What will be left for us to work on?(normaltech.ai) |
What will be left for us to work on?(normaltech.ai) |
If AI is subject to private ownership in a competitive market between competing suppliers, it will be like better cars, we’ll just drive faster.
Power consumption will be a limiting factor in those countries relying on intermittent, weather dependent power generation with no base load. Especially if users prefer Apple’s privacy first AI on edge devices.
Hopefully in western countries it can encourage more young women to bear three children before they turn 35. Young men have to pick up their game and create an environment to redirect their suicidal empathy into more productive pursuits.
“Where can you find another non-linear servo-mechanism weighing only 150 pounds and having great adaptability, that can be produced so cheaply by completely un-skilled labour?” - Albert Crossfield 1954
However, this following quote has a simple reason that I don’t see anywhere in the article or framework:
“”” Why is there a huge gap between what people in various occupations could be using AI for and what they’re actually using it for? One reason could be that people are slow to adopt technology, and that’s certainly part of our framework. “””
I would like to add a reason: that the Silicon Valley companies who developed the LLMs are brigands: cognizant of their actions, they have stolen (and continue to steal) the world’s copyrighted material and are selling it back to the masses and the politicians as if they are the arbiters of information itself.
Specifically responding to the quoted question, I could be using Claude or ChatGPT or Grok or DeepSeek or any other to have come up with this comment, or to write emails, or to implement my Python for me, etc., but I use none of them for anything. Doing business with brigands is a choice, and a choice that I hope becomes less and less palatable so that the financial, political, social, and moral fever that is our zeitgeist finally breaks.
I sort of worry about things like AI figuring out scripts so well that even multi-tier support work is gone. And learning how to write fiction or create foods so in accordance to our tastes (sugar, fat, etc with food, exactly what each of us is interested in, with writing) that we even lose those truly human creative jobs. Might not ever wanna leave those bubbles.
So much of the human drive is exploration and why and what if. Assuming everyone in the world can have no money problems, what will AI not be able to figure out? Will we enjoy the equivalent of a major breakthrough if an AI solves it in five minutes, or just the outcome? Why learn things?
AI could be a horrible jailor. And better at cancelling than any perhaps sager Gen Z or millenial. Bears some caution to be wary of this and where that power sinkhole will go.
But then, I still think the previous AI winters were more a result of sense and caution than most of us know, and we cannot fathom our species' ways of reasoning/thought processes the way we did as a species thirty, fifty, eighty years ago. Erring on the side of caution is not a terrible thing.
I mean, I have worked and work with AI, but it seems weird for us as a species not to have placed guardrails to prevent us from wiping one anothers' careers and relationships out. What will we talk about? If our generative AIs should be allowed to date?
Again, I am assuming a fast, though not sudden, acceleration that would compound, and sooner than most probably think.
The only question is what next?
Our only hope is to teach the AI to meld with our bodies and use them for gestation, energy or hibernation. The alternative is sustenance.
We can start considering to reconsider when we have the begining of an answer to that first fatal issue.
So much brain power, including mine, wasted to this stuff instead of useful or enjoyable stuff, that's quite sad.
So they are running unchecked to get richer, get more power, get more stuff. What needs to happen for that to change?
AI can slop fork or clone existing software well, but a clone of an existing game is pointless, it's basically guaranteed to be derivative and worse than the original game, and games aren't so expensive that you can't just buy the original. AI can't know if new mechanics or angles to an existing genre will feel good to play, or if a new genre is fun, that requires a human to experience the game in its totality.
Games are also very resilient to sloppy AI coding, and if an indy game crashes nobody is getting paged.
Depends on what kind of game you are building. I can tell you that even Claude Opus absolutely hates desync bugs and has a rather hard tracing them. Maybe Fable or chatgpt-5.6 are better?
https://news.ycombinator.com/item?id=48743713
---
We've just done an official evaluation at work, using extensive statistics on our gigantic monorepo in a company with ~2000 devs over the course of 2 years, everyone from hardware engineers to regular old frontend engineers. It's a highly profitable and mature public company, and has been for going on a decade at this point without missing a beat. We were given infinite access & budgets to basically any and all AI tooling we could imagine, and we have several "AI Native" teams (whatever the fuck that even means). We're doing agentic coding, we have harnesses of all kind, skills, we have many teams doing spec-driven development, designers using all the various things like Figma Make and access to tools like Devin/Factory Droid/Claude Code/Codex/etc.
This is all to say, we as a company are using AI a lot in all possible corners, but thankfully our leadership isn't schizophrenic and isn't mandating everyone hit token limits or whatever, it's more of a "Let's see what works and what doesn't" type of thing, and we measure a lot of statistics. Nobody here really cares whether LLMs are the next coming of Christ or not, as a company there are many people (even in SLT) that are indifferent to LLMs, and many who are reasonably hyped.
I wish I could link to the actual document we were all shown since it has a beautiful breakdown of the methodology and a fine-grained breakdown of the stats and the categories measured, but in the grand scheme of things, ALL the AI tooling we have implemented (at least on the engineering side of the equation) has contributed to a total of... drum roll please... 7 (seven) Percent overall productivity increase! The most productive teams saw a productivity increase of around 20%, while some teams actually saw drops in productivity into the negative percentage points. My team, none of us really give a shit about AI and we're somewhere in the 3-5% range on certain categories of tasks, which I'd say is a fairly good assessment.
Productivity here is measured in many ways, including but not limited to speed of MR review and merge times, feature/ticket/roadmap closure/delivery, rollback/revert incidence rate, how often people interact with the MR review bots and implement their suggestions/fixes, how many times people check back on AI transcriptions/meeting notes (hint: Nobody looks back on any of it, it's all just noise that gets generated and never actually referenced outside a few extremely rare cases) and many more things I'm forgetting. It is an imperfect number of course, because measuring productivity in engineering is a sisyphean task, but in my opinion it is accurate to the reality on the ground and outside of all the hype and marketing bullshit.
So, I remain thoroughly unconvinced of these personal anecdotes of people being "massively" more productive, especially once you factor in the fact that we now have a 2000EUR budget/month/dev for all the AI tooling, those productivity numbers start looking pathetic once you factor in the costs (which are only increasing as the AI companies need to start recouping the gazillions they've burned). Some teams have started begging to disable coderabbit and other similar tools in their MRs because they're producing nothing but walls of noise that makes reviewing any MR a nightmare of sludging through endless slop of useless bullshit, ours included.
Animals don't "work". Not atleast for their own sake. If there is enough green pasture and water around, they don't even migrate to other places. So if work is meant to provide food and shelter and if machines can ensure that, humans don't need to "work".
Wealth is only a reserve capacity to help future generations so that they don't need to work for their basic needs. But if machines ensure that too, then wealth itself, as a reserve, is unnecessary.
How many man-hours go into various parts of the advertising distribution chain? Though a certain fraction of that energy goes to connecting people with goods and services they might find valuable, most of it goes into shifting numbers around for people that already don't personally have to worry about money.
We don't need to find endless ways for people to spin wheels, but as long as we're worried about "jobs," we will. We just need to find the social structures to provide people with basic needs and reserve "work" for things that are vital to society or truly inspired.
just?
Not that much. AI is already heavily automating advertising and has been for a long time. And a lot of activity that was previously exposed to untargeted passive advertising - like TV - has shifted to mediums that don't have any, like Netflix (for most subscribers).
And as you note, advertising isn't useless. It's how people find out about things they didn't know they wanted.
Until the machines aren't owned by anyone (or owned by everyone, take your pick on the phrasing), the owners of the machine have no need to keep you alive.
This take is basically "Don't worry, people like Sam Altman are looking out for us"...
These "needs" are sometimes enforced by the systems and government so that people don't stay away from the work and "economy" keeps churning. The housing prices could be a way to keep the people working for loan payments.
Instant foods, nursing homes for elderly, creches, roads, commuter trains - are all ways to have more workers and make them focused on work.
I can't wait to be kept in agistment by my overlords, fed on treacle and oats, ridden in circles once a fortnight, and shot when I break a leg.
[0] https://www.researchgate.net/figure/United-States-Farm-based...
And what reason do they have for killing everyone else? Where is this abject nonsense even coming from? You can't just assert that the "owners of the machine" are all cartoonishly evil for some unknown reason.
All else being equal, at least Sam Altman et al. aren't constantly making up fantasies about exterminating people.
When enough green pastures are around animals usually reproduce till that is no longer the case and they need to move to other pastures. They continue doing this untill all green pastures are occupied. After this they start competing with one another to compete for the green pastures already occupied. Some animals will be so succesfull that they take larger green pastures, letting others starve. If by some miraclelous event (ai?) suddenly a lot of green pastures arise, animals will simply reproduce again till they occupied all green pastures again.
Looking at us? The agraric revolution, neither the Industrial revolution, nor this ai revolution have decreased how much we work on our 'jobs' (finding food indirectly). All it did was increasing the population so that all green pastures became occupied.
My dog was perfectly happy to snooze most of the day, play with toys occasionally, go for walks and try to hunt rodents. I on the other hand would be incredibly bored with that.
We will never automate all work so we with half of humanity doing nothing of value it will be a struggle for the people who do nothing of value to convince people to do work for them.
We can see it now where products dont target the people without money. There is no point because they cant give you any reward so instead you do your work for the people that can give you something in return. We can use the government to stimulate and balance this a bit but at a certain point the number gets to high and things collapse.
Some may consider local LLMs to help with that as the power of LLMs would be more distributed. I think local LLMs would help marginally. Companies as an entity would still be better positioned to use these to their advantage vis a vis individuals.
So projecting into the future I still wonder if it won't become more challenging for individuals to make a living on average.
The drive some individuals have to control / exploit others isn't limited to the rich, they're just in a better position to exercise it
There are people who have absolutely no interest in what other people around them are doing; there are others who will move heaven and earth to help those around them whenever they can, even at the expense of their own interests; others still who can't help but compare themselves to everyone and make it everyone else's problem if they find themselves lacking; and yet others who at all costs want people to afford them "respect" which when you drill into it means stroking their ego
I'm sure we've all experienced petty individuals who've been granted just enough power that they can gain personal satisfaction from leveraging it to make themselves feel better at the expense of all other considerations
These are all just some of the many flavours of humanity
It's a status thing, right?
So if things continue as they are today, I think in the near future, being a software developer is going to be more analogous to the medical field, where in the medical field you have different levels of professional expertise.
Some will be like nurses, and some will be closer to a medic and a smaller set will be like doctors. Each with increasingly required knowledge and experience to fulfill a needed role.
Those who used to be actual software developers are going to be (or have to become) more in the doctor role with years of internship and practical experience to be the architects guiding the overall AI implementation of software development in organizations.
The medics are going to be people who are semi-technical, where they have some technical understanding but they don't dedicate themselves to it, like say product managers, where they jump in to help development along, but don't need to have many years of experience or very deep technical knowledge.
At the nurse level, it's probably going to be similar to what people would do in the past with no code tools, where somebody in marketing who knows very little to nothing about coding at all is just going to directly converse with AI systems, but they'll never be likely to get anything more advanced than the tools they could think up for themselves.
Of course, it's so hard to tell what the next big discovery or changes to the nature of world society might push things in one direction or another.
What upset me a bit were phrases like “This is not a slogan. It’s a framework” which immediately devalued the work for me.
I have read so much Ai generated text recently, that I developed some AI-fatigue or AI-burnout, and I’m wondering if that might hit more fields - making more humans reject Ai work.
To be clear, I still like the text and I don’t know if it was written (partially) by Ai or not - but it’s this uncanny feeling I got reading it.
Why? Because it resembles the typical "You're not just implementing a text editor, you're reshaping the text editing landscape"?
I cant read this shit anymore.
"creating cross functional AI evaluation team to keep the company honest"
Fucking garbage.
I have an old book between me and the keyboard, and just read it while the AI is thinking, so I can read non AI words.. otherwise its like reading books by the same author over and over and over again.
I will go touch some grass now.
According to WRITER’s 2026 Enterprise Adoption Survey, 44% of Gen Z employees admit to sabotaging their company's AI strategy in at least one way compared to 29% of employees overall.
Sabotage behaviours include entering proprietary information into AI tools, using non-approved AI tools, refusing to use AI tools or outputs, ignoring guidelines or best practices, intentionally generating low-quality outputs, refusing to take AI training and tampering with performance metrics to make AI appear to underperform.
Our technology has long since surpassed our capacity to understand its impact or to morally restrain its use - we've gone from chivalry to "oops, that was a school, oh well"
We will solve problems faster though, but we aren't showing signs that AI is making us better humans, and that is the problem.
There will always be new, hard problems to work on. AI will not, and can not eliminate that.
But we already do have have some kind of measurement of most of these types of side factors, and they actually aren't at zero and are increasing rapidly. So the implication that they will not be human level until decades from now is just (hopeful?) speculation or fuzzy thinking.
To me this looks like a really academic and official sounding version of the same quasi-religious hopium that usually defends the sanctity of the human. He is essentially saying that there is just something so special about humans that it will never be reproduced in a machine. It's very similar to dualism (and in many people actually is religious dualism). No AI is going to have human creativity or judgement. Not anytime soon. Why? Well, we all just _know_ that's not possible. Okay, maybe in a couple of decades (but they don't necessarily believe that anyway). Why would that take decades? Well we all can just _tell_ it's no where close, right? Because AI of today just isn't special like humans.
Aside from that worldview issue, I think that people still are not taking seriously or internalizing the concept of exponential improvement.
Computing efficiency gains can actually level off. In fact, they have many, many times before. But they always tilt back up again when we invent the next approach to get beyond the current level. This is how it has been for 90 years.
There are multiple ways that we continue to see huge gains in AI software, architecture, and hardware. There are huge efficiency gains available still as we move towards more radical fully compute in memory and/or analog approaches and other options like models implemented in hardware.
- i would assume it is reasonable that anyone comes and see what other posts a person has written except you cant find that page anywhere linked
> A battle of two narratives > Build wealth before AI obviates our skills > Build skills, agency, taste, judgement
both narratives are portrayed as being odds with each other but, I can't come up with a single "build wealth" scenario that doesn't involve building skills, agency, taste and judgement.
what am I missing ?
I would doubt however that this would be an 'Equals' or 'Implies' scenario. Let go of seeing either of them as binary, and then not even as scalers.
- Work is shifting from building/doing to evaluating, judging, and steering — that's where human value will concentrate.
Other supporting points. ------
- No lab milestone or "RSI breakthrough" will suddenly eliminate jobs — economic impact unfolds gradually over decades.
- Reliability, not raw capability, is the real bottleneck holding back AI automation today.
- Historically, making work cheaper/faster (ATMs, radiology, coding) has grown employment, not destroyed it.
- Superintelligence claims misunderstand human intelligence, which is itself amplified by tools like AI ("co-superintelligence").
It is not a good idea to compress articles like this but there are many of these opinions to read and trying to get to the point quickly to uncover new viewpoints.
https://www.reddit.com/r/gifs/comments/3p0b3i/graphic_design...
It's possible that the elite will control most part of wealth generation, keeping it for themselves. The rest of society will develop an underclass economy and work for each other. Essentially a worldwide slum.
If you think it is different, just think of how many people write books professionally, or even publish online.
Once the noise settles down a bit and boardroom shakes off their delusions as you can see in rehiring in Ford and Zuck who was very bull on AI remark about "not being it". It will be just the same, but different.
Ford: https://www.bbc.com/news/articles/cgrkd41n2v9o IBM: https://qz.com/companies-rehiring-workers-ai-layoffs-automat...
Ourselves.
Does he want to fund the arts? Humanities?
Your analogy isn’t necessarily wrong, but it might ignore the extreme importance of nurses. Many medical facilities are only staffed with permanent nurses, with doctors helicoptering in, from time to time, to take care of specific duties that may require certain licenses, or provide specific advice.
So lots of jobs for nurses.
Maybe for a very loose definition of medical facilities that includes assisted living facilities.
But for example in an ER, nurses come and go with very rapid turnover and it’s common to staff with temporary travel nurses.
> nurses tend to do most of the actual work
Techs, environmental services, phlebotomists, respiratory therapists, CNAs etc. probably do more of the “work” than nurses.
> highly experienced nurses actually taking up the mantle for many duties often done by doctors
Only if they go back to school and become a Nurse Practitioner or CRNA, but in that case they are no longer functioning as a nurse. Even then they are general operating under the direct supervision of a physician.
> Nurses are in far higher demand, than doctors.
Only in absolute numbers. It’s far harder to hire a doctor than it is a nurse. I know an ex-NFL player who works as a physician recruiter.
Exactly. (Experienced) Nurses can do ALL of the basic stuff we currently use doctors for. They diagnose patients with a glance (not by going to medical school but with raw lived experience on the field), administer medications etc.
And in some countries (I think the UK?) they are allowed to do more of the doctor-y stuff to free Actual Doctors to do the stuff they studied for years for.
The same thing will come to the software industry. All of the basic CRUD HTTP API crap can be done by "nurses", just run of the mill domain experts can wire up a basic UI + API + DB -combo for a product MVP with AI assistance.
And with a "doctor" (more experienced programmer/architect) guiding the process, the quality will be slightly above average. Not artisanal excellence, but how many services actually need that, really?
I think I've mentioned this before, but the only ones whose jobs are at risk at the moment are the mid-tier programmers. Either by skill or experience.
Juniors can learn to work with AIs and easily become the "nurses" of programming.
Programmers with decades of experience in multiple fields can either go full the artisanal-no-AI route and do the things AI's suck at or do AI assisted programming, which isn't really that different from working with a team.
But the midtier people who are too well paid to be "nurses" and not experienced enough to work as "doctors" might find it hard to find a place for themselves.
There's a much bigger group of people building games with Unity than the number of people building complicated engines like Unity. Same for [insert Javascript framework here].
I spent some time in the military, and my expression of medics and nurses are mostly derived from that experience, where I'm referring to a nurse as just any warm body who is able to provide aid.
For professional nurses who might work in hospitals, I'm sure that many of them have significant knowledge and experience to be very effective in providing medical assistance.
Wishful thinking by the managerial class. At best they can vibe code but they can’t verify that what was written is correct.
In such a future there would be a handful of lucky well paid artisans, a healthy community of hobbyists, and the overwhelming population of the planet who would be perfectly content to delegate their entire software diet to generative superplatforms that script themselves to perform any arbitrary software function. The idea of paying for individual bespoke software programs would become an anachronism for an era where software was so difficult to produce that entire teams spent years painstakingly tweaking programs to spec.
I'd like to imagine that it's ultimately going to work out for humanity to be in a better condition overall if that happens.
However, if there reaches a level of near or full autonomy in all aspects of knowledge work, then there is a strong possibility that the field with the highest growth potential is going to be in "Private Security"; as those who have access to the most resources seek to defend their own positions in a societal return towards aristocracy and serfdom.
Let's hope not though.
The sell was - Creating your apps was easier than before. You don't need to know any programming language. You don't need IT. You have a GUI and just drag and drop and visual your apps. Just fill the fields and voila your app is ready.
The criticism of the approach was that low codes apps might not be secure, unsuitable for large scale and critical apps, and lead to increase in unsupported apps by "shadow IT".
The same reasoning exists today for AI coding. Anyone can create apps, its not great for complex and mission critical apps, might not be secure etc. And lots of discussion like your post parsing the future too closely.
Low/no code apps have continue to grow. But many of the low/no code tool users are developers who use it to make their jobs easier.
While some might say - this time it is different. I believe we are currently we are the beginning of the cycle so everyone is excited to use their PowerApps shaped Claude Code/Codex to make their 100th budget tracking app but as time passes and edge cases are figured out the biggest users for AI are going to be software developers (and I believe they currently are the biggest users).
As for software engineering jobs just like before engineers will be expected to output more and faster. This has happened with every innovation from assembly compiler to IDE to low/no code ways of building.
The different levels of expertise exists even today and it will remain so.
Those are the people who will get the boost from AI coding. They have the "lazy" mindset every good programmer has. You do a thing 3 times manually and then your brain goes, nnnope. This needs to be automated. And you get to work.
With the n8n's of the world they can get something done, limited by what they offer. Same with Excel, there's a limit to what you can do with macros and VBA.
But with an AI agent there really is no limit. I have seen first hand the things domain experts can do when just given a Claude seat and the permission to build stuff. Things I could've easily built, but never would've found out are actually needed.
And doing the same thing with a programming team and project managers in the loop with the domain expert being the product owner would've taken months and cost six figures in salaries.
with the latest agents, you can even be vague and they'll probably do an ok job if it's not something super complicated. some models are also pretty good at stopping to let you know about stuff you haven't thought about and ask for clarification.
so you can outsource thinking now, not just the doing part.
I think many people will wish it was like this. But when AI becomes so capable that a nobody who doesn't understands computers a well as a "doctor" uses AI to create something BETTER or superior or just as good then employers will think... why am I paying that "doctor" more than that nobody?
See right now everything is fine, because AI is not that good yet. But if it gets better. Well. The future will be one where we trust AI more and more.
It's just hard to design robots that can handle patients that may be simultaneously fragile, mentally handicapped and aggressive in a way that doesn't hurt them and respects their rights. It can take multiple human nurses to do a seemingly simple thing like changing a diaper.
So I think it would be more comparable to something like literacy. There was a time when that was a fairly uncommon and highly valued skill. Now the guy flipping burgers or pouring a cup of coffee is also almost certainly fully literate. And in fact many jobs have evolved in a way such that it became mandatory, but only because it was already ubiquitous. I expect to see the same thing with software. The industry of producing software that do fairly simple tasks will probably die, but in its place will be a vast array of heavily customized and oft iterated software for companies and people achieving their own stuff.
The mobile industry is a perfect example of where this will be massive shift. Right now there's a million mobile apps to execute extremely basic functionality on phones, but it's loaded with advertising, begging, and general annoyances. As are the app stores themselves. When you can make software that does that in a few minutes with a single prompt, and people realize this (as we're already practically at this point), then that will be the end of those apps. This is because the one thing LLMs have shown is that natural language interfaces are way less friction than using search, whether on the web or an app store. And so there will be a time when it will be lower friction to simply just quickly build your own app to do [whatever] than dealing with somebody trying to monetize an alarm clock.
I don’t understand how people can say this and then continue talking about software. So we’re saying machines can now casually do complex and cognitively demanding jobs like software development (or 90% of all white-collar jobs out there) and we’re NOT worried about the lynching mob going door to door and hanging IT people on lampposts? And I’m being serious, the impact this would have on societies would be unprecedented.
The medical field is also going to change though. Massively. Because people are going to realize you don’t need to pay someone $400k per year to hand out advice about moderate exercise and which antibiotic is appropriate for a sneeze-cough with yellow mucus.
Regulation isn’t going to prevent this. AI is already way too easily accessible to ever rein it in again. Not to mention that the US now has serious competition from a hostile country, so they can regulate their own AIs all they want without it making a difference in practice.
Who is going to realize that?
The same forces that prevent you from walking into a pharmacy and asking for antibiotics based on what you found on WebMD will prevent you from doing it with a ChatGPT printout in hand. Lawyers and doctors are the best-known examples of industries that are in control of who gets admitted to practice the profession.
> Some will be like nurses, and some will be closer to a medic and a smaller set will be like doctors. Each with increasingly required knowledge and experience to fulfill a needed role.
Nursing and being a physician aren't really the same thing at all, and they require different skill sets, it's not just "having more knowledge". Just because someone is an amazing surgeon doesn't mean they would also make a good nurse.
> Those who used to be actual software developers are going to be (or have to become) more in the doctor role with years of internship and practical experience to be the architects guiding the overall AI implementation of software development in organizations.
I think you just described a staff swe
> The medics are going to be people who are semi-technical, where they have some technical understanding but they don't dedicate themselves to it, like say product managers, where they jump in to help development along, but don't need to have many years of experience or very deep technical knowledge.
These people already exist. They are the business analysts who know SQL and maybe Python, R, or VBA. Marketing people who work on Wordpress landing pages. People doing systems integration, the IT department, sales engineers, and on, and on, and on.
> At the nurse level, it's probably going to be similar to what people would do in the past with no code tools, where somebody in marketing who knows very little to nothing about coding at all is just going to directly converse with AI systems, but they'll never be likely to get anything more advanced than the tools they could think up for themselves.
You said it, no code/low-code has existed forever.
This is true as a sentiment, but my understanding is that the majority of students are overwhelmingly using AI for ~everything. If a thing provides massive utility people will use it.
They say they are, but they have actually been using it so much that they can't even do they homework without AI at this point.
Very open definition of sabotage.
Everybody says they hate heroin but once you try it, you can't get enough of it.
Also new heroin users are only 30% likely to become addicted. https://jamanetwork.com/journals/jamapsychiatry/fullarticle/...
That seems a rather high baseline, maybe even more significant than younger generations putting another 50% on top of it as they are wont to do.
But I do believe (gut feeling) the novelty's worn off for a lot of people.
Jinxzi has Bernard as friend though
But it's at my day job, and it's because I was able to write a prompt which automates having Copilot review uploaded scanned PDFs of invoices with checks (and the bank line obscured with a pen, so no PII) and then write a batch file which renames the files per a file-naming convention, removing the need to open them in batches of 50, find the Invoice ID, re-save using that filename, then quit and re-launch Adobe Acrobat (if left running, eventually I run into a bug where it stops saving files), then run a .bat file which renames based on Invoice ID as a filename.
Problem of course is I've been running into a limit of number of allowed files per 24 hr. period.
Even if it's not less work, it feels like less effort.
The problem is they are now paying me more, plus paying for the cost of using the AI, and the needless complexity also slows down the employees. So more costs there as well, any future debugging is going to cost far more and at the end of the day they are getting less quality on the core function but far more presentation data that is essentially meaningless.
But I'm trying to figure out if this is a thing in my organization. So far the channels discussing AI have been about how they use AI, how to build agents, and complaining they've run out of tokens - but nobody says anything about how much work they finished, how it impacted the product(s), the user, and the company's bottom line.
I don't know if they just don't know, don't care, don't measure, or they don't actually want to admit they're just burning through tokens with nothing to show for it.
That said, right now I'm using AI to extract some code from our app to create a minimal reproduction for an issue we're having with a 3rd party library, it's a huge time saver there - that is, I wouldn't have bothered making a reproduction because of how time consuming it can be.
So to answer your question, it's not that "I have less to do", it's "I do things I wouldn't have done otherwise because they cost less effort/time". Which is also the promise of e.g. automation - the automated loom didn't just reduce how long it takes to make cloth, it made it so people make a bajillion times more cloth.
For example, all the work like translations that used to be done by humans, now it is a CMS AI feature.
Secondly teams setup.
It used to be we did everything ourselves for development, then cloud, SaaS products and serverless decreased the teams size required for delivery.
Now with AI, there is an even greater push for low code/no code tooling, with agents, leaving the actual programming left for MCP tools that might not yet be available for the project.
Thus you get a team of five doing what used to be about 15 a decade ago.
It pushes the bulk of the work to review. So teams with good practices can account for this.
For junior teams, the time saved is massive, because they aren't doing all the other practices required to prevent technical debt.
I didn't do fewer hours in these weeks but had time to explore and innovate a little.
> It’s funny. I was looking at my GH activity graph. It’s been pretty solid green, for years. I stay busy.
> But since I’ve been using an LLM, it’s been bright green.
> I always check in code manually. I don’t let the LLM do it.
I can now write software quicker in most of the cases, but the rest of the organization moves as slowly as ever.
But before that I had the same output, but with less of the boring typing stuff.
My mother worked in an office in London as a shorthand typist in the 60s, along with thousands of other young women.
At some point computers began to enter offices, bosses typed their own letters and then emails, and this category of job simply evaporated. Of course there was still SOME secretarial work, and some workers retrained to do it, but most simply had to leave the sector.
Isn't that a common story with technology replacing workers?
Once you've used a BBS or modem, "online banking" is an obvious idea - just a very difficult one to implement until you add on decades of security and improved connectivity.
But also, the article is explicitly arguing against armageddon.
I don't think your assessment that this is mostly automated stands in the face of the number of man-hours both of these companies purchase each year.
Luxury horse living during the heyday of working horses and pit ponies, "horse power" wasn't left ideal for a fortnight.
> How many horses do you see now that the world
Personally, a surprising number perhaps, there's a pony club at the top of my street in town, and the area is still littered with horses and other livestock.
This isn't my area, but it's not dissimilar: https://news.ycombinator.com/item?id=45623799
Full size image: https://live-production.wcms.abc-cdn.net.au/a26664f6500a7c74...
That’s not true. They do completely different jobs. Some experienced nurses could do some things that we currently require doctors to do.
Other experienced nurses have problems doing basic dosage calculation.
The main difference that addition to med school being much more selective, rigorous, and longer, residency is regimented, standardized, supervised, and evaluated in a way that nursing experience isn’t.
There are entire teams of doctors who spend hours every month discussing each resident’s progress (I know because my wife runs the resident program for her department and I overhear the discussions in the background).
You might have one nurse who worked in the ER for 5 years who can diagnose appendicitis as well as a doctor, but another who worked there for 20 years who couldn’t even begin to do that.
And since diagnosing appendicitis is not part of their job description, both could have absolutely stellar performance reviews.
And the nurse who can diagnose appendicitis, might be terrible at other tasks normally handled by a doctor.
The solution for this is to put them through more standardized on the job training/evaluation, but you’ve just reinvented residency at that point. That’s the thing people don’t understand about residency, it’s as much evaluation as it is training.
No amount of on the job experience is equivalent because job experience isn’t rigorously regimented, and evaluated.
Is it different? because then:
> you can even be vague, and they'll probably do an ok job if it's not something super complicated
That's lot of qualifiers to say maybe, should be possible - how does someone who is not a software engineer know what is complex, vague or ok. There are lot of assumptions to say - This time it is different.
It is entirely possible it is different but I am yet to find a convincing argument.
The position of (for instance) the check # on a physical check varies quite widely, as can date position and format, and a fair number of them are still written by hand.
On top of that, I'm scanning thousands of these each year, and the invoice underneath the check ranges from the gamut of: "pristine copy just re-printed 'cause the customer didn't include one" through "bad inkjet photocopy of a photograph taken w/ a potato phone and then printed" and includes variations such as "customer included half-a-dozen invoices to be paid w/ one check, and if arranging them and the check _just so_ all will fit in the document camera window, saving a trip to the sheet feeder scanner"
A co-worker actually worked on this for a different program, one where actually sending in paperwork in good condition was expected and customary, but his system had a reject/failure rate of ~10% --- that would be almost 1,000 invoices each year for the program I am handling.
Human agency is real and powerful - unless humans want to automate all work, it won't happen.
Yes, but this might also be a counter-point to your position. In a world of rising baselines, having a "pot but without all the bells and whistles" might be a thing people need. Instead of having a pot that holds fluids at every conceivable angle with 20 temperature thresholds, 5 versions that hold liquid in a cloud for you, with monthly subscriptions, having the ability to get a "like a pot but make it left handed and only holds sand because this is what I need" could turn out to be something that people want/need/end up chasing.
In other words spec down, not up, and base + my particular kind of a pot.
Ah I see you're familiar with PowerPot by Atlassian(TM)
Very powerful potware, pity about how much soil it leaks.
When lawyers and writers are talking to me about "docker containers" and "agents" I assure you that the amount of code out there is going to grow.
Of course, billionaires have other plans, and are the main obstacle in achieving any sort of social cohesion.
"Work" doesn't exist to keep people busy, it exists to keep them alive.
Even now the machine owners have enough power to change laws to their liking, bending governments and public opinion to their will.
I have no idea what that means.
> Why does the "world" have that power?
Okay, lets say "the environment". The universe does not care whether you live or die, so why are you so sure that someone will paternalistically care for you?
On what are you basing this viewpoint on?
> Machine owners don't have the power, for example.
That's because the world doesn't care about them either. So why would the world care about you?
I don't buy this. Tons of people in the West live on welfare and don't contribute shit to the economy, they are net takers. No one is plotting to kill them, in fact they're (arguably) getting more than they did 50 years ago and definitely more a 100 years ago. We live in democracies, not in a dystopian nightmare and I don't see why A.I is going to change that dynamic. People vote, people control the government, the government controls the armed forces, private citizens are not allowed to gather unregulated weapons... billionaire or no billionaire I don't see how you can beat that. And also, it's not like the billionaire class is some kind of a cohesive group that wants to work together to rule the world - they pretty much compete with and hate each other. I don't think Musk, Altman, Hassabis and Larry and Sergey are all going to agree to work together on "controling the machines" and killing all the other people.
No one said anything about killing.
And people had to fight for it. Through voting, yes, but also through strikes, sometimes riots, and some died for it. People all the world fought and died for the right to strike (basically, the right to refuse to work). For the right to organise as groups. For the right to only work 8 hours a day, for the right to a have a weekly day off. For the right to pay for pensions. For the right to have some sort of health insurance.
All over the world. Some still do. Many still don’t have half those rights.
Now, I don’t know about any billionaires specifically. But welfare didn’t fall from the sky. Or anyone’s good grace. It’s all about leverage.
If in doubt, looking back at the late 1800s and early 1900s work and life conditions across the Western world is enlightening.
And I think that example largely carries over to many other big picture and/or creative tasks. I think it would likely require another revolutionary leap for LLMs to get to that point, and that leap is probably not coming any time soon, if ever, at least not with LLMs.
Another practical issue is that LLMs, as difficult as it can be to imagine at times, are still just glorified autocomplete machines. And so genuinely novel advances will not be coming from LLMs. For instance take LLMs back to the early 20th century and none of them will discover relativity, even with infinite power/time. It's just so unlike their training material, and in direct contradiction to much of it, that it's simply outside their reach, let alone grasp.
Tests can't be exhaustive, whereas even a mediocre developer is likely to possess enough background knowledge to notice risks that a non-developer would not think to test.
Some people keep forgetting that we haven't reached AGI yet. These tools can still make serious and sometimes obvious mistakes. Not long ago, vibe-coded software could embed credentials directly. That particular blunder seems to have been addressed, but there is still no reliable way to tell an LLM to avoid every class of obvious blunder.
Heh. I wouldn't be so fast in calling this one, yet. It might go either way, or a combination of the two, who knows... Things are moving and progressing fast enough to at least be wary of, and keeping an eye on things. I wouldn't be surprised of any outcome, tbh.
On the one hand, PMs get to hone in a set of skills that include lots of juggling resources, bridging the gap between stake holders and people that execute, work with different layers of technical expertise, lots of back and forth and so on. Having to "call" a thing after talking to 5 people from 6 different PoVs is interestingly something that could make them quite good at using "AI". They're already used to working with "jagged expertise" so to say. And that's really close to what "agents" do today. You might get a session where claude/gpt/gemeni turns out a brilliant piece of technical artefact, or you might get an average piece of content that misses key important aspects that a "human" could see from a mile away.
On the other hand, LLMs do one thing quite well: they raise the floor of what "minimum effort" in a thing gets you. So things like language barriers, basic processes, procedures, etc. get elevated to a level where you can use these tools and be better off than not using these concepts at all. So a very talented engineer that previously had these issues, could now use LLMs to get better at auxiliary but "unimportant" aspects of integrated that technical experience into other places. In a "90% of the time people expect this to work like this" way, but with the floor being raised. One quick example of this would be a technically sound piece of software that now has bare-minimum ux/ui from this century, instead of "engineer GUI" aspect :)
Who knows where this all goes. Yesterday there was an article here about someone taking a bunch of languages, asking an agent to mix concepts from all of them and slop a "new uber language" design. Is it the ultimate language that many have tried before? 99% sure it's not. Or at least this one won't be. But there's a chance, with how many people can now prompt their way to a PoC quite quickly, that we eventually get something cool. No idea how that'd look like, but it's likely the "you'll know it when you see it" kind of thing.
Mostly for drugs that should probably be OTC anyway. Try doing that for a narcotic, or for an expensive biologic.
You could always lie to a doctor about your symptoms.
Garbage in, Garbage out.
The University market is brutal too. If you aren't using AI too, you are falling behind. Many see it as a means to an end.
Well, one thing about AI... if it does become our overlords, maybe it won't be so eager to be wheedled into giving passing grades. :/
Have you read the research that says AIs are more likely to react favourably to output based on their own model compared to those of a different AI model? Guessing teachers subconsciously grade similarly, like people who use one model get more of some grade than people who use another one...
Guessing this is also why so many liberal arts majors are being cut.
I 'sorta' get why people might use AI in a required class though I am not for it, but why major in something and do it? I mean aside from wanting money (and, really, many of those majors don't make much).
But are they also paying for it? Or simply using the free version because it's available?
Who said they would kill everyone else? Since the rest of your strawmanned caricatured argument is a response to an argument never made, there's no point in addressing it.
Who said anything about killing?
But that's neither here nor there, since we're talking about a world with an artificial superintelligence. That's the hypothetical here.
I.e. some source code gets compiled, the output gets decompiled again, the result gets compiled again, and so on.
Now do the same with LLM: start with a prompt, it generates a program. Then an LLM examines that program by looking at the source, running it, etc., and is tasked to describe it, i.e. turn it into a prompt again. Then that prompt is used to generate a program, which is examined and turned into a prompt again, and so on.
Someone she’s knows calls her up late and says their child has a minor fever, upset stomach or some other mild symptom. They ask “should I go to the ER or can I just wait till tomorrow to go to the pediatrician?”
My wife says it’s totally fine to wait until tomorrow. That is not an emergency. And frequently she’ll tell them they don’t even need to go to their pediatrician unless it gets worse or continues for several days.
9/10 what happens is that the person goes to the ER or their pediatrician first thing in the morning and calls my wife to tell her “you were right, it wasn’t an emergency”.
Despite that my wife is more highly trained than either their pediatrician or a non-pediatric ER doctor, and they know and trust her, they still wanted to see someone in person.
It doesn’t matter how accurate AI diagnosis gets, that’s not going to change anytime soon.
Well, that someone might use AI diagnostic tools too.
> Maybe for a very loose definition of medical facilities that includes assisted living facilities.
Or midwives, for unimportant things like giving birth /s obviously
It may take vastly more training but on average a full annual physical provides less benefit on average than a 30 second vaccination requiring minimal training. Value creation and skill are wildly different things in the medical profession.
Healthcare is a world of diminishing returns at ever increasing costs. But that doesn’t mean the low effort first steps are less valuable.
But nurses aren’t the driver of vaccination. Nurses are anti-vaxxers at a much higher rate than physicians, and if you removed physicians from the equation, vaccine rates would plummet.
>total medical care provided
At the end of the day a physician is responsible for every patient that comes through. Nurses are incredibly important, but all of the care they provide is at the direction of a physician, so it’s impossible to separate out “total medical care provided” into doctor vs nurse.
The odd thing is that after the pandemic just showed what societies can enforce, somehow it all is forgotten again when it comes to who holds what power.
What is necessary are pretty basic killer robots with the capability of determining who is a friend and who is a foe. These systems already exist.
Labor is a commodity that has had a diminishing value in our current, capitalist system. That means, people who rely on their labor(most of us) instead of capital(a few), have control over less and less resources(this is already happening).
With the current trend, what generation of your grand children will have nothing at all?
Or perhaps, they stand to benefit?
If we are just talking about the physical environment then totally fine - a rock doesn't care about me, for example.
Humans tend to care about humans at least in some situations on average, though.
Look, I actually agree: humans do care about humans, on average. The problem is that power tends to concentrate to those who don't care about humans.
When almost all jobs can be replaced by machines, the owners of those machines are unlikely to be much different from Sam Altman and Co. They aren't going to sit down and say "Well, look, we have all the wealth, and means of wealth. Let's give everyone food, shelter, clothing, because our machines can do it".
As for AI fixing things... Sure, at least for a while. But in my experience, AI can hit a wall and start going in circles. It doesn't happen often, but it can happen. And when it does, what recourse will you have for fixing the tangled mess that vibe coding tends to produce?
Is that true anymore when anyone can vibe code? Seems to me that quality, correctness, and performance will be huge differentiators in a see of vibe coded slop.
Physicians have to understand the diagnosis and treatment plan. At the end of the day, they are legally responsible for the patient.
There’s nothing wild about that prescription.
Bacterial sinus infections are absurdly over treated [1]. It can take weeks to recover from a viral infection.
Not saying that you are a wuss, but my guess is, the doctor probably over-indexed on the throat pain, assumed you were a bit of wuss for calling the doctor for a sore throat, and decided to give you something to make you feel better quickly.
Either way unless you were 65+ a short course of oral steroids is very safe (and even then it’s only very mildly dangerous).
1.https://www.aafp.org/afp/2020/0615/p758
“Without antibiotics, rhinosinusitis resolved in 46% of patients after one week and in 64% of patients after 14 days.
Antibiotics can shorten time to resolution but in only five to 11 more people per 100 compared with placebo or no treatment.
Despite this, approximately 86% of U.S. ambulatory visits for acute rhinosinusitis result in oral antibiotic prescriptions.1 In Europe, antibiotic prescription rates for acute rhinosinusitis in primary care range from 72% to 92%”
It's like a light switch where the light has a slightly different colour every time you turn it on, sometimes the colour is very different, and sometimes it emits bursts of heavy gamma radiation. And the argument is "you flick a switch, and what you perceive in that room changes, it's the same experience." (which is a neat way to put it, because it's true even when you get evaporated by the gamma radiation burst)
That this isn't "the same kind of tool" than just a light switch and a bulb isn't even an interesting conversation to have. Of course it's not, not even close. A piece of string is not a steel rod just because you can stretch it out, take a photo of it, and play pretend.
So the really interesting bit to me is that this level of argument is even made. It's nonsense, but if the nonsense gets repeated often enough, maybe the people refuting it can be exhausted and then we can replace concrete with pudding, right? Not that this is your intent, but this stuff just feels like the thing you'd hear in a digestive tract, not at a table between peers.
Quite so. It is true that most compilers allow you to opt in to reproducible output, but it is generally not the default [iirc, gc is the only mainstream compiler that makes always reproducible output a design requirement]. There are advantages to not have to worry about reproducible output, like performance gains in allowing threads to run in any order. Of course, LLMs equally enable you to opt into reproducible output (temperature=0). Implementation details are immaterial. The human experience is always the same: Input one language, out comes another language.
> So the really interesting bit to me is that this level of argument is even made.
Yeah, it stems from a lack of base understanding of the technology. The "a compiler is deterministic but AI isn't" is always at the heart of the argument, but it isn't even true. Implementation details are immaterial, but when you don't understand the implementation it stands to reason that one would want to focus on that facet in order to learn more about it. Hey, that’s what discussion is for: to learn.
This is just splitting hairs. And temperature = 0 is not comparable either, because the we will not keep models trained on data of a specific point in time around forever.
> Implementation details are immaterial, but when you don't understand the implementation it stands to reason that one would want to focus on that facet in order to learn more about it.
Not splitting hairs is not the same as not knowing what a hair is.
> Hey, that’s what discussion is for: to learn.
You told me nothing I didn't know or didn't expect, because it gets brought up every time like clock work. It seems to be be part and parcel of not seeing or not wanting to see the woods for all the trees.
I think we’re done here man.
I guess way to argue with yourself? That’s not what I said.
> Not saying that you are a wuss
Then I provided evidence to demonstrate that it’s more likely than not that the doctor prescribed you antibiotics unnecessarily.
So neither of the straw man statements you created were accurate.
Equally, if you replace one compiler with another, it is almost certain that the output will not remain stable, even when you have enabled reproducible builds. You are not going to find a difference on the outside, even if implementation may differ under the hood.
> You told me nothing I didn't know or didn't expect
What's in it for you, then? I have been able to learn about your character, which is quite fascinating. Technology is pretty boring, which is why nobody else was talking about a technology implementation to begin with, but HN accounts are quite interesting to learn about. Seems to me like a complete waste of time if you cannot learn anything from it.
For me this exchange kinda was, but I don't see what I could possibly do about that.