A challenge with this kind of study is that coding agents (Claude Code, OpenAI Codex) only started working really well in late November, which for most people meant early January due to the December break.
General agents (OpenClaw, Anthropic Copilot, ChatGPT "Work") started working even later than that.
This category of software may have a much more meaningful impact on work than the mostly-chat systems we were using from 2022-2025.
Studies that mainly focus on 2022 to end of 2025 might be missing out on a material uptick in capabilities.
Anecdotally, I (along with the rest of my team) got laid off from a big tech company at the end of January. The stated reason was "AI", although I think as we all know the real reason was "we overhired in 2022". I managed to get a job by April without putting in a lot of work, though (mostly just listened to recruiters until something hit my fancy). My work experience is good, but I don't know if it's so good that I would be immune to market effects. The company that hired me is still hiring other software developers, and I was told that it took them a long time to find someone with my qualifications. (We use Claude, although so far I haven't seen a lot of AI psychosis like I have in other places.. that was also a big filter in my job search). Long story short, I think narrative of job loss is way overblown. It's more doom trolling from anthropic and openAI and their mouth pieces as far as I can tell.
I won't believe in it, because I have seen it happen in agency work.
What used to be contracts for translations or asset creations for CMS projects, is now handled by the AI tooling of CMS products.
The adoption of SaaS, iPaaS and serverless, meant delivery projects could be done with smaller teams.
Now with iPaaS adoption of agentic workflows, and AI tooling to transform those serverless actions into MCP tools, the team sizes have been further reduced.
Most agencies aren't winning enough projects to justify those that land on bench.
It's so early game, is the real problem. My experience so far, has been that it's barely a junior dev. I've met so many in my career that think reading stackoverflow, or watching a youtube video mysteriously makes them an expert. AI reminds me of this sort.
Well anyone can use (prior to AI) a simple linter and learning to code isn't that big a deal. It's learning the pitfalls, the traps, that's the issue. And so far Opus just seems to fall into them again and again. I guess the best way to put it, is that it's not an architect. I sees no big picture, and that's not really a surprise with (compared to a human) an incredibly small context window. When I'm on a project, or working with a codebase, I often have years of "context window". And I have a career of "don't do this" context window.
So what I wonder is, will this be resolved? Will that awareness of larger scope be solved? If that happens, we'll be in another ballpark of competency.
Some companies have massive codebases. Are these companies slowly gaining rot in those codebases, a swiss cheese effect, which eventually will result in collapse? Because I've worked where a bad hire had just this effect over time. And what I worry about isn't using Claude to speed one up, it's the DEV that uses Claude and just "meh" and submits because it passes regression + other tests, and then a manager or code reviewer uses Claude and "meh" because it's a pass too.
> Are these companies slowly gaining rot in those codebases?
Yes!
And that "yes" holds regardless of whether they use AI or not.
Let's not pretend tech debt accumulation is somehow an AI problem. Some of the world's biggest companies routinely ship code that reeks of years of rot and decay.
It's safe to say that most people romanticize their past experience. Anyone's "golden age" is someone else's "end times".
Software has always sucked [1]. "Software crisis" was coined in 1968. There has been no period of time since that it got any better. Trauma just doesn't carry the same way that triumphs do.
Perhaps people do not actually learn from their mistakes after all.
> Let's not pretend tech debt accumulation is somehow an AI problem. Some of the world's biggest companies routinely ship code that reeks of years of rot and decay.
While that is true, there's a large difference you are ignoring: a company can survive this type of rot indefinitely if it takes place at a certain rate. Their customers get time to acclimatise , the processes take into account the breakage, and adapt to compensate, etc.
If the AI boosters are to be believed (2x, or 10x or even, in some cases, devs claiming 100x productivity), stuffing 2 decades of rot into the next two years could prove fatal.
Our production incidents have dramatically increased over last 2 years after management pushed AI. In 2023, we never failed our SLA. Now, we continuously keep failing it due to some stupid outage. There was no real change in architecture or strategy - only the adoption of AI.
There have already been several high-profile AI-root-cause incidents at many companies this year. I guess we can look forward to even more deterioration in the next 2 years.
> My experience so far, has been that it's barely a junior dev.
At least on greenfield projects, Fable is more like having an endless succession of fly-by senior devs who lock themselves in an office for a week, and who come back with very reasonable code that I then need to maintain somehow.
For longer-term maintenance, I am actually slowly warming to Sonnet 4.5-era models (so October 2025, right before the Opus revolution). They need to be given clear instructions and watched carefully. But since they force a human to stay in the loop, you don't have the institutional knowledge loss I see in some Opus projects, or the code that was one-shot with no human interaction at all that's a constant temptation in Fable projects.
And yeah, the bit rot is painfully real once the humans step back too far. I've been dealing with a compelling prototype that someone built, and that stakeholders love (for good reasons). But it had to be put on a tech debt repayment plan for a couple of months.
I have seen a bad hire having instant impact on team and after person was let go, team was still getting back on track to pre-hire productivity for another 3 months.
We know all managers that say "stack doesn't matter, we can swap devs". We had C/C++ dev hired to do web dev with C# and Typescript.
Guy was utterly incompatible even if he was smart but management wanted someone who is local and can come to the office and that was this guy selling point.
> General agents (OpenClaw, ...) started working even later than that.
There is nothing special about the OpenClaw thing other than the enormous astroturfing campaign that benefited various "crypto" influencers and other scammers. Anyone who endorses it is either manipulated or trying to manipulate you.
I get the point having read much the same from Tesla (and fans) regarding self driving cars that still haven't done half the things that Musk said was just around the corner pending regulators a decade ago and repeatedly since then.
And myself I keep making comparisons between AI and the progress in 90s video games where every minor improvement got called "photo realistic" and then forgotten with the next game engine: https://archive.org/details/nextgen-issue-26
So I'm not gonna say "this is it" when the software quality really matters, and I absolutely won't speak to progress (or lack of it) outside of software.
But I will say "you can look around and easily see small businesses using AI to generate posters, quite a lot of small business software and websites are in the same category: the mistakes are real but increasingly don't matter".
>the mistakes are real but increasingly don't matter...
I think it would start to matter once again. People will get fed up of AI posters and art. I think they already are...and once some threshold is crossed, the business won't dare to use AI generated assets/designs.
Turns out humans are much better at recognizing patterns in stuff that is generated ONLY using patterns from human generated content.
> People will get fed up of AI posters and art. I think they already are
Agreed, but will this look like a meme/fashion cycle? If so, re-prompt each year with a different look. Yes, there are still issues here, a friend found an image he was amazed was AI generated, but to me it was obviously so, so I showed him a screenshot of ChatGPT making something just it and included my prompt:
create image: hand drawing of cute springer spaniel puppy looking sideways, various geometric shapes drawn in layer behind and in front of the puppy, all done in style of 7 year old using crayons with mediocre colouring-in skills
As I said to them:
yeah, the line thickness feels AI, to me, the bad colouring-in scribbles feel like just the art style it was propmpted with
it's like: it gets the big picture of the composition, and it knows how to colour in badly, but it doesn't know how to draw a dog as badly as the colouring in
> Turns out humans are much better at recognizing patterns in stuff that is generated ONLY using patterns from human generated content.
We're better at recognising patterns full stop. All biological brains are, and needed to be better than the current state of the art in machine learning because if a living organism was as poor at learning patterns as the SotA in machine learning, the organism would starve to death before being able to pick up anything and eat it.
AI also has a second disadvantage, because there are so few models: the laziest of ChatGPT "thinkpiece" blog posts being everywhere is hard to miss, and 5000 fake bloggers all prompting the same model with "find biggest news story of today and write a blog post about it in a way that maximises my ad revenue" will get 5000 almost identical posts. This will remain true while each instance of the most commonly used AI fail to talk to each other in a way that at least mimics them collectively getting bored with writing the same thing 5000 times, it does not depend on e.g. quality.
It's the exact opposite. Humans will hedonically adapt to anything. If they see enough AI posters pretty soon they will start to not care, at least the majority, while some minority will disagree. We've seen this pattern many times, from data privacy and people still using social media, to using non free software.
>If they see enough AI posters pretty soon they will start to not care..
Wrong analogy. A marketing item wants you to look at it. It takes advantage of an involuntary impulse. A repetitive boring AI artwork will not trigger the impulse to look it. So it is not about "caring"....
I digress with your last quote, I feel like some people who are aware of AI are starting to develop a quasi allergic reaction to slop, and while the mistakes might not matter for most of the population, for others it does and will take notice.
I’m probably illustrating your point but as a FSD fan it really got ”good enough” recently with version 14. The tipping point was suddenly, much more often than not, it can drive end to end from start (my garage) to finish (parked at destination) with no interventions. I can text and watch videos on my phone and as long as I glance up once a minute, it doesn’t complain.
Handling highway driving with lane changes was great when it got there years ago, but just in the last year or so it has gone from a nice to have to “from now on I will never buy a car that can’t do this”.
AI has hit some milestones for replacing work as well. There’s still many more to go and maybe some of them will never get hit (much like I don’t think a coast to coast drive with zero interventions during winter conditions is ever going to happen) but there are points at which it forever meaningfully changes some field of work. I think it’s there for writing code.
> I’m probably illustrating your point but as a FSD fan
Half-and-half. I'm not denying that self driving cars (and LLMs) are improving, I'm comparing it against the standards set by the biggest proponents. But yes, I have heard basically the same thing you just wrote for the previous several major releases of FSD.
Where we agree is that, while you are a fan, you do explicitly give as an example of something you think it will never do, something very close to what Musk has promised:
"Ultimately you'll be able to summon your car anywhere … your car can get to you. I think that within two years, you'll be able to summon your car from across the country. It will meet you wherever your phone is … and it will just automatically charge itself along the entire journey."
(That said, I think Tesla's FSD will never get there, not that it's impossible. The way Musk is behaving, there's going to be a financial scheme named after him in whatever passes for a textbook in 20 years, and it won't be the positive kind of example).
What matters is the benefit it affords to drivers, not what "biggest proponents" say. Elon got overly excited in 2016, but for my day to day usage, I don't need it to self-charge on a drive across the country to get me. I need it to safely drive my family on my daily errands or weekend trips, which I now consider solved.
I'm mentally prepared for the next US administration to exact retribution on his companies, and I expect FSD will be neutered after that. Hopefully other car companies are able to catch up. I'm more of a self-driving fan than a Tesla fan, so as long as the thing works as well as what I have now, I'll be fine with it.
Yes. He was before and remains so to this day, but he did so then, too.
> I need it to safely drive my family on my daily errands or weekend trips, which I now consider solved.
This is probably unwise. As with LLMs, the statistics suggest a spikiness in the intelligence, with it being mostly good but also sometimes still making some very odd mistakes that humans would essentially never make.
As with your other comment, you know that if it gets into a crash you're responsible; while you consider this a win, I suggest waiting until the company you buy the car from (in this case Tesla) puts their money where their mouth is on quality and takes liability for crashes due to the AI upon themselves.
> Hopefully other car companies are able to catch up.
From the stats I've seen, they're much closer to the goal, relatively smoother/less spiky all-round driving intelligence. They may not be as impressive at their best, but when they fail the failure modes are themselves much safer.
The data shows that FSD (supervised) is safer than human alone driving. Given that I value the well-being of my family, FSD (supervised) it is as often as possible. Separately from the data, it is just so nice to hit a button and sit back and relax for the entirety of the drive.
The usual HN nits at this point are that the data is unreliable/biased and that the way I’m using it isn’t supervised enough. I’ll admit to the latter. In my experience the mistakes tend to be navigational but I’ve seen a few of the “nearly drove through the lane closed gate” videos so I definitely keep a closer watch when there is complicated highway stuff going on. I also have my foot at the ready for when I go through the gate to my community, similar failure mode and about 1% of the time it forgets to wait for the gate to close and reopen.
I look forward to the other car companies catching up, competition and more options are good. Elon is a loose cannon, so I need alternatives if Tesla ceases to be.
I would deeply implore you to not put to much faith in this technology as there are many very dangerous edge cases, including as I personally experienced, jumping red lights after coming to a full stop: https://news.ycombinator.com/item?id=48928102
The fact of the matter is that it drives reliably end to end. I don't care who's responsible or what it's called. I only care about the benefit it provides my family, which is immense. If it gets into a crash I know I'm responsible, but given that I consider it a better driver than I am now, this is an improvement over before.
People's right to make stupid decisions end when they reach public roads and/or borrow children to make a point. That's why we have things like driver's licences.
I just spoke to a fried who is a headhunter and who's been trying to automate his processes for a while (he likes to fiddle and certainly has skills, but he's not an engineer). He kept trying, but it just wasn't good enough.
Now he said with GPT Work and Sol, it worked, but the key point is: all of it suddenly worked.
The problem was one of reliability, of handling edge cases. All previous attempts / model-harness-combinations were too brittle and needed too much observation and fiddling - cheaper to do it yourself.
Now he says "I don't know why I would ever hire a recruiter [the folks doing the cold outreach] again. I can focus on the candidate screening and acquiring projects, everything else is fully automated".
This doesn't come from an engineer or an AI lab, but a technically inclined power user, and I think this is where things get interesting.
You heard it from someone with no experience developing software. A lot of AI hype comes from that, even (or especially) from people in the actual business of developing software - a surprising amount of managers and adjacent or supporting roles in the field actually have very little clue about software development.
It's cool that 'regular' people can now create solutions to many small problems, and automate stuff - genuinely a step forward. Like Excel, only vastly better. But for bigger projects, real software engineers know that what LLMs do today is only a tiny part of development. And it solves it in a way that might well make the rest of the lifecycle a lot harder. It's like that saying about tools that make easy things easier and hard things impossible.
IMO it doesn’t replace software engineers, but it certainly changes the job. The problem lies with VPs and executives that rose to their positions through managing phones and printers and shit who don’t appreciate all of the work that isn’t “write code”, and those people are the majority outside of big tech in my experience.
Currently in a re-org justified by AI, AI changing the roles people will need to play. It’s disturbing how much content in the materials about the new org structure and roles and whatnot is clearly ai generated and contradictory. We’re laying off about half of 500 people.
At my employer (a reasonably well known company) every internal communication appears to be AI generated, from emails to slide decks.
It gets better: The internal AI gateway chat thingy where you can ask questions has AI autocomplete that pops up after your write more than 10 characters. What the fuck!?
Who are these "real" software engineers and what are they working on? Among software developers I know, the biggest users of LLMs are the seniors. I don't think any of them would agree that its usage is a tiny part of development.
If you're a high-end executive, people would not approach you directly but try to find a connection to you (which also serve as references).
For all other roles the first few messages are similar: this is <role X> with key challenges a,b,c. You seem to be a good fit because of d,e,f. Would you be open to explore this? This requires relocating to <place>.
Traditionally, this was done by entry-level people. In either case, this isn't the person who will jump on a call with you.
"Desperate" and "clueless" seem very strong words here in reference to someone who gets actively approached from a recruiter.
Then it just becomes a new baseline (everyone have access to the same LLMs), and recruiting moves up the philosophical ladder where human can add more value. What will it be? I don't know, I'm not a recruiter.
Again I've heard this since 2022 when gpt3.5 came out.
This is like microprocessors in the 80s. Sure they double in capability every 18 months but the start is so pathetic it will be 30 years before they are good enough for everyday tasks.
CPUs in the 90s were amazing! They were over-specced for "everyday tasks." Our problem is that we overbuilt CPUs too much, so software is now written with ten unnecessary layers of abstraction because there's no real reason to simplify.
3gb of ram aren't enough everyday tasks at any sane resolution. That we still don't have 1200 ppi desktop monitors is as stupid as using black and white screens in 2000.
It seems to be true this time though; I have observed it myself and heard it from several experienced developers I personally know and respect. It feels like some threshold was crossed with Opus 4.5 and Gpt 5.3, where the models are now able to reliably solve certain classes of problems that were previously unreliable.
Time will tell of course, and it’s early, but inflection points do exist with progress.
Thing is. You can find an extremely similar paragraph written about Claude 4.x or some equivalent gpt. And simultaneously, many people expressing their frustration and the shortcomings of <insert any model>
“But it’s different this time” - several people, several times over the last couple of years.
This is not at all a dig at you, I’m very sorry if it reads that way. My point is these things only get truly better in anecdotes. The ways in which they fail is yet to change. Just yesterday I had gpt 5.3 generate completely awful code for the Cinema 4D Python API. Also an anecdote. But for all of the people saying they are truly intelligent and truly reason, they still make obvious mistakes, write around problems, fail entirely at architectural decisions, fail at random, generate FAR too much code.
And no amount of harnesses, methodologies, loops make much of a difference. If you listen to people on the internet they say it’s all working. You listen to people on the job and they mostly say it’s creating tech debt and a review bottleneck. Also burnout, so much burnout.
I think LLMs are mediocre. I think it’s fine they’re mediocre. You can work with low expectations. But the hype cycles are so tiresome.
I don't buy into the huge LLM hype but I certainly think late 2025 was the inflection point. Up until then, I thought LLMs universally sucked at code. GPT-3.5, GPT-4, o1, o3, Claude 3.5, Sonnet 4, the whole bunch. Each iteration got marginally better, hallucinated APIs less and so on, but my overall evaluation of them all was that they wrote crap code and were unsuitable for anything other than one-off scripts.
Then the incremental improvements did, in my experience, cross some kind of threshold in late 2025 where the things became useful. It is of course anecdotal and personal judgment. But I asked LLMs to implement a small feature in my codebase (my usual test) and finally it produced code I was happy with. They've also been able to locate and diagnose a problem based on logs. In my view it's now a markedly different level of capability than we had a year ago, though I would call the previous two years equally useless.
You can find an extremely similar paragraph written about Claude 4.x or some equivalent gpt
Yes, as a product gradually improves there will always be many people for whom version X didn't work well for them and version X+1 does. It turns out that Opus 4.5 and GPT 5.1 were larger than average improvements that cross that threshold for a significant number of people.
My point is these things only get truly better in anecdotes. The ways in which they fail is yet to change.
If your claim is that there's no substantive difference between Sonnet 3.5 and Fable, then we live in very different worlds.
LLM capabilities are spiky. They're amazing at some things, and poor at others. Over time, the set of things they're amazing at has grown, while the set of things they're poor at has shrunk. If you think LLMs are just "mediocre" without any nuance, that's a sign you haven't spent the time to evaluate them in order to make an informed opinion.
> You can find an extremely similar paragraph written about Claude 4.x or some equivalent gpt
Even 3.7. I remember when it came out and people were claiming that that was now the model that was going to replace engineers. Cue Fable years later and people still claim that this one is the one.
Not shape, but even with napkin math, you can work out the conclusion. Take email vs snail mail. Online ordering vs relying on catalog. The spreadsheet, the word processor, the various CAD software.
The key thing is that it's easy to contrast the old way vs the new way and the evidence become obvious.
The thing with LLM tooling is that they're not reliable. I can do fine with risks, but only when there's a way to manage it so that if the disaster happens, it's practically a black swan event.
Typing more code or solving one task has never been the core problem. The core problem has always been to encode a whole system into the computer AND then provide a control interface for it. It requires both an understanding of the system you want to encode (especially how it behaves over time) and empathy to know what would be the best control interface for the users.
That understanding does not rely on the amount of code, and the best control is found through communication.
An understanding of the system could be the following: A conference schedule consisting of events (time, place, speaker, description,...) stored or presented in some format. The interface would be: A web app with a mobile first UI that presents the information in an accessible manner (highlighting, filtering, exports,...).
A relatively quick (I haven't tested it), would have been to open the web inspector and extract the data using the dom API (requires knowledge of the dom api and a desktop browser), put the data into some json or a tsv file, then write a php script or a python script and then serve that. The interface could have been built with the standard elements of some css framework (bulma?).
Not saying the above is better. But the thing is that is doable from even a raspberry pi. And more it's repeatable and extensible. And the individual piece of knowledge are reusable in different situation.
Okay thanks. But what I'm saying is all of these anecdotes should add up to something measurable if there's something to the idea that an inflection point was reached in November 2025, right? There's a lot of marketing and hype and motivated reasoning going on, so that's why I don't trust broadly-reported notions.
Con: "Where's the evidence that this actually is improving productivity?"
Pro: "The evidence hasn't shown up in statistics yet; it's too new!"
Con: "And won't this destroy maintainability?"
Pro: "Show me the maintainability disasters caused by AI."
Con: "I can't yet; it's too new!"
Both sides are playing the "it's too new" card when asked for actual evidence to prove their claims. In fairness, it actually is too new for there to be much statistically-valid data, especially if the inflection point was November 2025. So both sides are trumpeting their position, neither with actual trustworthy data.
Everybody has their anecdote. Nobody has data yet.
I really don't think that's how it works. Smart, experienced developers who thought coding agents were junk for most of 2025 and think they're useful now in 2026 are not saying that because they got "worn down over time".
It tends to be when the training data wanders into their area of expertise temporarily and they go “OMG, they hype is real. I was so wrong” and then a few releases later they’re on the train and furious that the skills in their domain space have not just stopped improving, but regressed. Cue someone else in a different part of the world starting the same cycle.
Meanwhile the guy who leaned in a year ago and gave up reading the output is beginning to see work grind to a halt and throwing more agents at it is increasingly not working.
You can see these tropes all over social media near constantly.
> November was, for me and many others in tech, a great surprise. Before, A.I. coding tools were often useful, but halting and clumsy. Now, the bot can run for a full hour and make whole, designed websites and apps that may be flawed, but credible. I spent an entire session of therapy talking about it.
> The real annoying thing about Opus 4.6/Codex 5.3 is that it’s impossible to publicly say “Opus 4.5 (and the models that came after it) are an order of magnitude better than coding LLMs released just months before it” without sounding like an AI hype booster clickbaiting, but it’s the counterintuitive truth to my personal frustration. [...] A year ago, I was one of those skeptics who was very suspicious of the agentic hype.
> Six months ago, in an episode of the Lex Fridman podcast, David shared how he doesn’t use AI tools to write code: he types out all his code. But things have changed a lot since then.
> In this episode, we discuss his approach to building software, how it’s changed in the last six months, and why he now takes an agent-first approach, and how he barely writes any code by hand.
> There are other questions around AI (like what the economy of it will actually look like in the end), but "is it useful" is no longer one of those questions. Anybody who doubts that clearly hasn't actually used it.
> Shock! Shock! I learned yesterday that an open problem I'd been working on for several weeks had just been solved by Claude Opus 4.6 - Anthropic's hybrid reasoning model that had been released three weeks earlier! It seems that I'll have to revise my opinions about "generative AI" one of these days. What a joy it is to learn not only that my conjecture has a nice solution but also to celebrate this dramatic advance in automatic deduction and creative problem solving.
Okay so you have no names of experienced developers saying that coding agents are now useful when they weren't before. The only two you have actually saying that are... not working as software engineers.
> 2025 was a weird year for me. If you had asked me exactly a year ago, I would have said I thought LLMs were amusing toys but inappropriate for real software development. [...]
Today, I would say that about 90% of my code is authored by Claude Code.
> I don't think you are arguing in good faith here.
Funny coming from a guy using Linus saying "There's no doubt AI is useful today" and interpreting that as "Yeah last year AI coding agents were garbage but now they're really useful".
There's no denying AI is useful. I use AI, my project at work is all about using LLMs. They're even useful for coding, when I don't give a fuck about the quality of the output (which is almost never in a work scenario).
But to claim that there has been some magical shift late last year is simply laughable. Anyone working as a software engineer and claims that is either lying to themselves or simply bad.
Oh and don't try with the "you're not prompting it right".
The world needs to SEE what is coming out of all this spending. Guess what? It aint much! Thats why firms are in a hurry to review their expenditures and switch to chinese offerings.
I hate to break it to you - but wake up. You are not living in reality.
I didn’t start using Claude Code until late 2025. Prior to that I would use ChatGPT to give me snippets of code but I was still doing most of the actual code writing. Coworkers told me in late 2025 about how they hadn’t written a line of code in “months” and just use Claude Code/agentic “whatever” so I tried out Claude Code and was pleasantly surprised. It is passable to have entire apps written by LLMs (I’ve made several that I otherwise never would have had the time to create by hand), but I wouldn’t say maintainable or easily extendable. It’s hard to be specific, but there’s something about LLM code that doesn’t look “natural”, and I’m not talking about the excessive use of comments in code. The code itself is unnatural. Functional, but unnatural. I wouldn’t want to suddenly lose LLMs and have to read through and understand and continue enhancing a codebase created by an LLM.
For me it feels a lot like generated images or video. I've made lots of things now, but those that are 100% LLM written "work" but are uncanny, weird and the details are wrong everywhere you care to look in detail.
This has been my experience has well. For code generation, you have to really constrain these LLMs on your coding style, design goals, test cases, and overall expectations. I mainly use these tools to help my understanding of the code and to generate code for very specific problems. Even after all that setup and careful review I'd say it's still a net big speed up for certain software engineering tasks.
Jason Turner gave an excellent talk at last year's CppCon explain how he thinks tools can be used to make generative AI coding assistance safer and more productive. https://www.youtube.com/watch?v=xCuRUjxT5L8
I think how I'd put it is like, it's a deterministic common denominator thing iterated on, but it's not really intuitive, it's not really "what makes sense" but rather "what would x look like", similar to how you point out it is when it comes to image generation
it's useful for scaffolding but after that I'm not sure how you could rely on it without being in the loop and directing how the code should be like
I was getting useful coding work done with GPT 3.5. I think devs saying "the models are finally good enough" this year are just trying to save face from their own previous irrational denials.
Useful, yes, sometimes, but it wasn't fully automated "Here's our JIRA board URL, fix everything that's rated 1-3 story points and in the current sprint".
Since so many people are doubting you here, here is a post from ~a year ago that's pulling the same "LLMs 6 months ago were crap, now they're awesome" shtick: https://fly.io/blog/youre-all-nuts/. There's more posts along this vein being put out from 2024 on or so.
The idea that "yeah, you said technology had improved in the past, and now you're saying it has improved again" is a gotcha just seems incoherent to me.
I think people are, rightly or wrongly, identifying goalpost moving re: coding agents are only just this year working well enough to impact work. I'm reminded of a quote I read last year from one of my favorite writers, Noah Smith:
> The debate over whether AI is taking people’s jobs may or may not last forever. If AI takes a lot of people’s jobs, the debate will end because one side will have clearly won. But if AI doesn’t take a lot of people’s jobs, then the debate will never be resolved, because there will be a bunch of people who will still go around saying that it’s about to take everyone’s job. Sometimes those people will find some subset of workers whose employment prospects are looking weaker than others, and claim that this is the beginning of the great AI job destruction wave. And who will be able to prove them wrong?
The context for some of these posts are yet-another-null-result coming out, saying that AI isn't having the expected objective result, with the posts defending AI by saying that it totally changed just after the cut-off the research. Evolutionary step changes are definitely a thing, but the problem isn't that AI is having such step changes, it's that you're saying that the most recent step change is the one that shows the improvement, and every study that comes out saying "no, that wasn't," the response to that is, "no, there's a new change that is totally the key inflection point!"
Or, put differently: if a study comes out next year saying they don't see major impacts from AI in 2026, will you admit your viewpoint as being wrong, or is your response going to be "no, there was a massive inflection point in September 2026 that completely invalidates the paper"?
I keep running into this as well. Someone shared a link to a study from last year telling me that AI doesn't make programmers as productive as they think. The study was obviously done months prior to it's release. So evaluating the state AI even further back.
But this didn't seem to concern them. The study said X, therefore it applies to today.
I'm just not sure it's worth it to argue with others about it at this point. Not that I'm 100% all behind AI coding, but I'm just shocked people are still this resistant.
> I'm just not sure it's worth it to argue with others about it at this point.
This is the correct response. Let closed-minded people do their thing. Makes them less competitive against you. There's nothing for you to gain by trying to help them understand what they are missing.
A year ago it was capable but required much more hand-holding. I thought it was as good as it would ever get and was still happy and productive, but only because I didn't know how much better it would get.
Nobody was saying coding agents started working in 2023 or 2024, because the category was defined by Claude Code which was first released in February 2025.
I would say that Aider is what defined coding agents. That was at least multiple months before Claude code. I remember seeing a coworker use aider for a hackathon project Adeline November-December 2024 , and it was already decent and pretty close to the DX we consider coding agents to have
Devin and Cline were the original true "agents" IIRC. Aider was the original terminal "agent" but it was't a true agent, it was sort of a hybrid agentic chat, since it had limited recursion ability (3 turns by default) and you had to configure it carefully to make it consistently take multiple turns in a row.
The timescale is well established: Late '25 was the start of agentic ai when capabilities of model + api + scaffold reached autonomous state. Any study comapring events before that timeframe is comparing apples with oranges.
The quality of the output is so variable. It depends on the model, “effort level”, prompting, probably even the programming language/app functionality, and libraries involved. For example, I find LLMs are best at making simple web apps. These web apps, while simple, would still take a senior engineer perhaps a week or two to create, but LLMs can spit them out inside of an hour. Conversely, LLMs struggle with things like Docker or local model stuff. Parallelization of code is a mixed bag. In these areas I think it often would have been faster for me to write the thing by hand.
I'll give a recent example: the past month I had been using Codex cloud and it was working great writing Rust modules. Suddenly last week they made a few changes and everything went to shit, couldn't get a single good result out of it. Same text box, same prompting, same model, totally different results because they changed the cloud tool's resource limits.
So whereas before my debugging time would have been spent in Rust docs looking up traits and such, these days I feel more like an AI therapist trying to figure out why it's not feeling up to task on any particular day. It can be anything from regular service outages to geopolitics that on any given day my workflow is fucked up.
In my before-AI workflow, I was never restricted from compiling Rust code because of concerns that Cargo is a national security threat. So you really have to broadly scope the notion of "reliability" with these AI tools; it's much larger than whether it can give a good output but whether it can do so consistently enough to depend on.
"unless you're content with producing low quality work." - With the right guiding hand, it is a productivity multiplier without compromising quality. As a fully autonomous developer, it is a disaster.
> With the right guiding hand, it is a productivity multiplier without compromising quality
This just reads like another variation of “it’s the user not the tool,” which is just endless runway for always blaming people and never acknowledging the limitations of LLM’s.
I’d be curious to hear how the recipients of your work enabled by the “productivity multiplier” feel about the quality.
You can't play an entire orchestra's sheet music on a single guitar either, but your playing ability still matters a lot.
I would say that as of July 2026, with the right scaffolding, you can get reasonably good output out of a LLM, or better a combination of LLMs. For example, it pays off to prepare an implementation plan with one LLM and then let another LLM check it for flaws, then again. After several iterations like this, you will have a plan better than whatever you could come up with yourself.
It often is the user and not the tool. LLMs are complicated, have nontrivial failure modes, and the user needs to steer them carefully. They might be the most complicated tools on the planet right now.
Anecdotally, the recipients of my work have become visibly more happy in the last months. LLMs are great at diagnosing subtle problems which tend to appear at Friday night only, and this is the sort of problem that bugs actual people the most.
> I would say that as of July 2026, with the right scaffolding, you can get reasonably good output out of a LLM, or better a combination of LLMs.
Totally agree, I don’t think I said or implied otherwise.
And yes can it can be the user and often even is, but when it comes to any LLM conversation I’ve been a part of it seems people think the only answer is “you’re using it wrong.” Evangelists swear it’s a 100x multiplier and anything counter to that means you’re either a Luddite who is blinded by politics or are too dumb to use the tool.
How are junior devs becoming qualified “guiding hands” these days? If the expert with LLM assistance is multiplied, what’s a company’s incentive to pay for a junior, and how would they train to get good in these conditions?
Juniors are still useful, but not as pure "coders". The juniors you want to hire now can wear product, engineering and QA hats, and are self starters with good attention to detail and AI skills. These types of people are actually a bargain.
I produce code that is significantly higher quality with the assistance of coding agents, because I no longer succumb to the temptation to cut corners due to lack of time.
One example: everything I do is properly tested and documented now, even the most trivial of changes. Previously I would have weighed those tradeoffs and sometimes decided not to bother with the tests because they weren't worth the time.
> because I no longer succumb to the temptation to cut corners due to lack of time
No offense, but that says much more about the way you approach programming than about the quality of LLM outputs.
In my experience, LLMs are the ultimate corner-cutting tool. With LLMs, I now succumb to the temptation to cut corners, build something I haven't properly researched and don't fully understand, prioritize shipping quantity over quality.
Without LLMs, I have to understand the domain and the tools and ultimately my full solution (with all its warts and limitations). When I really care about the project and consider it "my baby", LLMs are out of the picture.
Software engineering is about tradeoffs. The single most important tradeoff is always time: this feature would be better if I did X, but doing X would take me Y hours, is it worth the time or should I spend that time on something else?
Not having to type in and then iterate on the code manually has a material effect on those time calculations.
I absolutely agree that you need to understand the domain and code and tools. That's what I want to spend my human time on - not typing in the code and sweating over every line of syntax.
I agree, good software engineering is largely about making good trade-offs.
However, I completely disagree with your characterization of LLM coding as saving you "typing in the code and sweating over every line of syntax".
Claude will happily architect complex layers of OOP, multithreading, SIMD (a "personal favorite" of both Claude and ChatGPT).
If you don't immerse yourself in the details, you will never know if you made the correct engineering trade-offs, let alone whether the various "optimizations" and clever solutions by the LLM actually work.
I think you're referring more to product strategy trade-offs than engineering trade-offs here; so yeah, getting something you don't understand out the door is often the fastest path to sales. This is also how we get so much low quality and unnecessary software out there (it was true before LLM coding, just amplified now).
I think a reasonable tradeoff, in theory, would be: use the LLM to help you explore the space of solution, then dive into the details and choose the path that feels right. However, I highly doubt most people do that. It's also easy to let the LLM's first solution influence your understanding of the problem, so that you will never truly explore novel solutions that fall outside of the most likely token sequence prediction.
Oh and LLMs when properly constrained can also do pretty good "auto-complete on steroids", which can definitely save you typing a lot of boilerplate. If that's how you use them, then your characterization is correct.
> I think a reasonable tradeoff, in theory, would be: use the LLM to help you explore the space of solution, then dive into the details and choose the path that feels right.
Oh absolutely. One of my favorite prompting patterns is "suggest options for X", followed by serval rounds back and forth to discuss tradeoffs and maybe introduce new ideas into the mix.
I also use LLMs to build exploratory prototypes. I'm working on a project today which I've had different coding agents build a total of four times already. Those explorations gave me the confidence to dictate a finished solution that I'm sure will work well against all sorts of future cases.
Here's the rub. They produce low quality work according to your rubric. If that was the universal rubric, they would already have RL'd against it, and you'd like the work they produce.
I wouldn't say it like that exactly, though you're not wrong either. People just heavily discount the value of future efficiency gains due to nuanced architectural/stylistic choices. I think this is in part because people have a limited expectation that the code will survive long enough for it to matter, or that they'll be at the company long enough to reap the rewards of the up front work.
Claude 4.5 was it (nov 2025?), without a doubt. It went from frequent hallucinations to highly usable with much less garbage output. If you were making demos of AI tools around this time your demo/pitch/product was saved and you probably looked like a genius.
The capex is still "early" for sure (i.e. data centers are still being planned and built out, we're hardware/energy constrained).
If model scaling holds out, we're "early-ish" in terms of the reliability and performance of these systems, just based on utilization of the compute from the planned capex. If we hit hard diminishing returns and we don't find architectural/data workarounds, that would put a wrinkle in things, but I suspect that the AI we have now is capable of helping us find those workarounds and keep things moving.
I haven't. Around the start of 2026 is pretty widely mentioned as when they went from "this is broken slop" to "huh this is actually 90% what I would have written", which matches my experience.
Anecdotally I'm seeing a lot more recruiter activity/interest now than this time last year.
But it seems more correlated with hype-cycle-stage than anything else. Right now a lot of founders seem to be convincing a lot of VCs that they can make $LOTS by replacing/changing $BIG_INDUSTRY/$BIG_PRODUCT with an agent-first blah blah replacement, and then using that money to hire more people to manage/execute/coordinate the coding agents...
Last year, by comparison, there seemed to be a mood of "software will stay the same but will require less people" while right now there's a lot of hype around "we can build different types of software or build it in different ways" and those early-stage things are in growth-mode. That guarantees nothing about how many people they'd need in the future, or their success at all, ofc.
The news that I'm getting from contacts in non-startup-land is a bit different - still layoff threats. Still pressure to use AI tools more. Mixed confidence on whether or not longer-running "agent" modes are that much more effective-without-breaking-things in legacy code if not used with care.
Is the recruiter activity you see very… Human? I’m getting a lot of emails but when I do my research, it all appears to be bots. I don’t trust any of it and I don’t respond.
However, companies have been using AI as an excuse for layoffs since well before January 2026, which corroborates the study's conclusion. (Source: https://layoffs.fyi/ai-layoffs/) There is certainly an uptick starting 2026, but that could be explained either by AI actually causing more layoffs, or by AI becoming an even better excuse for layoffs.
This isn’t the first study showing this though. It’s pretty simple, programmers and IT were severely overhired during the pandemic, there are massive job losses now, and it’s easy to blame AI when in reality there are a lot of economic factors and AI isn’t increasing productivity as much as anyone would think.
Maybe the future will change that for very specific things, but I think people should be learning and preparing for that, which isn’t any different than what everyone has been told in every job market since the start of the Industrial Revolution.
I personally hope that AI continues not to result in a noticeable negative impact on employment and that pandemic over-hiring turns out to be the major factor for all of the layoffs.
I'm nervous that the studies which show that so far don't seem to be taking the 2026 improvements in coding and general agents into account.
What the heck would I be "stealth-marketing" here?
It's genuine concern. I do not want to live in a dystopia where AI results in mass unemployment. That would suck, even for the people who manage to stay employed.
> What the heck would I be "stealth-marketing" here?
Cynically, "thought-leaderishness".
I've been spending a lot of my time these days outlining why we can't "just make an agent for it" to CEOs who read blogs like yours. They can't distinguish a production system from a quick HTML tool from a guy whose job doesn't depend on it working.
One of the goals of my blog is to help CEOs who read it not make stupid decisions. I aim to be a counter to the breathless LinkedIn hype they are exposed to everywhere else.
No offense but wouldn’t being an AI booster run contrary to this goal? I imagine being an Ed Zitron type would be more conducive with that, albeit dishonest if you don’t genuinely believe it.
> I think this is going to have a huge impact on society. My priority is trying to direct that impact in a positive direction.
> It’s easy to fall into a cynical trap of thinking there’s nothing good here at all, and everything generative AI is either actively harmful or a waste of time. [...]
> I’m going to continue exploring and sharing genuinely positive applications of this technology. It’s not going to be un-invented, so I think our priority should be figuring out the most constructive possible ways to use it.
That's been my focus ever since: try and use whatever influence I have to nudge things in a positive direction, and help people figure out how they can use this technology in a way that is beneficial both to them and to other people.
>programmers and IT were severely overhired during the pandemic
1) I hesitate to believe that losses were disproportionately technical roles as opposed to administrative.
2) Over-hired by what metric? It's well known that hiring never fully recovered after the GFC; was the recruitment post-pandemic just bringing us to parity with where we had been 20 years earlier?
Not to say that I disagree with your following point. The AI overspending and the layoff cost-cutting are not in a direct causal relationship; both are rather symptoms of a common corporate pathology.
They were hired at a higher rate than anytime since the 90s. In the US they ballooned by almost 50%. So when you take layoffs in consideration, the US is still has 25% more programmers than it did prior to the pandemic. And looking at the hiring rates and who is getting hired in programming, it looks like companies aren’t hiring entry level basically at all at this point, so shows the market is still over saturated. No wonder coding camps and CS majors are not being pushed any longer because it’s not a great market for them.
> This category of software may have a much more meaningful impact on work than the mostly-chat systems we were using from 2022-2025.
Mostly an impact on software development - I'm not seeing broad automation and layoffs in industries like law, finance etc. It will gradually happen but due to issues with memory, reliability and long term planning of LLMs there are real barriers. Even in software development - while it has completely transformed the field I don't think many people still believe we won't need devs in 2027 or that their amount will shrink by 50%.
I'm optimistic that coding agents will not obsolete software developers - my own experience is that coding agents have made my work harder, because the scope of projects I can take on has increased.
My ideal version of all of this is that nobody loses their job and everyone gets to take on more ambitious projects.
I'm not quite naïve enough to assume I'm right about that though!
> my own experience is that coding agents have made my work harder, because the scope of projects I can take on has increased
Likewise. I find the types of problems I need to tackle and the challenges they represent are actually quite exhausting, too. And the agents unblock me relentlessly so I’m constantly pressed to do relatively difficult things. Either that or code review. I’m hoping it’ll only be an adjustment phase but this is the most challenging my career has been in over a decade.
I think other industries are late, but also will see less productivity gains. In software engineering agents have the benefit of very short feedback loops that provide extremely actionable insights which allows the agent to very quickly iterate towards the final solution. This is much less the case in other industries such as law and finance I think. The first pass of the agentic loop is usually something that barely works (you don't even get to see it with the latest models). If that first pass is the primary output in other fields with no tight feedback loops, then the productivity boost will be much lower than in software.
For whatever anecdotal evidence it's worth - that has been my experience as well. For the first time last December I noticed the harnesses performing like actual workers.
Everything impressive has happened in the last six months.
And this is why the labs cannot just "stop training and become profitable". I can't imagine they would like it if studies like this will actually become credible.
"Move fast like a blur so people can't see that you have no clothes"
Any research on the impact of AI would have been lagging indicators. It's not the researchers' fault[0], but the nature of a field moving at neck breaking speed. Remember there was a paper saying programmers were 20% slower with AI?
That early 2025 METR study was particularly interesting because participants self-evaluated themselves as 20% faster, but the measurements showed they were actually 19% slower.
All the reports of productivity since then are self-reported, or using questionable measures such as SLOC and PRs, so it’s reasonable to say that productivity improvements are still unknown.
Unfortunately, METR hasn’t been able to replicate the study because they couldn’t find enough willing participants.
> Remember there was a paper saying programmers were 20% slower with AI?
Key thing was not that they were X% slower, but that they were slower while being convinced they are faster. Of course, any analogies with the current hype cycle are completely unfounded.
Well, in addition to that you also have to consider that companies are now paying (increased) API pricing. It would be an understatement to say that my clients in the F100 range are skeptical at best regarding the gains they’ve seen compared to the costs.
This has led to many of them instilling dollar limits or demanding proof of increased productivity (not just output) with the implication being if you don’t provide value with it it’s getting taken away.
So that is to say, if they aren’t happy with the price now, how will they feel when it goes up again compared to just keeping a certain headcount?
Oh wow. You just hit something there. There are government programs for R&D where I live. Grants. Tax credits. This sort of thing. I wonder if people are trying to expense to those programs too.
>This has led to many of them instilling dollar limits or demanding proof of increased productivity (not just output)
They should have done that from the beginning - demanding proof of increased productivity - if that was their goal. otherwise they were not using their brains well enough.
And you doubly don't want to work with them, first because they confused output with productivity at first. and second, because they're parroting the productivity metric.
You only need one guess for whose pockets the productivity benefits go into.
When I was a teenager, a friend had an old gas guzzler from the 70s. I live in a rural area. One time, my car broken, he drove to pick me up to go to College.
This cost him an extra $40, in today's dollars. No, I'm not joking. That thing ate gas like a dry camel drinks water.
This is what Fabel5 feels like. Crazy expensive. 10 minutes work pulled almost $80 is usage credits yesterday. I'd be exceptionally skeptical too, on costs, if I still had the DEV I had last week, but they were also eating that kind of cash on a very-improved, but still used as a linter.
For $200+/hr, or ~$400k/year, I'd want to see a tripling of output at least. In a lot of US markets, you can hire 3 junior devs for that.
Yes, there are cheaper options. Opus, etc. But it's really over-priced, and frankly I think the real gold now is making open models fully functional. Anyone predicating their business upon tie-in with the big boys is just going to fail, hard.
This. Also I'm finding out that, after being blown away by agent mode lately, non-agent mode still kind of sucks across frontier models. Using GPT and Gemini in non-agent mode is asking for inaccurate information confidently presented as the truth. Turning on agent mode fixed a lot of that for me.
I think it's because of the lack of feedback. Humans also can't do much without feedback. E.g. I doubt most people could write 100 lines of code that works first time without even compiling it once.
Even more extreme than that, if you were working on a large code base where compiling it took forever, or you needed to rely on a very slow CI pipeline, it became very important to git gud and write shit that worked first time in order to deliver when promised. I'd argue if you've never had one of those moments where you sunk a bunch of time into some changes and it all compiled/worked flawlessly the first time, you're missing out.
Indeed. I wonder how much AI will do in industries with longer feedback loops that are much noisier. I think the field of software is a unique field in that sense.
Ah, Simon with another load of bullshit! Your own blog articles contradict your comment. But sure, keep moving the goalpost for your overlords Anthropic and OpenAI.
Also in July 2025 I was successfully using one of the earliest async cloud environments - Codex Cloud - to scrape websites.
By October 2025 the most recent Claude could one-shot a simple Datasette plugin.
What changed in November 2025 was that the level of hand-holding needed to get good results dropped to the point that it was no longer credible to deny their utility.
> Can't wait for your Summer 2027 comment where you'll claim that coding agents really only became usable in November 2026.
Oh! I think I see what happened here. I said this:
> coding agents (Claude Code, OpenAI Codex) only started working really well in late November
"Working really well" and "only became usable" are not the same thing.
Your detractors don’t care about the facts Simon, as you’re probably well aware. They’re afraid of LLMs and need to deny that they can do anything well or have any potential to take their jobs.
In their world AI never has an impact on jobs, just like AI can never be an equivalent to a human, and the goalposts move every time a goal is scored.
And yet I have felt a general code quality decrease and overall enshittification of software products since 2023 when people were already using copilot. Or I am just biased to use that to justify any overengineered piece of shit code with that because I refuse to think any sane person would come up with such contrived code and I am looking at the wrong places.
General agents (OpenClaw, Anthropic Copilot, ChatGPT "Work") started working even later than that.
This category of software may have a much more meaningful impact on work than the mostly-chat systems we were using from 2022-2025.
Studies that mainly focus on 2022 to end of 2025 might be missing out on a material uptick in capabilities.