this recursively self-improving AI that is so powerful, so autonomous that they advocate that every company should be prepared to "pause" the effort. And their Fable/Mythos has this specific restriction as mentioned in their model card[1] that they are going to reject requests to tune and train models because, well you guess it, the models are too powerful to be used by mere mortals.
[1] We’ve implemented new interventions that limit Claude’s effectiveness for requests targeting frontier LLM development (for example, on building pretraining pipelines, distributed training infrastructure, or ML accelerator design). Using Claude to develop competing models already violates our Terms of Service, but enforcing this restriction through our safeguards avoids accelerating the actors most willing to violate these terms. Unlike our interventions for cybersecurity, biology and chemistry, and distillation attempts, these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT).
> 90% of the code is 90% of the work. The last 10% of the code is the other 90% of the work.
Don't think I've heard that one but certainly rings true to my experience.
Reminds me of "ninety percent of the game is half mental"
I like this analogy; raising children well like delivering products well pays dividends. They’re less likely to cause problems and if they do, they tend to be smaller in scope.
I skimmed the article, guilty, but I think what I got from it is that CEOs will CEO? No disrespect meant, I’ve seen your name here often and thoroughly enjoy the folklore that you share, but I don’t understand what context you reacted to. Cheers.
The context that they think that shipping is simple. Shipping (what you need all those annoying peons for) is really terribly difficult, and has a lot of moving parts that designers often fail to take into account, until the deployment people lock them into a restroom stall, and refuse to let them out, unless they listen.
That's common with newer engineers (and now, non-engineers). I believe that Mr. Dunning, and Mr. Kruger had something to say about it.
I also spent most of my career at hardware-oriented companies, and shipping hardware is orders of magnitude more difficult than shipping software.
> Conceiving them is fun. Delivering them is painful. Raising them, is a lifetime of work.
Then there's the technical debt!
Shipping is frankly the easy part. It's the operating overhead that often breaks you.
I liken it to free puppies.
This is true.
I have always prided myself on writing concise, high-Quality code, because it tends to be quite debt-free.
So far, LLMs seem to deliver code with "Louie Da Loan Shark"-levels of tech debt.
You can actually get high-quality code out of them -- at least with Claude; not had a great experience with Gemini -- but for complex tasks requires riding them very, very hard and really understanding where things can go wrong and poking at them repeatedly. Iterate, iterate, iterate.
> Iterate, iterate, iterate.
That describes my last week. What made it most annoying, was the need to release through TestFlight, because the memory issues would not appear, when tethered. Also, I was checking in constantly, because I had to revert and reset the context, several times.
Every line of code is a liability
Yup. In another post, I was grumping about having to accept a truly obese bunch of code from an LLM. This particular issue is the only one I could imagine even considering accepting that much pasta, but I sort of have to, because the LLM was able to quickly solve a problem that would have taken me a couple more weeks to address.
I remember a .sig that went something along the lines of:
now it's closer to 95% of work can be done by AI and requires 5% mental effort, but 5% of the work requires 95% of the mental effort to finish because of all the unoptimial decisions AI has taken. I find that AI works best in small micro-service type architecture where each component has a clear goal and doesn't have interconnected parts within the same application that can break. But you do run into an issue where changes in microservice a need changes in microservice b and updating it is not ideal since it usually cascades thru the entire system or requires stacks of legacy support.
IME it’s possible to have good clear APIs, limited scopes/goals, etc in a normal (macro?) service. But it requires a level of discipline and process many teams are unwilling to engage in.
There are a lot of bad CEOs, though. It's a lot like a politician -- it's quite difficult to become a CEO, and the skills to make it to that position don't always intersect nicely with the skills necessary to actually do the job well.
What skills? I've met several.
Most of them got into a prestigious school on legacy, paid for by wealthy parents. Many were above average IQ, but by no means geniuses. They had access to computers earlier than others, due to said affluence. They seem unable or unwilling to comprehend they're overwhelmingly on average, "nepo babies" to steal a term from the world of entertainment.
CEOs do get there with lots of politics in almost all cases. It’s all about who’s ass you kiss and who’s ass you don’t and if you’re lucky with timing things might just fall into place.
I think it’s exceedingly rare that a CEO is actually competent at their job. In most cases it’s the labor class propping the company up, and in some cases the workers are doing so against the wishes of the CEO. Not that executives want to ruin the company, they’re just incompetent and therefore make terrible decisions constantly.
Knowing which ass to kiss at the right time is an important skill not everyone has.
Kissing ass: $1
Knowing which ass(es) to kiss when: $9,999,999
And that's how CEOs justify their exorbitant compensation
You’re making the case for worker-owned cooperatives. Love it — we need more of them!
I'm very sympathetic to cooperatives, have traveled/know the Mondragon people (largest coop federation), etc.
However, I think there's a reason why coops seem to succeed at smaller scales, but there are essentially no large innovative coops.
There are a few large boring coops, and some small innovative ones, but seemingly something is making the CEO/investor board model the one large innovative companies are all using.
I suspect that it's both (1) access to capital is far harder for coops, and (2) that workplace democracy and hardcore mission focus aren't fully compatible. That is, "you cannot serve two masters" without losing focus on one of them.
If a company accumulates capital, it becomes vulnerable to the principal agent problem, and coops are way more vulnerable here than centralized companies.
If a company doesn't accumulate capital, it doesn't scale in complexity. It can grow by having more people do more of the same things, but it can't move into markets that demand anything complex.
What are the concrete benefits?
Do they tend to make greater revenue or profits? Pay higher wages and offer greater benefits to employees?
The problem is that knowing the right people to get investment does seem to have utility coops struggle to get, I think? maybe CEOs are basically like producers on movies who are just there to network for you.
In software I can imagine a worker-owned consultancy, but not a product company. It would imply staying in one place working on one product for your whole life, which doesn't sound inspiring
A company need not be a single product, and working in a worker-owned cooperative need not be a lifetime commitment to a single firm (though cooperatives ideally will have less turnover than firms owned by capital separated from labor.)
No, it implies that you give workers the means to dictate the direction of the company. That is what workplace democracy does:
https://en.wikipedia.org/wiki/Workplace_democracy
Acting like centrally planned dictatorships is a good form of collaboration is just so off base. There's no reason to think that introducing democracy into the work place wouldn't immediately benefit both workers + customers.
If this sounds crazy the C suite + board already vote on who gets hired into the executive team, vote for the direction of the company, and vote for their compensation packages (hint, they never decrease them).
Why shouldn't workers be legally enabled to do the same? What is the justification to this? I'm curious to hear it because the only way people can justify the current system is declaring that some people are actually more deserving of prestige, money, and benefits while others deserve to suffer.
With income inequality increasing, healthcare outcomes worsening, and children literally becoming stupider isn't it time to question the current system and ask ourselves if this is the society we truly want?
Where’s your real world evidence of all these benefits of coops?
Because I would love for it to be true.
Well workplace democracy has only been tried in a few corporations. If you want an interesting business case look into Semco Partners in Brazil:
https://en.wikipedia.org/wiki/Ricardo_Semler#Semco_1990%E2%8...
There is academic research on this too if you're curious but it's mostly in English, Spanish, and Portuguese.
But yes, there isn't much "evidence" because this system hasn't been tried en masse; however if you look at our current neoliberal hellscape, it's pretty hard to imagine it doing worse. Also neoliberalism wasn't really "tried" either, it was thrusted upon us by a group of individuals that wanted it.
One thing to keep in mind is that society can change quite quickly if you want it to. I'm sure the children that died working in factories during the 1800s never imagined such a society where children are valued, cared for, educated, and protected but it did happen.
It has happen before and it can happen again. It only happened because people were willing to fight for it.
The rules are allowed to be changed at anytime if we deem so, a better world is possible.
“We are sure it will work because it’s never been tried!”
I believe the Germans have had success with including labor representatives on corporate boards. Maybe we can start there.
Don’t you still need someone to make high level decisions?
Worker owned cooperatives have a variety of ways of doing this. Voting directly, electing people, etc. The main difference is that the cooperative typically doesn't buy the myth that the person making the high level decision needs to be paid 1000x the workers.
Maybe, but not necessarily for this reason. Even in a worker-owned coop, someone sets the overall direction. And how is that person going to be selected? It's still going to be largely politics.
Politics among the workers vs politics among the shareholders.
> it's quite difficult to become a CEO
on the contrary, it seems to be one of the few jobs that seems to require absolutely no qualifications to have.
What you need to do to be CEO is.... convince someone to lend you money in the hope that you'll get it back to them.
I've worked under some absolutely awful people who wouldn't pass an interview anywhere, but somehow they're CEOs, because they can smarm there way into more money consistently.
>convince someone to lend you money in the hope that you'll get it back to them
And it should be noted that many of these people lending money are in a similar situation of not being required to have any qualifications. Sure, some of them have worked their way up through sound investment after sound investment, but many of them were either born into their position or simply got lucky at some point along the way. Just think of all the money investors threw away pursuing crypto and NFTs for example. Many of those investments were transparently stupid from day 1.
To be fair, raising money takes a certain skill, that few people possess; and in many cases, it’s essential for a startup to even exist.
Often, they are good at taking things, keeping things, misdirecting and setting boundaries (especially communication boundaries). They are good at keeping their positions.
This is a broad range of skills and to actually be a CEO, you need to really hone these skills and be among the very top. To be good at those, just enough to qualify for a modest CEO role at a small start-up, you generally don't have the time to be good at anything else.
Saying that you don't need any skills is mischaracterizing it. You don't need any value-creating skills, yes, but you need significant value-capturing skills.
I can imagine a world were all companies become empty of workers and only executives remain and they would just have meetings with each other while they starve and would explain it away as a new diet they're on. There would be no petrol and they would be forced to walk to work and would say that it's their new fitness routine... And they would all believe each other.
It's not difficult at all to become one, and the work involved in being a CEO is not particularly difficult in comparison to senior technical work at all. The only thing that is harder about being a CEO is the responsibility. I'm sure being the CEO of Microsoft or whatever is plenty difficult and demanding in many ways, but most CEOs are not that, and speaking just from experience most CEOs and CTOs are clueless morons.
With that said, I've been programming for 25 years and I've only been a CEO for 3, so take what I said with a pinch of salt.
I do think people overestimate titles like this a lot, though, and it really comes down to what the company actually does and what is demanding for that company at that position/role. The CTO of a some-bullshit-as-a-service company may as well be straight out of college, because they're likely doing something trivial that literally anyone (including LLMs) could put together. The CTO of a well-used and reliable streaming service that handles a meaningful part of the world's Internet traffic is obviously solving a more interesting and demanding problem, and their decisions are going to be more important.
what about zukerberg he didnt have to do any politics to get to ceo. yet he is the face of ai layoffs and bad ceo.
What about him?
doesn't meanThe Winkelvoss twins would beg to differ.
He's the face of bad CEOs because people like to make up things about how bad he is. The Social Network, a major 2010 biopic about the early days of Facebook, famously cut his college sweetheart and now-wife out of the story in favor of a fabricated character arc involving an ex-girlfriend who does not exist.
Poor Marky Z. always getting a bad rap. Let s/he who hasn't made zillons getting billions of people hopelessly addicted to social media while facilitating a genocide or two and generally destroying the world order as we know it throw the first stone.
From your tone, it sounds like you may just be intending to warn me that it's cringe not to agree with any criticism of Zuckerberg? If that's so, I have to respectfully disagree; I think this is a bad attitude that leads towards being poorly informed about the world.
If you're interested in discussing the specific claims you're making, I really don't think that billions of people are hopelessly addicted to social media, and I would love to hear your basis for claiming this. My understanding (from e.g. https://www.nature.com/articles/s41598-025-27053-2) is that genuine compulsive use of social media is quite rare, and most people who describe themselves as "addicted" are just regular users who enjoy it but kinda feel like it's a waste of time.
https://gizmodo.com/meta-settles-lawsuit-that-claimed-social...
> Meta Settles Lawsuit That Claimed Social Media Addiction Screwed Up Schools
> On Thursday, Meta settled a lawsuit brought by a Kentucky school district that claimed the tech giant’s social media platforms have created a mental health crisis at its schools.
> The case was considered the first of its kind and a bellwether (a case that is representative of a large pool of lawsuits and will be a test for future litigation). The plaintiffs argue that social media platforms have had a major negative impact on the mental health of school-age children, which in turn has caused a burden on the education system, as American schools were forced to redirect resources to counter this problem.
> The settlement comes shortly after Meta lost a key bellwether social media addiction trial. Back in March, a judge in Los Angeles ruled that Meta was liable for the adverse mental health effects a now 20-year-old suffered after getting addicted to Instagram from an early age. The representatives of the young woman argued successfully that it was Meta’s deliberate design choices, like the infinite scroll and face-altering filters on stories, that had exacerbated her addiction and subsequent mental health issues like self-harm and depression.
---
https://en.wikipedia.org/wiki/Careless_People#Myanmar_genoci...
> The military junta in Myanmar was facilitated by Facebook to post hate speech that sought to foment sexual violence and promote genocide against the Rohingya. "Myanmar would have been a better place if Facebook had not arrived" Wynn-Williams writes.
> Wynn-Williams argued that Facebook failed to moderate hate speech against the Rohingya in Myanmar, including the use of the racial slur kalar. She noted that the company only had two Burmese language moderators, both based in Dublin, for the entire country, and claimed that one of the two moderators gave a pass to hate speech while removing pro-human rights content. She further claimed that she raised concerns that the moderator was "in cahoots with the" junta, only to have her concerns dismissed by the content team. Additionally, she claimed that her efforts to have Facebook's Community Standards rules translated into the Burmese language were resisted by the company communications team, who told her that "Myanmar isn’t a priority country" in the region.
---
https://en.wikipedia.org/wiki/Facebook%E2%80%93Cambridge_Ana...
> In the 2010s, personal data belonging to millions of Facebook users was collected by British consulting firm Cambridge Analytica for political advertising without informed consent.
> The data was collected through an app called "This Is Your Digital Life", developed by data scientist Aleksandr Kogan and his company Global Science Research in 2013. The app consisted of a series of questions to build psychological profiles on users, and collected the personal data of the users' Facebook friends via Facebook's Open Graph platform.[2] The app harvested the data of up to 87 million Facebook profiles. Cambridge Analytica used the data to analytically assist the 2016 presidential campaigns of Ted Cruz and Donald Trump.
> Other advertising agencies have been implementing various forms of psychological targeting for years and Facebook had patented a similar technology in 2012.
---
https://www.the-independent.com/tech/facebook-manipulated-us...
> Facebook manipulated the emotions of hundreds of thousands of its users, and found that they would pass on happy or sad emotions, it has said. The experiment, for which researchers did not gain specific consent, has provoked criticism from users with privacy and ethical concerns.
> For one week in 2012, Facebook skewed nearly 700,000 users’ news feeds to either be happier or sadder than normal. The experiment found that after the experiment was over users tended to post positive or negative comments according to the skew that was given to their news feed.
> The research has provoked distress because of the manipulation involved.
> Studies of real world networks show that what the researchers call ‘emotional contagion’ can be transferred through networks. But researchers say that the study is the first evidence that the effect can happen without direct interaction or nonverbal clues.
> Anyone who used the English version of Facebook automatically qualified for the experiment, the results of which were published earlier this month. Researchers analysed the words used in posts to automatically decide whether they were likely to be positive or negative, and shifted them up or down according to which group users fell into.
Yeah and Justin Timberlake was a co star. You’re really beefing about a movie almost 2 decades old?
It's on my mind because they're releasing a sequel in October, and Aaron Sorkin has not (as far as I've seen) acknowledged that much of the original was not true nor promised to be more accurate this time around. I'm pretty confident that it's going to be about the same mix as last time, and I'm going to have to go around saying "actually the scene where Zuckerberg did suchandsuch terrible thing wasn't real", and people will respond by insinuating that it's lame for me to care.
CEOs understands that AI offers potential productivity increases. Using that productivity boost to cut staff is an unimaginative approach. Bolder approaches include using that boost to exceed the expectations of current customers, or to increase sales without proportional increase in staff, etc.
I think sooner rather than later, AI will replace CEOs.
A custom-built AI would be pretty good at replacing a CEO. Think of all the things a company could do if they reduced overhead by that much?
You don't need custom built anything, ChatGPT could generate corporate initiatives and PR statements all day and no one would notice.
100% been saying this for a while now. the main thing AI will be able to replace is a C suite.
I think it will handle communication between the C suit and the underlings.
maybe this is the way forward. Imagine how many tokens one could burn given C suite salaries.
it's kind of like saying like you could replace a king with an AI. the position is a relationship to power more than it is a function with a productive output.
a bad king and a bad CEO could be replaced with a spinning top with no loss in productivity (and maybe some gain).
I still remember when I used ChatGPT the first time to write an email. I thought to myself “Oh. This sounds like 99% of the corporate communication from above”. We were joking that corporate had BossGPT for years and just didn’t tell anybody.
It also made me realize that most the so called “creatives” in marketing and PR also just repeat variations of the same few templates. Not much real creativity there.
The hallucinations would be a feature in this case.
AI means mediocre. Mediocre companies are pretty bad to work in. So it would guarantee a soulless and pointless company imo
My charitable word would be average. AI means average. And an average CEO isn’t that bad. It’d be a deliberate trade in exchange for less spending on salaries. (just one very large salary)
https://replaceyourboss.ai
(discussed 6 months ago https://news.ycombinator.com/item?id=46072002)
I am building a project now and then will create an AI to manage it and be the CEO.
The code is human + AI, the management is only AI
You can get a lot of tokens for a CEO. I'd say it's worth a try.
Robotics aren't there yet, it needs to go on golf playdates with investors and board members.
Maybe a virtual / Zoom agent could allow the elderly investor to stay in the game. When you get too old to go golfing you can still stay in the game.
Who takes responsibility when the AI does something unethical or illegal? Do we put the computer in jail? Or do we just look the other way like we do with human CEOs?
I think Grok on x.com is a reasonable case study. There’s the stuff we heard about and then there’s a separate category of things that weren’t newsworthy, and only the newsworthy things prompted changes to be made.
It sounds like you already know the answer.
To be fair here, CEO dont take responsibility and dont go to jail. In rare case they do, they will be pardoned.
Remember the printer in Office Space?
lol CEOs in jail.
While I agree that AIs would do a good job…
Would you rather take instructions from a ruthless robot or ruthless flesh sack?
Weirdly enough, I'd take the robot. At least we can pretend the robot doesn't know any better. The human is actively choosing to be a dick and profiting off it.
The flesh sack is choosing to be an insufferable twat, and the robot either doesn't have any choice or has a decent statistical justification for what it does.
There are a LOT of bad CEOs.
There are also a LOT of bad software developers.
When they meet, the software developer is fired.
The CEO exits after a while, after exercising their stock options...
Why can't we get AI models that replace these CEOs? I bet they're pretty good at running a company.
If AI makes you more capable, it’s basically like having a capital injection.
CEOs that look at that and think they need to reduce headcount seem to also be signaling they don’t know what to do with increased resources
What about employees who think AI replaces their CEO
Upper management material written all over them.
>> Yes, the tools are powerful, but a CEO who thinks they replace the work of employees is simply a bad CEO.
This is a broad generalization of employees. There will be some "routine tasks" that can be done by AI, now that is a lot more powerful.
There won't be as many employees needed for routine work - for example L1 and L2 support work. For example, many companies had ML engineers building models for various models. Companies can get that off the shelf from AI companies. They don't need a big team of model builders now.
If L2 support work is the example then I doubt we’re near replacement.
L2 issues are already involved in some way often revealing some kind of system failure, requiring context and exploration to understand, and judgement (and perhaps even system overrides) to fix.
I could see “automated L2 is the new L1” improvements, but without a big capability jump and/or a resource bonfire, I don’t think even frontier models would effectively replace good L2 staff.
They might magnify good L2 staff so fewer are needed (and maybe even help L1 staff become L2).
I often use AI chatbot to generate step-by-step instruction for setting or repairing up some bit of tech. It's incredibly empowering, and saved me a a lot of money that would have been spent on buying replacement tech.
You know who can't do that? People who call L1 support.
Wait, tech CEOs don’t understand why employees are valuable?
Astronaut holding up gun to other astronaut
Always have been.
It’s hilarious to me that when you stop investing in juniors and seniors who use your AIs retire, what are they going to do then?
You bring in contractors.
Ideally contractors that benefit you personally (eg: your buddy who now owes you one), but definitely contractors that let you outsource the responsibility.
Even better if you get some management consultant to suggest the idea and/or do the subcontracting.
Definitely buys you a few quarters of bonus and some time to land your next gig.
Who wants to be CEO for that long? The company will be sold off to a larger conglomerate long before that happens.
Get your state governor to go on tv and ask for people who know how to program Cobalt.
Wow the token leaderboard idea is nuts. It's similar to trying to measure the productivity of software engineers based on number of lines of code.
The message isn't subtle, and isn't meant to be: "we don't care how, but we expect you to stick your nose into AI tools and find some way to fit them into your workflow".
Which indicates: the management believes there are productivity gains from AI use, but adoption lags due to inertia and reluctance to change existing workflows.
> Which indicates: the management believes there are productivity gains from AI use, but adoption lags due to inertia and reluctance to change existing workflows.
Methinks adoption lags due to management's inability to align incentives such that productivity gains are rewarded.
more of "we whined and cried and screamed that we needed new budget in order to buy these tools or we would literally die. now we have them, they don't work as well as we hoped, they aren't leading to productivity gains, and they're actively alienating our workforce and users alike. we're so screwed we literally have no idea how to do reverse this."
CEOs who think that excavators replace their hand-diggers are just bad CEOs
If the excavators only worked half the time, sometimes goes out of control bonking someone in the head and costs a billion dollars, then yeah they're pretty bad CEOs
A story from today - I needed a small utility to remap my logitech buttons under windows without installing their horrendous GHUB. Logitech Onboard Memory Manager still required ghub to be put into onboard mode.
The solution - linux has utility called piper. I downloaded the repo and just told codex - figure out what piper is doing and create me a small utility to do it under windows. So the jolly critter started experimenting with hex commands, then pulled some other repo on which piper depended figured out how to enable said onboard mode and 10-sh minutes later I had small python script that did what i needed to do.
That would have taken probably half a day of work for a human.
There are many stupid CEO and organizations which are not committed to quality. And a lot of employees that are too set in their ways. But the instinct that underinvesting in AI is more dangerous than overinvesting is right. Doomed if you do, doomed if you don't
"To err is human, but to really foul things up requires a computer" - this is from the 60, but right now is turned into overdrive.
This is a great article, and I agree with most of it.
The problem is that the wrong eyes are seeing it.
We need these kinds of articles to be published in places that executives read, and tailored to their audience.
AI recovery is going to be a big wave of consulting over the next several years, maybe very publicly or maybe quietly, but it's going to be a thing. That doesn't mean "all AI is bad" or any other such nonsense, it means that there are a lot of companies out there right now that are doing it wrong and will need help.
The executives that get ahead of this are going to be the winners.
It's our fault for stupidly naming everything AI:
A* search -> AI
Backtracking -> AI
Neural Networks -> AI
Fuzzy Logic -> AI
Genetic Algorithm -> AI
Deep Learning -> AI
Generative "AI" -> AI
Similar to Tesla naming it's driver assistant "auto-pilot" in 2015 and your average Joe thought he would be able to sleep while the car would drive him to work.
The CEO just hear AI and think of AGI. They expect Skynet.
Aircraft autopilot is over 100 years old, and Testla autopilot is an automotive version that fits the definition by analogy to what aircaft autopilot does and what the human metaphor means (doing simple tasks aithout higher-level cognition or handling of surprising stimuli). Autopilot is not end-to-end driverless.
> the definition by analogy to what aircaft autopilot does
It's a bad analogy. Autopilot just maintains the aircraft in some state, then there's the flight director which maintains the flight path, and you can connect or disconnect the two at will. When connected the director can change the autopilot state.
To use the flight director you must fully specify your flight. The weight, the fuel, the weather, expected winds, takeoff and landing runway length, runway conditions, expected brake demand, as well as every single waypoint you're going to cross and the expected state at that crossing.
> what the human metaphor means
We learned after high levels of cockpit automation that maintaining situational awareness was still required. Pilots are freed of some stress during high workload portions of the flight, provided they planned correctly in advance, and that zero changes to their flight plan (not likely) occur.
As a result pilots are mostly told and trained to hand fly the plane during take offs and landings if the weather allows for it. You should only use high levels of automation if the situation demands it.
The point isn’t whether the aviation term "autopilot" has a technically defensible meaning, but that "Autopilot" predictably made normal people think "the car drives itself", just like "AI" now makes non-technical CEOs think "AGI employee replacement machine".
You are arguing the dictionary while I’m arguing the predictable, if not very well calculated and purposeful, misunderstanding.
The CEOs that think AI replaces their employees are the same that at the same time don't want to pay the AI costs.
CEOs are probably the most replaceable position, period, by human or AI. Everyone just gives them information. They don't know any information themselves.
Problem is, a CEO can fire employees, find out it was a dumb decision, then leave with a million dollar severance package. So they don't really care when they're wrong.
The primary product of AI is labor displacement and consequently wage supression. This is what OpenAI and Anthropic are really selling. It didn't start with AI but AI is accelerating it.
This is what layoffs have been about since the pandemic. People in fear of losing their jobs do extra unpaid work and aren't asking for raises. The theoretical potential of AI gives companies the excuse to fire more people. The investment itself is directly used as a reason of why they need to cut back on labor.
Any sufficiently sized business can only feed the insatiable hunger for ever-increasing profits ultimately by cutting costs and raising prices. And what do we have now? High inflation and a decline in real wages. CEOs are just following this playbook.
And the result is that society is bouldering towards collapse. We're seeing the first hints of this with the youth unemployment crisis [1][2][3].
Also, who is going to buy anything when nobody has any money?
[1]: https://www.americanprogress.org/article/americas-10-million...
[2]: https://www.brookings.edu/articles/twelve-ways-to-fix-the-yo...
[3]: https://www.bbc.com/news/articles/cy026x9jpd0o
> Also, who is going to buy anything when nobody has any money?
This assumes that a mass consumer economy is necessary, when it isn't. Mass consumption is relatively new, for most of history economies functioned with just a small consuming elite and large underclass that consumed very little. We are already approaching that again in the states given that the top 10% of earners are already responsible for nearly half of all consumer spending.
There's a floor even in a mostly automated economy where some services are resistant to automation simply because the human element is the product. Luxury hospitality, personal care, etc. That billionaire is going to want a human masseuse, not a robot.
A highly automated economy could stabilize like this with a small elite population consuming luxury goods & services, served by a low-wage economic underclass human workforce.
Its certainly not a pleasant society, but its also not unsustainable given enough oppression or pacification (bread and circuses anyone?)
Care to share some sources that corroborate any of these claims and explains how the dots connect?
Consumption by the top 10% of earners: https://www.wsj.com/economy/consumers/us-economy-strength-ri...
As for services increasingly unaffordable by the workers providing them, that's Baumol's cost disease (https://en.wikipedia.org/wiki/Baumol_effect)
I don't have any sources regarding someone preferring humans for certain services over robots, just intuition there, and the fact that consuming human labor and time is itself a status signal of wealth, and the current growth of personal services.
As for connecting the dots, look at Brazil, one of the world's most unequal economies having a small consuming elite and a much, much larger low-wage service underclass. The gulf states as well. Granted, their circumstances don't map cleanly to a post-automation Western economy, but it does demonstrate that a largely bifurcated consumption based economy can exist and can be mildly stable.
Whether the US falls into that direction too will depending on politics. A bunch of mid-career knowledge workers aren't going to willingly to flip burgers in a service economy without some serious surveillance and oppression. But when thinking about it in those terms, the recent push for mass surveillance laws and tech along with the increasingly dangerous rhetoric around protests and "domestic terrorism" start to make sense.
it necessary for ceos doing layoffs. i thought thats context we are talking about.
Previously:
Tech CEOs are apparently suffering from AI psychosis
https://news.ycombinator.com/item?id=48295679
I believe there are entire companies right now under AI psychosis
https://news.ycombinator.com/item?id=48153379
So much of this hype feels like astroturf in preparation for the upcoming IPOs:
https://tomtunguz.com/spacex-openai-anthropic-ipo-2026/
and I don't know what worries me more - a burst in this bubble (and maybe some other tech stocks), or a failure of these valuations to be burst somehow, and even more concentration of capital and power around those corporations.
Most CEOs are not special. They are not especially smart, or skilled, or technical. The role self selects for sociopathy. That's not a quality that has any kind of linear relationship with intelligence. Quite the contrary.
A common misconception about AI is that it is intended to fully replace humans, which is incorrect. The purpose of AI is to reduce the need for human labor, and it has already been doing so. For example—though this is not an exact figure—a task that previously required 15 people might now only need 10. In no instance has the human element been completely replaced; rather, the reliance on manual labor has simply been reduced.
It's not exactly a misconception, when companies are pitching AI as a full and complete replacement for human employees. People are just reading the billboards on the side of the road.
I always found advertisements for AI to be so strange, why would you advertise your AI to the public as a danger for humanity that will also put everyone out of work? Such advertising would only appeal to sociopaths, but of course that's because it's intended to appeal to CEOs.
> though this is not an exact figure
You mean, this is an entirely made-up figure.
So what if it is? The example still stands.
A "unit of work" that required X people to complete in Y time can now be done by X/Z people in Y time, where Z is whatever efficiency you are able to get out of applying AI tooling to your business.
For some companies, Z might be less than 1 though. ;-)
So you still need skilled people, just not the same amount as before, because you have different tools available to you.
This has happened before with other advancements in industrial/technological automation. It's not a new concept.
That supposes that AI has a positive impact to efficiency. So far I see the exact contrary, at least for software engineering.
That sounds like 5 humans got replaced by AI. I don't think most people worry about whether all humans will be replaced, simply whether or not they will be replaced, or people they care about.
Or all 15 people are still employed doing the work of 22.5. Or even more people have been hired now that each person can generate 50% more value than they previously could. Or people are reallocated from the AI assisted task to another. Or some combination thereof.
Which is very short sighted. You or anyone close to you might not be replaced but it should be clear that you don't want to live in a society with 20% unemployment.
I think Spain had unemployment in that range for many years.