- Development velocity is very noticeably much higher across the board. Quality is not obviously worse, but it's LLM assisted, not vibe coding (except for experiments and internal tools).
- Things that would have been tactically built with TypeScript are now Rust apps.
- Things that would have been small Python scripts are full web apps and dashboards.
- Vibe coding (with Claude Desktop, nobody is using Replit or any of the others) is the new Excel for non tech people.
- Every time someone has any idea it's accompanied by a multi page "Clauded" memo explaining why it's a great idea and what exactly should be done (about 20% of which is useful).
- 80% of what were web searches now go to Claude instead (for at least a significant minority of people, could easily be over 50%).
- Nobody talks about ChatGPT any more. It's Claude or (sometimes) Gemini.
- My main job isn't writing code but I try to keep Claude Code (both my personal and corpo accounts) and OpenCode (also almost always Claude, via Copilot) busy and churning away on something as close to 100% of the time as I can without getting in the way of my other priorities.
We (~20 people) are probably using 2 orders of magnitude more inference than we were at the start of the year and it's consolidated away from cursor, ChatGPT and Claude to just be almost all Claude (plus a little Gemini as that's part of our Google Whateverspace plan and some people like it, mostly for non-engineering tasks).
No idea if any of this will make things better, exactly, but I think we'd be at a severe competitive disadvantage if we dropped it all and went back how things were.
It's all romantic, but a bunch of devs are getting canned left and right, a slice of the population whose disposable income the economy depends on.
It's too late to be a contrarian pundit, but what's been done besides uncovering some 0-days? The correction will be brutal, worse than the Industrial Revolution. Just the recent news about Meta cuts, SalesForce, Snap, Block, the list is long.
Have you shipped anything commercially viable because of AI or are you/we just keeping up?
Has it occurred to you that there might not be a correction, and that the outcome would still be brutal, at least on par with the industrial revolution.
It's physically impossible to build out the datacenters required for the "AI is actually good and we have mass layoffs" scenario. This Anthropic investment is spurred on because they've already hit a brick wall with capacity.
$40B goes a long way, but not for datacenters where nearly every single component and service is now backordered. Even if you could build the DC, the power connection won't be there.
The current oil crisis just makes all of that even worse.
The next level of layoffs is probably still 25 years out.
Hasn't even been 25 years years since the previous layoffs before the current ones.
But all the economic indicators suggest those are "bad economy" layoffs dressed up as "AI" layoffs to keep the shareholders happy.
And that's without accounting for the various wars (and resultant economic impacts) that are already in progress. A large part of what drove the meat grinder of WWI was (very approximately) the various actors repeatedly misjudging the overall situation and being overly enthusiastic to try out their shiny new weapons systems. If one or more superpowers decide to have a showdown the only thing that might minimize loss of life this time around is (ironically enough) the rise of autonomous weapons systems. Even in that case as we know from WWII the logical outcome is a decimated economy and manufacturing sector regardless of anything else that might happen.
I think that just means the relative civilian loss of life will increase once again.
russia is really and empire of the dumb and subjugated serfs at this point (again, history repeats), but they are far from only such place.
Dont expect more, most people are not that nice when SHTF.
Bubbles like the AI bubble are a game theoretic outcome of a revolution. Many players invest heavily to avoid losing, but as a whole the market over invests. This leads to a bubble.
But right now, the difference in developer experience between a dev on a team at a business which has corporate copilot or Claude licenses and bosses encouraging them to maximize token usage, vs a solo dev experimenting once every few months with a consumer grade chat model is vast.
Meta seemingly has a constant stream of product managers. If llm’s really augment the productivity of engineers, why isn’t meta launching lots more stuff? I mean there’s no harm in at least launching one new thing.
What are all those people doing with the so called productivity enhancements?
What I’m calling into question is how much does generating more code matter if the bottle neck is creativity/imagination for projects?
The only thing I’ve seen is a really crummy meta AI thing implemented within WhatsApp.
Only solution I can think of is to drastically cut headcount so productivity is back to prior levels, and profitability is raised. Big Tech is mostly market constrained with not much room to grow beyond the market itself growing.
As for startups, seems like AI tools have drastically reduced their time to market and accelerated their growth curves.
Hobbyist solo dev, counting tokens, hitting quotas, trying things on little projects, giving up and not seeing what the fuss is about.
vs
Corporate developer, increasingly held accountable by their boss for hitting metrics for token usage; being handed every new model as soon as it comes out; working with the tools every day on code changes that impact other developers on other teams all of whom have access to those same tools.
I might be missing a lot of self-evident assumptions here but I feel like I'm still missing so much context and have no idea what this difference is actually describing.
I'm talking more about why threads like this seem to be full of people saying 'this has completely changed how corporate development works' and other people saying 'I tried it a few times and I don't get the hype'
My impression has always been it's more important the build the correct thing (what the customer needs/wants) rather than more stuff faster.
The process of learning what the customer needs/wants is a heavily iterative one, often involving throwing prototypes at them or betting at a solution, then course-correcting based on their reaction. Similarly, the process of building the correct thing is almost always an iterative approximation - correctness is something you discover and arrive at after research and prototypes and trying and getting it wrong.
All of that benefits from any of its steps being done faster - but it's up to the org/team whether they translate this speedup to quality or velocity. For example, if AI lets you knock out prototypes and hypothesis-testing scripts much faster, you can choose whether to finish earlier (and start work on next thing sooner), or do more thorough research, test more hypothesis, and finish as normally, but with better result.
(Well, at least theoretically. If you're under competitive pressure, the usual market dynamics will take the choice away, but that's another topic.)
why do you think restaurants rarely change their menus.
Thats just one set of costs but a good starting point.
It's an absolute tornado of PRs these days. Everyone making the most of these tools is effectively an engineering team lead.
I’m making a team version of my buildermark.dev open source project and trying to learn about how teams would like to use it.
Backends handling tens to hundreds of thousands of messages per second with extremely high correctness and resilience requirements are necessarily taking a different approach to less critical services that power various ancillary sites/pages or to front end web apps.
That said there's a lot of very open discussion around tooling, "skills", MCP, etc., harnesses, and approaches and plenty of sharing and cross-pollination of techniques.
It would be great to find ways to better quantify the actual value add from LLMs and from the various ways of using them, but our experience so far is that the landscape in terms of both model capability and tooling is shifting so fast that that's quite hard to do.
It hardly seems worth it to try to iterate on design when they can just build a completely functional prototype themselves in a few hours. We're building APIs for internal users in preference to UIs, because they can build the UIs themselves and get exactly what they need for their specific use cases and then share it with whoever wants it.
We replaced an expensive, proprietary vendor product in a couple of weeks.
I have no delusions about the scale or complexity limits of these projects. They can help with large, complex systems but mostly at the margins: help with impact analysis, production support, test cases, code review. We generate a lot of code too but we're not vibe coding a new system of record and review standards have actually increased because refactoring is so much cheaper.
The fact is that ordinary businesses have a LOT of unmet demand for low stakes custom software. The ones that lean into this will not develop superpowers but I do think they will out-compete slow adopters and those companies will be forced to catch up in the next few years.
I develop presentations now by dumping a bunch of context in a folder with a template and telling Claude Cowork what I want (it does much better than web version because of its python and shell tools and it can iterate, render, review, repeat until its excellent). The copy is quite good, I rewrite less than a third of it and the style and graphics are so much better than I could do myself in many hours.
No one likes reading a bunch of vibe coded slop and cultural norms about this are still evolving; but on balance its worth it by far.
He did a writeup: https://buduroiu.com/blog/ai-lent-end/
Don't leave the kicker out of the story
Mainn blockers are still product, legal, management ... which Claude code didn't help with.
At work, what I see happening is that tickets that would have lingered in a backlog "forever" are getting done. Ideas that would have come up in conversation but never been turned into scoped work is getting done, too. Some things are no faster at all, and some things are slower, mostly because the clankers can't be trusted and human understanding can't be sped up, or because input is needed from product team, etc. But the sorts of things that don't make it into release notes, and are never announced to customers, those are happening faster, and more of them are happening.
We review server logs, create tickets for every error message we see, and chase them down, either fixing the cause or mitigating and downgrading the error message, or however is appropriate to the issue. This was already a practice, but it used to feel like we were falling farther behind every week, as the backlog of such tickets grew longer. Most low-priority stuff, since obviously we prioritized errors based on user impact, but now remediation is so fast that we've eliminated almost the entire backlog. It's the sort of things that if we were a mobile app, would be described as "improvement and bug fixes" generically. It's a lot of quality-of-life issues for use as backend devs.
At home, I'm creating projects I don't intend for anyone outside my family to see. So far things I could theoretically have done myself, even related to things I've done myself before, but at a scale I wouldn't bother. Like a price-checker that tracks a watchlist of grocery items at nine local stores and notifies me in discord of sales on items and in categories I care about. It's a little agent posting to a discord channel that I can check before heading out for groceries.
Or several projects related to my hobbies, automating the parts I don't enjoy so much to give me more time for the parts I do. My collection of a half-dozen python scripts and three cron jobs related to those hobbies has grown to just over 20 such scripts and 14 cron jobs. Plus some that are used by an agent as part of a skill, although still scripts I can call manually, because I'll go back to cron jobs for everything if the price of tokens rises a bit more.
I was super-skeptical, and now I'm not. I think companies laying off employees are delusional or using LLMs as an excuse, but there is zero question in my mind that these things can be a huge boon to productivity for some categories of coding.
https://en.wikipedia.org/wiki/Jevons_paradox
In the end only profit matters
We are definitely reaching the point where you need an LLM to deal with the onslaught of LLM-generated content, even if the humans are being judicious about editing everything. We're all just cranking on an inhumanly massive amount of output and it's frankly scary.
I presume I'm not the only one.
Barely an hour goes by without a new 4-page document about something that that everyone is apparently ment to read, digest and respond to, despite its 'author' having done none of those steps, it's starting to feel actively adversarial.
With good management you will get great work faster.
The distinguishing feature between organisations competing in the AI era is process. AI can automate a lot of the work but the human side owns process. If it’s no good everything collapses. Functional companies become hyper functional while dysfunctional companies will collapse.
Bad ideas used to be warded off by workers who in some shape or form of malicious compliance just would slow down and redirect the work while advocating for better solutions.
That can’t happen as much anymore as your manager or CEO can vibe code stuff and throw it down the pipeline for the workers to fix.
If you have bad processes your company will die, or shrivel or stagnate at best. Companies with good process will beat you.
I just went and deleted it because it's completely broken at every edge case and half of the happy paths too.
This was possible before but someone would maybe notice the insane spaghetti. Now it's just "we'll fix it with another layer of noodles".
edit: LOL called it, a bunch of useless garbage that no one really cares about but used to justify corporate jobs programs.
Still useless in the sense that if you died tomorrow and your app was forgotten in a week the world will still carry on. As it should. Utterly useless in pushing humanity forward but completely competent at creating busy work that does not matter (much like 99% of CRUD apps and dashboards).
But sure yeah, the dashboard for your SMB is amazing.
Your rant just shows you don't understand why people pay for software.
I'd been fighting to make this for two years and kept getting told no. I got claude to make a PoC in a day, then got management support to continue for a couple weeks. It's super beneficial, and targets so many of our pain points that really bog us down.
Or, Excel > Data > Sort > by the Date column. No dashboard needed, no app needed.
In the startup world something like "every emailed spreadsheet is a business" used to be a motivating phrase, it must be more rough out there when LLMs can business-ify so many spreadsheet processes (whether it's necessary for the business yet or not). And of course with this sort of tool in particular, more eyes seeing "we're paying $x/mo for this service?" naturally leads to "can't we just use our $y/mo LLM to make our own version?". Not sure I'd want to be in small-time b2b right now.
If you are using an LLM to create an application to grab data from heterogeneous sources, combine it and present it, that is much better, but could also basically be the excel spreadsheet they are describing.
And what’s worse is that when someone does build a decent tool, you can’t help but be skeptical because of all the absolute slop that has come out. And everyone thinks their slop doesn’t stink, so you can’t take them at their word when they say it doesn’t. Even in this thread, how are you to know who is talking about building something useful vs something they think is useful?
A lot of people that have always wanted to be developers but didn’t have the skills are now empowered to go and build… things. But AI hasn’t equipped them with the skill of understanding if it actually makes sense to build a thing, or how to maintain it, or how to evolve it, or how to integrate it with other tools. And then they get upset when you tell them their tool isn’t the best thing since sliced bread. It’s exhausting, and I think we’ve yet to see the true consequences of the slop firehose.
I run a team and am spending my time/tokens on serious pain points.
This is in a real-time stateful system, not a system where I'd necessarily expect the exact same thing to happen every time. I just wanted to understand why it behaved differently because there wasn't any obvious reason, to me, why it would.
The explanation it came back with was pretty wild. It essentially boiled down to a module not being adequately initialized before it was used the first time and then it maintained its state from then on out. The narrative touched a lot of code, and the source references it provided did an excellent job of walking me through the narrative. I independently validated the explanation using some telemetry data that the LLM didn't have access to. It was correct. This would have taken me a very long time to work out by hand.
Edit: I have done this multiple times and have been blown away each time.
> The explanation it came back with was pretty wild. It essentially boiled down to a module not being adequately initialized before it was used the first time and then it maintained its state from then on out.
Even without knowing any of the variable values, that explanation doesn't sound wild at all to me. It sounds in fact entirely plausible, and very much like what I'd expect the right answer to sound like.
This the the difference between intentional and incidental friction, if your CI/CD pipeline is bad it should be improved not sidestepped. The first step in large projects is paving over the lower layer so that all that incidental friction, the kind AI can help with, is removed. If you are constantly going outside that paved area, sure AI will help, but not with the success of the project which is more contingent on the fact that you've failed to lay the groundwork correctly.
it's crazy that the experiences are still so wildly varying that we get people that use this strategy as a 'valid' gotcha.
AI works for the vast majority of nowhere-near-the-edge CS work -- you know, all the stuff the majority of people have to do every day.
I don't touch any kind of SQL manually anymore. I don't touch iptables or UFW. I don't touch polkit, dbus, or any other human-hostile IPC anymore. I don't write cron jobs, or system unit files. I query for documentation rather than slogging through a stupid web wiki or equivalent. a decent LLM model does it all with fairly easy 5-10 word prompts.
ever do real work with a mic and speech-to-text? It's 50x'd by LLM support. Gone are the days of saying "H T T P COLON FORWARD SLASH FORWARD SLASH W W W".
this isn't some untested frontier land anymore. People that embrace it find it really empowering except on the edges, and even those state-of-the-art edge people are using it to do the crap work.
This whole "Yeah, well let me see the proof!" ostrich-head-in-the-sand thing works about as long as it takes for everyone to make you eat their dust.
I'm not trying to marginalize your or anyone else's usage of AI. The reason people are saying "such as" is to gauge where the value lies. The US GDP is around 30T. Right now there's is something like ~12T reasonably involved in the current AI economy. That's massive company valuations, data center and infrastructure build out a lot of it is underpinning and heavily influencing traditional sectors of the economy that have a real risk of being going down the wrong path.
So the question isn't what can AI do, it can do a lot, even very cheap models can handle most of what you have listed. The real question is what can the cutting edge state of the art models do so much better that is productively value added to justify such a massive economic presence.
It's the same model as Uber, and I can't afford Uber most of the time anymore. It's become cost prohibitive just to take a short ride, but it used to cost like $7.
It's all fun and games until someone has to pay the bill, and these companies are losing many billions of dollars with no end in sight for the losses.
I doubt the tech and costs for the tech will improve fast enough to stop the flood of money going out, and I doubt people are going to want to pay what it really costs. That $200/month plan might not look so good when it's $2000/month, or more.
You can use "API-style" pricing on these providers which is more transparent to costs. It's very likely to end up more than 200 a month, but the question is, are you going to see more than that in value?
For me, the answer is yes.
The "costs" are subsidized, it's a loss-leader.
> This whole "Yeah, well let me see the proof!" ostrich-head-in-the-sand thing works about as long as it takes for everyone to make you eat their dust.
People will stop asking for the proof when the dust-eating commences.
I personally noticed this. The speed at which development was happening at one gig I had was impossible to keep up with without agentic development, and serious review wasn't really possibile because there wasn't really even time to learn the codebase. Had a huge stack of rules and MCPs to leverage that kinda kept things on the rails and apps were coming out but like, for why? It was like we were all just abandoning the idea of good code and caring about the user and just trying to close tickets and keep management/the client happy, I'm not sure if anyone anywhere on the line was measuring real world outcomes. Apparently the client was thrilled.
It felt like... You know that story where two economists pass each other fifty bucks back and forth and in doing so skyrocket the local GDP? Felt like that.
Claude is a tool. It can be abused, or used in a sloppy way. But it can also be used rigorously.
I've been beating my team to be more papercut-free in the tooling they develop and it's been rough mostly because of the velocity.
But overall it's a huge net positive.
well, isn't that what AI can be used effectively for - to generate [auto]response to the AI generated content.
I guess you gotta look busy. But the stick will come when the shareholders look at the income statement and ask... So I see an increase in operating expenses. Let me go calculate the ROIC. Hm its lower, what to do? Oh I know, lets fire the people who caused this (it wont be the C-Suite or management who takes the fall) lmao.
You could argue that all the spending is wasted (doubtless some is), but insisting that the decision is being made in complete ignorance of financial concerns reeks of that “everyone’s dumb but me” energy.
The real thing to look at is whether or not the future outlook for company AI spend is heading up or down?
Are they peeking over the shoulder of each team and individual? Of course not.
It can be the case that the spend is absolutely wasteful. Numbers don’t lie.
Oh, they were involved all right. They ran their analyses and realized that the increase in Acme Corp's share price from becoming "AI-enabled" will pay for the tokens several times over. For today. They plan to be retired before tomorrow.
Most firms are not a google or a Microsoft - a firms cash balance can become a strategic weapon in the right environment. So wasting money is not a great idea. Lest we forget dividends.
Moreover if you have a budget set re. Spend on tokens - you have rationing. Therefore the firm should be trying to get the most out of token spend. If you are wasting tokens on stuff that doesn’t create a benefit financially for the firm then indeed it is not inline with proper corporate financial theory.
People who work at VC-backed firms do not get to enjoy the same degree of liquidity, not even close. There can be some outliers but that is 0.1% of all.
Can't believe simple stuff like this has to be said.
Round-tripping used to be regulated. SPVs used to be regulated. If you need a loan you used to have to go to something called a bank, now it comes from ???? who knows drug cartels, child traffickers, blackstone, russians & chinese oligarchs. Even assuming it doesn't collapse tommorow why should they make double digit returns on AI datacenters built on the backs of Americans?
> “Im convinced none of these people have any training in corporate finance. For if they did they'd realise they were wasting money.”
This isn’t meaningful criticism. This is a vacuous “those guys are so dumb”.
[waits for chickens to come home to roost]
After all (Grug Chief reminds us), the only truly secure computing system is an inert rock.
"We are writing down X billions over 4 years, and have cancel several ambitious programs related to our AI experiments. We were following standard practice in the industry, so [shareholders] can't blame us for these chickens coming to roost. If everyone is guilty, is anyone really guilty?"
> Security is less or no concern, bugs are more acceptable, performance / scalability rarely a concern. Quickest way to get things done
This is literally how rest of the world works already, and always had. We'd still be living in caves otherwise. Fortunately most people (at least outside software) seem to understand that security is a trade-off against usefulness, and not an end goal in itself.
Even right now the difference with working with 'AI native' developers or with regular developers is day and night.
I certainly wouldn't want a non-clause enabled developer on my team now.
You only want to work with people who are hip with the North Pole?
I wonder what I’m doing differently.
I did spend quite a bit of time, mostly manually, improving development processes such that the agent could effectively check its work. This made a difference between the agent mostly not working and mostly working. Maybe if I had instead spent gobs of money it would have worked output tooling improvements?
Haven't found a process that beats this yet and I burn very few tokens this way.
I like writing code, I’m good at writing code. What I hate doing is dredging through logs, filtering out test scenarios and putting together disparate information from knowledge silos - so I have the AI doing that. It’s my research assistant.
Effectively I’m using it like an automated search engine that indexes anything I want and refines the results by using the statistical near neighbors of how other people explained their searches.
It's now trivial to fix these problems while still doing our day jobs -- shipping a product.
This will have previously been too ambitious to ever scope but we’ve been able to build essentially all of it in just two months. Since it sits on top of our other systems and acts as more of a window/pass through control pane, the fact that it’s vibe coded poses little risk since we still have all the existing infrastructure under it if something goes awry.
it's trivial to reimplement a better solution.
Also, I am not sure if it is trivial to implement. The code is injected into many scenarios and workflows, so replacement will be painful and risky if new solution break some edge case.
It's better than the "here's my code, it a giant pile of spaghetti but only luddites care about code quality and maintainability anyway" method, at least.
I've been using it to write tools that drastically facilitate spinning up local k8s cluster with an entire suite of development services that used to take two days to set up in Docker.
Coding velocity doesn't matter if it the net result is software that sucks massive schlong. The real world doesn't care if programmers can write code faster.
My hypothesis is that companies dont want to offer cheaper nor better services. Only want to cut costs and keep the revenue for investors.
I other news, TQQQ is pretty high!
Where I work, the power dynamics have shifted wildly. There are a number of senior engineers who refuse to touch the stuff, and as a result, they can barely keep up with their peers. Some of our juniors are now running laps around them.
When a stranger to your craft can now teach themselves what you know, how to do your job, and even how to automate your tasks in the span of the same workday as you, all while reliably being able to gauge the innacuracy of the output they're reading, how much longer do you really hold relevance?
Are the juniors increasing economic productivity or just pushing lines of code?
</retired from being measured against a random number generator>
And also because the Plan agent generates a huge plan, asks me a couple yes/no questions with an obvious answer, and then regenerates the entire plan again. Then the Build agent gets confused anyway and does something else, and I have to round-trip about 5 times with that full context each time.
But yea it's not gonna make facebook 20% better tomorrow just that you need 5 people instead of 40 to build the next facebook.
I'm at least 5x faster, if not more. With tooling I might be able to get to 10-15x.
That "more expensive" is someone's revenue. May be AI is the kind of technology that allows to make more and more revenue by making things more expensive and worse than by making them better and cheaper.
And yet.. building shit is no longer the sole domain of the software engineer.
That's the sea change.
I've literally had finance and GTM stand things up for themselves in the last few weeks. A few tweaks (obviously around security and access), and they are good to go.
They've gone from wrangling spreadsheets to smooth automated workflows that allow them to work at a higher level in a matter of months.
That's what all this AI is doing. The shit we could never get the time to get around to doing.
The only thing that matters is the impact on the financials. The shareholders (the people who employ you) dont care about any of this if it does not enhance value.
Another project I'm seeing in the same realm is taking an approved protocol and some study results and checking that the records of what was done match what they said they could do in the approved protocol. It can also make sure that surgical records have all the things they should have. This can help meet one of the requirements from the national accreditation organization to do "post approval monitoring".
Another way I've used it is to have it collate and compare a particular kind of policy across many institutions who transparently put their policies online. Seeing the commonality between the policies and where some excel helped me rewrite our policy.
This is work that just wasn't happening before or, more accurately, it was being spread over lots of people, and any improvement in efficiency or consistency is hard to measure.
Given the fact that both Altman and Amodei are pathological liars, there's absolutely no reason to believe that Anthropic has $30B ARR.
Can you explain how that’d work? What would the $30B figure be based on if they only have $100 in revenue?
(Run Rate = Revenue in Period / # of Days in Period x 365)
It's a forecast.
(That said, their numbers are much realer than that.)
That said, most people would use a monthly or quarterly period to estimate ARR. I'm not sure what Anthropic used. Probably monthly.
(I would then argue that he was re-hired specifically because others involved with OpenAI understood that it is literally his job to lie and that OpenAI would not be where it is today as a corporate behemoth rather than a research non-profit without a world-class liar marketing it, but that is merely conjecture.)