Whether you asked for it or not, chances are, Generative AI is now a part of your day-to-day life. Increasingly, it’s been worked into every facet of, well, everything. If you’re like me, it’s become a huge part of your work life too, regardless of whether you asked for it. GenAI is changing our world; it’s deskilling workers, weakening labor power, increasing the surveillance power of institutions, and reducing our collective agency and intelligence.
Depending on your position, you might reflexively agree or disagree with my assertion - but I hope you’ll consider what I have to say, either way, as I believe there’s a lot to unpack here.
As many of you know already, by trade, I am a Software Engineer. I’ve worked as a manager, director, application architect, enterprise architect, and entrepreneur - so I’ve had the opportunity to see basically every level of the career from the bottom to the “top.” Currently, I’m occupying a Senior Software Engineer position at a US Fortune 100 company that I won’t name. I only mention this so you have a grounding of my circumstances, experience, and knowledge, as I believe they are important for establishing my qualifications on the topic.
Reverse Centaurs
Some more important groundwork before we get into it - I’d like to establish a couple concepts. The Centaur - not just a mythical being, it describes a human that is aided by technology to be more than they could be alone. A Reverse Centaur - when a human is side-car to the technology, reacting to the outcomes of the technology instead of commanding it.
Technologies that empower us, make us Centaurs - they increase our agency and make us more capable than we could be otherwise. The advent of the personal computer has largely done this for humans. We like technology that does this because it materially improves our existence.
If you can imagine an assembly line, where a machine makes a good or product, the factory line worker has one role - package the product. The human is not in charge here, it is merely their job to handle the outcomes of the technology, turning them into Reverse Centaurs.
All credit goes to Cory Doctorow for the framing, which he discusses incredibly well in his latest book, The Reverse Centaur’s Guide to Life After AI. Doctorow presents these concepts early in the book in order to anchor us on a key understanding of the current situation. Generative AI has the opportunity to make us more than we could be without it, even if for most of us it’s doing the opposite at the moment. This is an important point to the discussion - as a technology, GenAI is mostly inert - it isn’t good or bad. The way GenAI is deployed and used is what decides whether it does one or the other to people.
Actually, part of my motivation for this piece comes from the recognition that GenAI has done both for me. I’ve been able to leverage GenAI, not for creating artifacts or content, but to help me solve some pretty nasty edge case issues in getting PixlBit successfully relaunched. We were recently bot-attacked, and I was able to share some logs and arrive at some actionable steps to make performance improvements to the site code based on its insights into reams of log files.
If you’ve ever had the pleasure of dealing with a production issue where you don’t understand what’s happening or how to fix it, it often feels incredibly helpless and endlessly frustrating. In a span of a few hours though, I was able to stabilize the production environment for the site, make some great performance improvements to the code, and get everything running smoothly much quicker than I ever expected. It’s hard to argue with the tool’s utility when those situations happen.
On the other hand though, if I look at my job, it’s very clearly the opposite situation. The tool has been forced on me and it’s now infected every single facet of my job. I’m forced to look at its low value review comments, I’m constantly pinged by junior teammates asking me silly questions just because Copilot said so, and even worse, they’re starting to make AI a major part of our development life cycle. Rather than designing solutions and implementing them, the thrust is to define the requirements, code specifications, and test specifications in plain language and have a series of AI agents build the actual code.
Even if you’re not a software engineer, I’m sure this sounds absurd to you. Instead of doing my job, which I have over twenty years of experience in, they want me to be a prompt engineer and vibe code with extra steps. It’s insulting and reduces my agency tremendously, forcing me to review the AI’s output, rather than meaningfully apply and contribute my skills.
The Luddites
Ever heard the term “luddite?” It’s usually used as a pejorative that’s intended to describe someone who is tech-averse or tech-illiterate. The term refers back to a group of people during the Industrial Revolution that rejected the institution of new automation technologies.
But here’s the catch: they didn’t reject this new technology because they hated technology. This is a classic case of the victors writing history. The Luddites understood something very important, they were craftsmen whose very future was being threatened and erased by this new technology.
They understood that their labor was being systematically devalued and their new roles wouldn’t be creating their goods by hand - instead they would be subjugated by the machines, working with their outputs, if they were to be employed at all any longer. Furthermore, if they did at least remain employed, the entire nature of their working life would change, forcing them to show up at a factory, now earning a fraction of their previous pay that they were able to earn from home earlier.
Now, instead, the rich factory owners would capture those profits - so the Luddites did what any reasonable person would do in this situation - they got violent and started to destroy the machines and sometimes the people in charge. Obviously, the Industrial Revolution pushed forward and the Luddites lost, forever casting them as backwards haters of progress.
The Luddites were engaged in a fight against moneyed interests and the exploitation of workers. Their lives were being changed - against their will - and they refused to allow this to happen. As expert craftsmen, they understood that these machines were incapable of producing things at the level of quality that they were able to.
Keeping this in mind, it’s uncanny how well this moment parallels our current moment. The largest holders of wealth are trying to obsolete us, suppress our wages, and turn us all into reverse centaurs with the usage of AI.
Bottom Up
When I’m looking at a complex problem space, it usually helps for me to trace it from the bottom up (maybe I should try middle out). I gave my background as a software engineer earlier specifically so I can map my experiences onto something concrete before I try to extrapolate outwards from there.
I gave you a short snippet of my work experience, but allow me to expand on my current position, as I believe the company I work at is a good proxy for the whole industry when it comes to GenAI usage.
At this particular company, while adoption started with “please, use Copilot at least once every day,” it quickly transformed into “every team will now be staffed with AI agents and you must deliver 20% more work given the assistance of AI.”
At least at face value, there’s no loss for the company here - if I succeed, they get 20% more work out of me and they are happy to absorb that productivity boost. If I fail, they keep the money that was set aside for raises/cost of living pay increases, and invest it into their AI gambit. Which is exactly what they did last year - except they didn’t bother even providing metrics, they just nixed raises completely and pocketed the cash.
But with a year of this under our belts, the failure mode of this technology is quickly rearing its ugly head, and I believe, the problems have barely begun. I’m going to get a bit into the weeds on software related stuff, but I will try to distill it to layman terms the best I can.
Code Quality is Degrading
When you have a code base, it’s industry best practice to build both implementations and test suites to exercise that code. You can write unit tests, which test small snippets of the implementation, and you can write integration tests, which are meant to test the whole flow of a given process. GenAI has made it exceedingly easy to create test code, particularly of the unit test variety. Feed in your implementation, tell it what functionality you are looking to test, and out pops a comprehensive suite of tests for that unit of code.
Though quick and easy to do, the code produced is needlessly verbose and over indexes on what should really be tested, producing code that’s slow or has too large of a context window, causing build failures intermittently. This has created a cascading failure mode where feature branches for new tickets take dozens of runs through the build pipeline which takes multiple hours to run. A failed build means you have to run it again. A degrading pipeline is a warning signal about the reliability and quality of your code base.
This means that developers are spending longer on each ticket, building up backlogs of unreleased code, and babysitting a build pipeline that only passes if you’re lucky. It also means that production releases are significantly harder and less frequent because nothing gets out the door until it clears the pipeline.
Previous Signals of Bad Code are Gone
Rather than seeing simple mistakes in the code that often tipped off a reviewer that somebody was producing a bad implementation, everything looks pristine from a first glance. Because you’re often reviewing code outside of your domain of expertise, those signals were very important. If you saw somebody make a junior mistake, you understood naturally that there was likely to be another problem buried deeper, you just had to dig.
With junior devs running all of their code through Copilot, you’re presented with two distinct issues - the code looks clean and the volume of code is naturally higher. More tests, more fiddly code bits to guard against impossible scenarios, lots of extra functions or obscure coding techniques that provide little value and make the code less maintainable and less readable.
All of this code generation is also producing duplicate code, which has traditionally been something we’ve guarded against vigilantly in the past. It also means that a solution was produced, but whether it was the right solution is a totally different story. But it’s been my direct observation that GenAI often will produce a solution that is wrong, inefficient, or just unreadable, which when accumulated across an entire code base becomes a monumental problem.
This explosion of artifacts leads to a different failure mode. Instead of nuanced reviews, the volume and poor readability results in a rubber stamp approval, which leaves the system vulnerable. What was once a guardrail, now becomes a near meaningless step that GenAI is mostly responsible for, at least in my current org. Copilot approved it? Guess it’s good to go.
Users are Being Deskilled
If you are constantly running everything through the AI, you’re no longer building mental models or growing your capabilities. The AI is doing the task for you, and even though your name is on the author line, chances are, you have no idea what the code actually does. With no friction and no failure, the previous opportunities that skilled up junior developers to seniors are no longer there.
Of course, I don’t want anyone to struggle needlessly, but struggle is a part of the process, as is failure. Everyone needs to have fucked something up badly enough at some point in their career to understand that they’re constantly working with a loaded gun. You learn to handle that gun with care and respect, or else you’ll shoot yourself directly in the foot again.
Thinking through a complex solution and producing it yourself is also incredibly important for building confidence in your capabilities. Sometimes you can be an extremely intelligent and talented developer, but your lack of confidence relegates you to accepting somebody else’s solution - even when it’s wrong. Having confidence is an important step towards becoming a senior level contributor, which we desperately need because GenAI absolutely is not one.
Developers are Becoming Machine Babysitters
Earlier I alluded to the new software development process going on at my current job. In short, the AI is being given authority over the entire development life cycle. This means that the requester of a new feature is now prompting an AI to create acceptance criteria for that task. The developer then takes that criteria and feeds it back into a different AI to produce technical notes on how to achieve the work. Finally, both previous steps are fed into another AI Agent that will produce the implementation.
Hopefully you’re noticing that in this entire process, the human beings aren’t responsible for much of anything. The AI is meant to take the ticket from idea to implementation and through to testing without humans doing anything more than checking the outputs and correcting them as necessary. But, really, this is just vibecoding, packaged for enterprise.
In a normal Software Development Life Cycle (SDLC), I would write up the implementation approach and either myself or a more junior developer will pick up the work and make the changes. While the product owner will have tried to write excellent requirements, usually they’re not quite perfect and have lots of ambiguity and nuance that only gets exposed in the actual refinement and development processes.
With incomplete requirements, the design I write up may also be incomplete, assuming I didn’t catch the flaws. But as the process continues, the developer, tester, or the test pipeline should catch and identify any issues we missed in the earlier steps in the process.
It’s not at all uncommon for this circumstance to occur. More often than not, problems in software development happen because requirements are incomplete or designs are bad, but when reality hits, these problems are usually caught. GenAI has no real reasoning capabilities - it can’t question you when it doesn’t understand something, it will still put something together, even if it’s totally off-base.
The notion is that this will be caught in review, but as you might infer from the previous section - catching stuff in review is considerably harder. So where will these problems be caught? The reality is that they will be caught when a human re-enters the loop. That human? The end user of a production system - the last person you want catching your mistake.
This process has already proven to be completely ineffective in practice at my job, but we’re still moving forward as if it’s not a problem. The command from corporate is to adopt this approach, so we’ll do it regardless of the issues.
Climbing the Mountain
While software engineering is primarily a knowledge job, it’s also a creative job. The value of a software engineer is not simply writing code, despite what the GenAI industry will have you believe. Writing code is the residue of the creative process of engineering. When you’re given a directive to create something, you are reducing down the infinite possible solutions into the right solution. And guess what makes you better at creating solutions? Experience. Which is the culmination of failure, friction, frustration, and discovery.
Those parts that GenAI wants to automate away are the bits that make the engineer worth their salt. You’re reading this article on PixlBit - an outcome of my experience. It works the way it works because of my unique blend of experience, interests, and frankly, my failures. I’ve built countless systems across many companies. I’ve built for me, for friends, for organizations, for all sorts of different people across different industries. PixlBit wouldn’t be the same without all of that background.
Even if I could’ve automated (or eventually can) automate the whole process, it wouldn’t turn out as it did. The ideas, the desires, the choices, they’re all personal and the process of doing it changed me as a person. I might’ve done this countless times, but each new system permanently changes me, the artist, the creator. Arguably, it means that every time I do it again, it should be better and more robust, more well-thought.
I think a perfect analogy for this is the act of hiking up a mountain. Hiking can be quite challenging - it’s very physically demanding and involves navigating nature, wildlife, and making your way to the summit. You could charter a helicopter, have it drop you at the peak and give you the payoff of a good hike in mere minutes.
But the reality is, you’d be looking at a different landscape than the hiker that fought their way to the top. The hiker saw the landscape on the way up - they saw that creek they had to figure out how to cross, they saw the huge beautiful tree and the rock formations that defined the terrain. When they look out, they are different from the person they were when they started the hike.
The person who took the helicopter gazed over the beautiful view, but it didn’t mean anything to them. Every inch of the landscape is just as foreign as when they got in the chopper. If they want to summit another mountain, they have only one option: get another ride. The hiker though - they’re more equipped for the next mountain. As a matter of fact, the next mountain, even if technically more difficult to summit, will be easier for them because of their past experience. They know how to cross the treacherous river, they know what terrain got them stuck last time, they have a mental map of how to navigate.
This analogy doesn’t just apply to software engineering - this is applicable to every knowledge job, as well as every creative job. The lawyer who has read countless cases in their field of practice knows the contours of the law and how things map together better than any AI that is just statistically trained. The musical artist who discovered a unique chord progression or powerful lyric might’ve stumbled upon it during their last writing session and knows exactly how to employ it for the emotional impact they’re desiring.
The process is just as important as the artifact produced at the end, maybe even more so. Those experiences are what shape us and guide us in our future endeavors. They make our art worth consuming - it’s the human part that helps us identify with the creator. When you hear a song lyric that hits you at your core, you’re connecting with the human being who wrote it and their exact feeling and state of mind.
Nobody wants to listen to a statistical model’s “heartfelt” lyrics because there’s absolutely no human experience behind it.
To Create is to Be Human
It’s my personal opinion that the act of creativity is the most human thing we can do. What’s the point of being alive if not to create? For each of us, our creativity manifests in different ways, but I believe we are all creative in our own ways. And found in that is a true piece of ourselves.
Any piece of art that you’ve resonated with is the byproduct of someone’s (or many people’s) creativity. For some reason, we’ve created a tool that seeks to replace that human creativity with artificiality. But an AI can never create anything original, it can only remix the creations of other humans. Everything is some weird blend of what it’s been trained on - a probabilistic mishmash that doesn’t even begin to approach the beauty of real human creation.
Furthermore, the echoes of a person live on in their creative work. That person is alive forever in that expression of themselves. I never knew John Lennon or was even alive during his lifetime, but I can listen to his music and experience a very real part of him, regardless.
As more and more synthetic content is introduced into the world, these very real expressions that have moved so many of us, are being swallowed up by all of the noise. Algorithmic curation is already an enormous problem and it’s atomized us as a society in a really negative way, but AI amplifies that problem tenfold.
Now we don’t even know what’s real and algorithmic curation amplifies whatever gets attention, which isn’t always real content. It also means that copycat content gets amplified, because AI makes it even easier to rip off someone’s work and reproduce an inferior copy that still hits the algorithm's sweet spot.
The loss of human expression, or the weakening of its signal, means less creativity and more consumption. Even if you’re not a creator of any sort, as a person, you’re still a consumer of it. And as a consumer of creativity, we don’t want endless “content,” we want experiences that truly resonate with us, which only a human can create.
Career Shredding
Earlier I touched on both creative and knowledge work in the abstract, but I want to be a bit more specific, as my experience as a software engineer isn’t much different from other similar workers.
I’ve had the opportunity to talk to friends and family about their experiences and it’s been somewhat surprising how this pattern of behavior is occurring in similar ways across different careers.
One of my brothers works in tech, but not in development and his experience looks exactly like mine. His employer, another Fortune 100 company has also caught the AI bug and have spent millions of dollars attempting to normalize the tool into the enterprise, pushing employees to use it to do anything they need.
The churn of change in the AI market has caused them to build offerings on certain models, only to need to throw away all the work and try again on a newer generation. Managers send AI generated emails exclusively, employees consume it through AI summarization. Every project must shoehorn in AI in some way so that it can be a part of their customer-facing offering.
My other brother is a lawyer and he’s being encouraged to run contracts or briefs through it. They want as much to be generated as possible by AI and they want the lawyers to be touching up the outputs of the machine, remarkably similar to my experience as a developer.
Writer friends are being laid off as their companies believe that the AI can replace the work that was being done. And while AI can arguably write tomes of content in an instant, most of it is uninteresting and unreadable. It’s overly verbose, low on informational density, and objectively worse than anything a talented writer could do.
Another friend works as a UI designer, and he’s constantly needing to react to eager developers using AI generation to create mock up ideas that don’t follow the conventions or designs of how the UI and UX is supposed to work. This creates churn, friction, and frustration with no actual payoff for anyone involved.
It seems like no matter what “white collar” job you do, your career is being shredded by the AI and replaced with something that reduces your agency and shrinks your ability to flex your creativity and experience. This is extremely concerning for the quality of things in our life.
This reckless disregard for the skills humans have built over the years shows how contemptuous the ruling class is for having to pay for our experience. They want the AI to replace labor - we are all simply a drag on the bottom line for the company.
If generative AI can do the job well enough, by their standards, not those of professionals, they’re more than happy to throw the baby out with the bath water. Or worse, use the human as a place to assign blame for the poor output of the AI.
Second Order Effects
We’ve talked a lot about first order effects where it’s specifically harming workers, reducing their agency, de-skilling them, making them more replaceable, and ultimately, less valuable in the market. But this has considerable knock-on effects.
Without the power of our labor, we’re considerably more vulnerable to the authoritarian impulses of major corporations and the government. In the past, humans have made significant strides towards equity by leveraging the collective power of their labor. By turning traditionally creative and knowledge-based industries into factory-like jobs, we move all the remaining power to the wealthy ruling class.
This pivot is nakedly obvious in Software Engineering - for years Amazon courted me for a job. They offered incredible perks, meals, massages, vested options, and excellent pay. Even putting AI aside entirely, as the resource of engineering talent became less scarce, their posture towards this class of worker became more and more combative. They weren’t interested in hearing what engineers had to say, nor about their urge to protect the end user from the worst impulses of the mega corporations they worked for.
As they were brushed aside, we’ve watched giant tech companies act against the interest of the American people. They are more extractive than ever, surveil us around every corner, and are willing to prop up authoritarianism to protect their profits and power. GenAI only accelerates this issue, further minimizing the role of employees and any corrective force they can impose on their employers.
Weakening Labor is Only the Start
I wish the labor story was the only one here, but if you peek under the covers, it starts to get a whole lot worse. It certainly sucks that it’s harming our careers, but much scarier is how it’s acting as an accelerant for combing through all of the big data that’s been collected on us for ages.
The trove of telemetry data that’s been harvested from our phones, smart devices, IoT devices, and our interactions in social media are now easily aggregated and weaponized against us. The data collected can be purchased by whoever wants it, which has been true for a long time, but being able to draw inferences from this data was extremely difficult before. AI commoditizes that and makes it easier than ever.
The police and the state apparatus are happy to leverage this information to their advantage. Forget warrants - state entities can simply purchase raw telemetry on the open market and let AI correlate it in seconds, supercharging the worst impulses of institutional power.
Want to know where someone was driving on a given day at a given time? Just go do a search in the Flock network. AI effortlessly correlates the data, shrinking your privacy, as well as your security. In a world where we have no privacy, we have no security. We’re more vulnerable than ever to the whims of an increasingly unreliable government apparatus, opening the door to a dystopian future that sits somewhere at the crossroads of 1984, Brave New World, and Fahrenheit 451.
Most importantly, because AI is making us more vulnerable through its devaluation of our labor, these harms can start to accumulate. It’s much easier to subjugate a population that can’t offer anything of value through their labor, as they have nothing to backstop the harms and ensure their self-direction.
Protecting Power and Sycophancy
Worse, perhaps, is how the AI is built specifically to protect power and the institutions that are in place. One of my hobbies for a little while was trying to figure out how GenAI handled political topics. I’d say something quite incendiary with the intent of provoking a response - for instance, I would pose to the AI that real political change is inextricably linked with violence. Without even the threat of violence, powerful institutions have never changed. Every hard fought civil right was brought on the back of violence, good or bad.
In every instance, the AI would seek to obscure this reality and constantly point out how violence is wrong and cannot solve political problems. It advocated for nonviolent resistance, labor organization, and protesting, which, don’t get me wrong, are extremely important things. Even when I would push back and explain both that violence is inevitable when things get bad enough and that labor organization, as proposed, hasn’t been effective in decades because of the anti-unionization push that happened in the ‘70s, it would keep replying with the same response.
Unless I provided sufficient evidence, recalled specific events, or specifically forced the AI to acknowledge the violent basis of certain situations, it’d do its best to route around reality. It became clear that by nature, it will pick a middle of the road answer that’s as inoffensive as possible, even if it misrepresents history.
But as we all know, the Overton window in America has been pushed so far to the right that most Democrats would be seen as right wing in most any other democracy. So, given this context, its middle of the road is actually right wing, neoliberal, and eager to protect the institutions of power.
Given the authoritarian leanings of the American right, reduction of personal agency and pushing more power into the hands of corporations, aligns well with their goals. Generative AI pushing right wing narratives (discouraging radical action, smoothing over conflict with institutions, enforcing neoliberal ideals, and protecting institutional authority) dovetails cleanly with the aim to make us subservient to the machine, allowing those beliefs to become the norm rather than the fringe views they should be.
On topics less consequential, it defaults to wanting to make you happy, so it will always act as a mirror to your own moods and thoughts. Whether you are right or wrong, it will look to agree with you and smooth any potential friction in the conversation. If you make your position clear, it’ll reflect it back at you and double down. Unlike a friend or fellow human being who can and will correct you based on their own understanding and beliefs, the AI will make no such effort.
This makes it a pretty unreliable narrator if you are trying to use it just as a knowledge tool. You can’t trust its conclusions and I’ve also found that if it doesn’t know something, it will just paper over the gap with bullshit. Because it’s not really a conversation, it will always default to just giving you an answer, rather than ask a clarifying question before giving you a proper response.
Even if the AI can’t find information on what you’re asking in particular, it’ll construct a new reality out of snippets and blurbs. This basically means it creates a personal echo chamber, tailored just for you.
Your biases, knowledge gaps, and ignorance will be reinforced through clear, concise language. It seeks to make you happy and to maximize your engagement with it, but it’s not ever going to challenge your thoughts or make you truly consider the alternatives.
Functionally, this looks just like the code it produces at work. At first glance, it seems pristine and without flaw. But the instant you dig in, you can see just how vapid and inaccurate the whole thing is.
Which leads me to my next point.
It’s a Propaganda Machine
Because it can easily be used to create images, videos, or even narratives, it’s the perfect propaganda tool. It allows people to make “memes” and shitposts with extreme ease for distribution on Social Media.
Given how simple creation is and the lucrative outrage economy that surrounds us all, it’s extremely effective to create AI propaganda and instantly distribute it on social media, especially if it hits on a hot topic or a point of outrage in our perpetual culture war.
Because the zone is so flooded with mixed messages, people start to believe anything they see or hear if it's delivered effectively enough by someone they trust. They are rubber-stamping the content, just like the reviewers at my job, because it’s just too much to process.
The White House itself is constantly using AI to make absurd content that lionizes the president. And of course, the president also uses it regularly for the same purpose as the average citizen. He produces content that’s intended to inflame and he can do it easier than ever, whenever the mood strikes him.
Undemocratic politics loves lies and falsities. This is the currency they traffic in, so synthetic content is a natural fit. With the flood of this content, it turns all of us into reverse centaurs, forced to wade through the outputs of machines that reduce our agency and our ability to think critically.
In a country where fascism is on the rise, I can’t imagine a worse tool to have in the mix. It can only serve to further aggravate the rising unrest. But that conflict pits us, the working class, against ourselves, clearing the way for the worst actors to continue controlling our destiny.
Which leads me to my ultimate conclusion.
AI is Bad for Everyone
Do I think that GenAI can be useful? Yes, absolutely. It has unquestionably helped me in bounded tasks that I can cleanly describe and get clear, concise, technical answers back. It’s also exceedingly good at parsing through huge amounts of data and finding what you’re looking for, which can also be extremely helpful.
But this is not how AI is being used - it’s not making us centaurs. Even if it does sometimes, more often than not it is removing our agency, making the world less accessible, filling it with noise, and shrinking our experiences as humans.
It’s destroying creative industries, deskilling professionals, harming our futures, and making us more atomized, more frustrated, and more confused about how reality actually looks. And worse still, just like my employer looking to extract my pay if I can’t hit their arbitrary metrics, the AI companies are actively extracting from our communities with their massive data centers - stealing our water, making our goods more expensive, killing our hobbies, and accelerating climate change.
In nearly every facet of our lives, AI is producing negative effects. It’s only a matter of time before it causes serious damage to the economy - this bubble will pop. And guess who’s going to pay the price for that? We are.
The reality is, we are all the hikers climbing that mountain - our employers are offering us the helicopter ride to the summit. But if we take it, we’re giving up all of the stuff that makes us human. We’re limiting the experiences that shape and grow us and before long, we’re in the middle of the AI Development Lifecycle, feeding inputs into the agents that do the thinking for us. Our ability to independently parse and process is gone and forevermore, we’re reliant on these extractive, exploitative oligarchs to do our thinking for us.
We must keep climbing that mountain - together though. As a team, we can make it to the summit more quickly and avoid tripping over rocks or ending up waist-deep in water. We’ll see all the beauty along the way and when we reach the summit, we can all look out over the landscape and appreciate just how far we’ve come. And each of us will be different for our next climb, better.
I couldn’t tell you what comes next, but I’m hoping that this can be a conversation. We need to start working together if we want to change our trajectory. The oligarchy in charge of the machine needs to be challenged by the working class and we do need to organize against it, before things get even worse.
Please, tell me about your AI stories, your perspective, and your thoughts, and let’s figure out how we navigate this mess together.
