News for hackers. Stuff that matters.
Re:How the Claude Code team thinks about verification by OtherShrezzing on Tuesday November 17, 2026 @06:35AM (#50530087)
This article feels like the result of having unlimited tokens at your disposal. It’s got some very expensive practices, where near-free practices would produce the same or better results.

There’s a long section in the middle where the article describes per-change verification via skill files and “ask Claude to use the chrome-for-Claude tool” when checking if changes have introduced regressions.

But there’s already tools which do this without spending tokens tokens. Playwright will achieve equal quality but deterministic results across the same verification attempts. It is write-once-use-often, so has a marginal cost near zero.

Re:How the Claude Code team thinks about verification by cyanydeez on Tuesday November 17, 2026 @06:41AM (#50530137)
Chinese gonna beat anthropic because they gluttonous use of AI.
Re:How the Claude Code team thinks about verification by bitlad on Tuesday November 17, 2026 @06:43AM (#50530152)
They are living in the future.

I think future AI cost will be as cheap as doing an API call in classic sense.

Re:How the Claude Code team thinks about verification by ben_w on Tuesday November 17, 2026 @07:04AM (#50530306)
Quite possibly. But, a thought I had some years back:

In the future when AI is able to do everything, there will not be one single super-AI that does it all, but a whole ecosystem of different AI. This is the same as how in the biological world, while there are various degrees of generalists, there's no single cell, plant, or animal which does absolutely everything. If our body needs more blood, it makes more blood cells, it doesn't repurpose intestine or skin for the task.

Similar argument in economics, comparative advantage: an AI which can do everything is best used for the things which only it and no other AI can do, and leaving to other (smaller, cheaper) AI relatively simple tasks like face recognition.

Even where AI can do this, that doesn't mean the best solution is AI rather than a traditional API.

Re:How the Claude Code team thinks about verification by brabel on Tuesday November 17, 2026 @07:14AM (#50530379)
You are dangerously close to dismissing the Bitter Lesson.

https://en.wikipedia.org/wiki/Bitter_lesson

Re:How the Claude Code team thinks about verification by ben_w on Tuesday November 17, 2026 @07:23AM (#50530459)
Hardly.

First, last I checked AlphaZero is a much better chess player than any LLM, and by a large margin.

Second, this isn't even about finding the best model for small tasks, it's about best use of finite resources, compute or energy, which means that it still holds even if a future AI (call it Alice) did everything better than AlphaZero does chess, it would still make sense for Alice to offload chess to a merely-good model like AlphaZero so that Alice could spend more time doing the things that nothing else besides Alice could do:

https://en.wikipedia.org/wiki/Comparative_advantage

Re:How the Claude Code team thinks about verification by antonvs on Wednesday November 18, 2026 @03:21AM (#50543377)
That’s not what the Bitter Lesson addresses. It looks at where advances in AI can come from and how different approaches, scaling vs. applying human knowledge to develop fancier models, compare in advancing AI. It doesn’t deal with comparative operating costs of different sizes of model.

Small, specialized models can still benefit from the Bitter Lesson in that they can take advantage of ever-cheaper computation. They don’t contradict the lesson.

Re:How the Claude Code team thinks about verification by tpetry on Tuesday November 17, 2026 @07:06AM (#50530322)
Many API calls can be answered by something as underpowered as an Raspi. But AI needs very expensive GPUs to run.

So no way it will be as cheap. That is just wishfull thinking.

Re:How the Claude Code team thinks about verification by slowhorse on Tuesday November 17, 2026 @07:12AM (#50530360)
So the future is about replacing deterministic procedures with stochastic procedures?

On some level you are right, LLMs and the agents built with the various models and architectures will get cheaper. But there are some significant drawbacks by introducing this risk, regardless of cost. The future will need the capacity to understand this too.

Re:How the Claude Code team thinks about verification by mleo on Tuesday November 17, 2026 @09:37AM (#50531976)
Just as likely, as models and harnesses get better with memory, they will recognize the process of building and rebuilding the same tool to complete a task can be optimized by keeping the tool around and using it to deterministically complete the task instead.

This is the process of increasing efficiency today, automate manual processes to reduce human effort. Why wouldn’t models get better at this as well?

Re:How the Claude Code team thinks about verification by jimbokun on Tuesday November 17, 2026 @05:27PM (#50539159)
Just in time compiler but at the source code instead of byte code level.
Re:How the Claude Code team thinks about verification by catchnear4321 on Tuesday November 17, 2026 @06:45AM (#50530157)
It behooves the token merchant to espouse the benefits of having more tokens.

Not that it is an explicit, conscious direction, but it very much is one.

You’re not wrong, in that free flowing tokens are enough to convince such a group it’s simply good advice.

Re:How the Claude Code team thinks about verification by adamddev1 on Tuesday November 17, 2026 @06:54AM (#50530234)
The more you use AI to write your codebase, the more you need AI to manage your codebase.

The more you use product X the more product X you need to buy. Nice to be selling product X.

I'm not working for a manager, so I don't need to use it at all.

Re:How the Claude Code team thinks about verification by catchnear4321 on Tuesday November 17, 2026 @06:57AM (#50530257)
This feels a bit like the other extreme.

It can meaningfully “type” faster than you.

Take the handicap if you really wish to, I suppose.

Re:How the Claude Code team thinks about verification by adamddev1 on Tuesday November 17, 2026 @07:00AM (#50530281)
I do understand that it is extreme, but I think the principled stand makes my code and my skills better in the long run. I'm banking on building less, but building better. Otherwise where do you draw the line in losing quality and understanding? And for what I'm working on, typing is rarely the bottleneck anyways.

I still use LLMs (cautiously) for searching and troubleshooting API things.

Re:How the Claude Code team thinks about verification by catchnear4321 on Tuesday November 17, 2026 @07:44AM (#50530651)
Letting go of the typing isn’t a light decision. And it is harder to get good output when it’s also simply a lot of output.

I’m banking on that being an evolving, but necessary skillset. For a while at least. And I’m still gaining experience in the process.

Plenty of exposure, plenty of prior experience, it’s a weird mix of management-ish context handling and sme mentorship while also providing the research benefits you mentioned.

Re:How the Claude Code team thinks about verification by RetroTechie on Tuesday November 17, 2026 @07:56AM (#50530745)
> I'm banking on building less, but building better

Same here. But as opposed to the "use AI" vs. "don't use AI" crowds, I'd prefer a more pragmatic approach:

Do whatever possible to maximize the quality of the software. And with quality I mean correctness, reliability.

AI is just another tool there. Where it helps produce more correct, reliable code, consider using it. Where it doesn't, don't.

Sadly it seems many companies have a different definition of "better" (featuritis, shortest time to market, developer productivity, etc).

Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @06:46AM (#50530167)
I agree completely, there was a much better talk on this by Lauren Tan: https://www.youtube.com/watch?v=O_zcc6Xl12g

A key pillar was introducing things like static analysis to provide cheap deterministic checks as the first goal and only then moving to other forms of verification.

Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @06:53AM (#50530230)
You are not agreeing with the OP. You are just saying that there are many levels to testing and each give their own swiss cheese level of verification.

Sure you can have linting, compilation, unit tests, e2e tests. Do you really think Anthropic doesn't do this?

Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @07:06AM (#50530318)
I was agreeing that this page as presented felt like it was from the POV of people will unlimited tokens to burn.

In particular constructing ( and often deconstructing on change ) differential tests is very expensive relative to cheaper things to set up.

I felt this article's argument was poorly presented because of this seeming contradiction. The video I linked was much clearer about what works and what doesn't when scaling to inhuman levels of PR merges.

Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @07:18AM (#50530416)
I don't agree with "very expensive" claim. My friend works in a f500. They do releases like 2-3 times a week. Assuming each release uses $500 of tokens (higher estimate) just for verification, that's like $6000 a month and that's just a small percent over the tokens used for coding. What's the big deal? Their team brings in way more money than that - its a drop in the ocean.

I also know of companies that run very expensive test suites using spot instances in AWS that cost around $100 each time. No one complains about it.

I think all the posts complaining about expenses forget to compare the orders of magnitudes.

Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @07:49AM (#50530702)
It's not the per-run costs I worry about, it's the costs of churn.

As soon as Oracle Tests fail, the LLM has three options:

1. Rewrite the test - This undoes the benefit of the test in the first place 2. Implement workarounds and quirks to match the oracle, often at the expense of recreating buggy behaviour 3. Implement both corrected behaviour and quirks-mode behaviour to meet the oracle, raise further jobs to formalise the switch over and change of behaviour.

( And this "change of behaviour" might be a few pixels out of place! Or it could be a much more significant change. When trying to meet a goal, LLMs will often treat these with the same zeal. )

If you're working interactively, you can catch it and steer it to what's best for the feature. If you're fully automating development, then you're at the mercy of whatever the model does at that moment.

Oracle-based testing has its place, but I don't think it's good for large scale automated testing.

Also, you're talking about $500/release 2-3 times a week. That's fine, that's less than the cost of an employee.

Anthropic and others are talking about how to scale to merge 2000-3000 PRs per week.

At that scale, that cost per merge is $6M/month, surely a cost you'd want to minimise if you can cut them by half with better type-checking, static analysis, and other cheap deterministic tests, even if that's still a small cost of the whole project.

If nothing else, it's pure waste. Our planet is (mostly metaphorically) burning, and I believe that more efficient IT processes and development practices would go a long way to helping mitigate some of that.

Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @12:06PM (#50534515)
It’s not a good sign that you had to bring up the environment to justify this..
Re:How the Claude Code team thinks about verification by nozzlegear on Tuesday November 17, 2026 @01:33PM (#50536027)
[delayed]
Re:How the Claude Code team thinks about verification by Rapzid on Tuesday November 17, 2026 @04:10PM (#50538232)
That's a weak man fallacy. The strongest parts of the argument are costs and efficiency.
Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @06:17PM (#50539599)
But the author also did a fallacy

https://kenrahn.com/JFK/Critical_thinking/What_do_many_piece...

If one’s original argument doesn’t have enough weight, they just add other weak ones to justify it.

Anyway, responding to the strongest version of it: the economics of any new system always looks bad before it becomes popular. I heard people say the same thing about AI driven coding and now there’s absolutely no going back. If you were a software engineer living in 1990s and you heard someone say that Uber, a ride sharing company, spends 100s of millions per year on just microservices and db infra, you’d be skeptical again.

Usually the costs justify themselves because of the productivity it brings. People spend 10x on cloud for this reason.

If you think I’m wrong, let’s make a bet of sorts. I’m highly confident that AI driven testing will be another normal and extensively used tool just like how CI/CD as a concept is today

Re:How the Claude Code team thinks about verification by cicko on Tuesday November 17, 2026 @08:03AM (#50530811)
Come back to Earth, please. Not everyone wants to suddenly spend $6K more per month on deployments.
Re:How the Claude Code team thinks about verification by tayo42 on Tuesday November 17, 2026 @10:37AM (#50532834)
People really work in different worlds.

Most jobs I worked in wouldn't bother discussing 6k of expenses. The meetings and decisions making involved probably cost more

Re:How the Claude Code team thinks about verification by theshrike79 on Tuesday November 17, 2026 @04:13PM (#50538272)
[delayed]
Re:How the Claude Code team thinks about verification by sceptic123 on Tuesday November 17, 2026 @07:17AM (#50530409)
If you take the post at face value they are saying some of those tests aren't valuable
Re:How the Claude Code team thinks about verification by rolosa on Tuesday November 17, 2026 @07:44AM (#50530647)
That developer brags about slopping out 2000+ PRs a month, how is anything they have to say useful?
Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @08:12AM (#50530901)
Slopping out 2k+ PRs/mo while also claiming to reduce defect rates.

That seems like something to pay attention to, even if you prefer to treat it as a cautionary tale with your popcorn at the ready for when you expect it all comes crashing down.

Re:How the Claude Code team thinks about verification by rolosa on Tuesday November 17, 2026 @08:49AM (#50531344)
At 2000 PR/mo, assuming a "standard" 8 hour per day, 5 days a week workweek with no time factored for lunch, meetings, breaks... literally all they do for 8 hours a day, 5 days a week, is submitting PRs. That works out to roughly 5 minutes to slop, review, and document a single PR, approx 92 PRs per work day. That tells me they aren't doing shit.
Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @09:09AM (#50531596)
I suspect that an ever increasing number of tickets being raised are being raised by agents themselves, nitpicking and being unable to deal with the idea of implementations having trade-offs.

It doesn't take 5 minutes to prompt claude, "Please pick up Ticket #1234567".

With modern models, you rarely need much more than that. Perhaps checking in an hour later with a quick prompt, "Progress report?" if it's still working on it.

Claude having written the ticket in the first place, defined acceptance criteria. Perhaps a human has skimmed it for a go/no-go decision, perhaps not.

The point is that at that scale you can't humanly read the code, so you need to define verification strategies that work for the machine.

Precisely so that you're not just running on pure vibes.

Re:How the Claude Code team thinks about verification by skydhash on Tuesday November 17, 2026 @09:34AM (#50531934)
> The point is that at that scale you can't humanly read the code

Why is that scale needed in the first place?

Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @09:59AM (#50532270)
I don't think it is, but I think it's something worth paying attention to.

With my skeptic hat on, it'll be interesting to see what happens when you look at their internal speed, in PRs/month vs the outward velocity in features and improvements actually shipped.

With my booster hat, it's something to learn from so I can better scale from juggling hand-reviewed claude sessions to a future it might be demanded I join at some point.

We're in a weird space in development, where the initial mechanical work that used to take a week now takes an hour at most, but what remains is a lot of manual verification and developer-testing.

If we can learn from what these mega-swarms are doing to reduce the need for manual developer testing ( because at that speed it's clear there must be close to none ), then that is a useful lesson that might reduce our work per task too and let us do more.

And yes, some of that may be stuff we've known for years, such as functional CI/CD, better linting, better automated e2e testing with playwright, etc.

But actually convincing management to invest in that kind of thing isn't always easy, especially if it risks instability of a process that from their perspective has been working fine for decades, even if that "process" is someone logging onto a server to manually fire off a script that was last touched 15 years ago.

So having talks such as this to point to is useful. You can show management an expert on a stage, and in some companies an expert on a stage can be weirdly convincing to management in a way that their own employees sometimes are not.

Re:How the Claude Code team thinks about verification by skydhash on Tuesday November 17, 2026 @11:01AM (#50533275)
> We're in a weird space in development, where the initial mechanical work that used to take a week now takes an hour at most, but what remains is a lot of manual verification and developer-testing.

I don't think the low speed of development was an actual bottleneck. Coding in the real world is actually a mix of mechanical typing and information gathering. The typing part is not that relevant as it's actually not much time to type in 500 lines of code (if you touch type) assuming that's the median size of PR. It is the gathering of information that is IMO crucial for the health of the software.

It's like robots in a chemical lab. You can probably build some mechanical apparatus that can mix whatever components you want. But you still need a scientist to design experiments and to check the result. And while you can help speed up some of the process, what's valuable is the knowledge that the scientist extract, not the amount of experiments being done.

This scale of work is only possible because there's no information gathering going on. Instead it's just busy work with wishes and prayers hoping to get a good result.

Re:How the Claude Code team thinks about verification by fg137 on Tuesday November 17, 2026 @09:48AM (#50532133)
While I have seen numerous tickets addressed by Claude/Codex in one shot with decent solutions, about half of their first implementation is completely wrong in the direction, is way over engineered/significantly expands the scope, or leads to very confusing or even incorrect user-facing behavior.

I can't see how I would be able to keep customers happy without actually reading the code and understand what it is doing.

Re:How the Claude Code team thinks about verification by rolosa on Tuesday November 17, 2026 @12:11PM (#50534616)
The argument that engineers like the one in question are needed because of their judgment, guidance, experience, etc. But on the other hand if their job is reduced down to "@grok review and fix issue #42069", every 5 minutes of their workday, then what value do they actually add? Why not just remove them from the equation? I'm saying that a developer who brags about doing 2000+ prs/day is effectively dead weight and brings nothing to the table.
Re:How the Claude Code team thinks about verification by senordevnyc on Tuesday November 17, 2026 @08:29PM (#50540719)
Well, she’s using the new tools available to us to pioneer these kinds of scalable software factories and using that process to build a very successful product.

It sounds like you’re doing things basically the way they’ve always been done.

Re:How the Claude Code team thinks about verification by magicalist on Tuesday November 17, 2026 @10:45PM (#50541606)
> It sounds like you’re doing things...

What sounds like that exactly? I don't see anything in the GP about what and how they're "doing things" one way or the other.

They're making an argument about someone's statements. Engage with the argument or don't, but don't make it weird like this.

Re:How the Claude Code team thinks about verification by senordevnyc on Wednesday November 18, 2026 @07:45AM (#50545644)
This comment: https://news.ycombinator.com/item?id=50531344

This is so far from engaging with how Lauren actually works that’s it’s barely worth even responding to. They’re describing a standard software engineer workflow as it existed decades ago. Lauren is playing on a totally different level, more successfully I suspect (though who knows because while she’s public, this critic is anonymous).

Re:How the Claude Code team thinks about verification by losteric on Tuesday November 17, 2026 @10:57PM (#50541700)
Maybe the value is in setting up that system and the skillset to know when to stop/sidestep/backtrack that mechanism?

Basically management I guess

Re:How the Claude Code team thinks about verification by Alacart on Tuesday November 17, 2026 @10:13AM (#50532477)
While I suspect your point about the value/quality of the work at 2000 PRs/month is directionally correct, isn’t their entire take that they’re no longer micromanaging the PRs at all? So talking about how much time they commit to each one doesn’t make much sense. It’s a bit like looking at a modern car and saying “haha how will they even get the horses to pull this?” or maybe “how do these developers know the compiler isn’t using GOTO statements underneath?”.

Now, are those PRs doing useless work? Are they introducing unnecessary risks or trade offs? Are they building towards something impossible to understand or that will fall apart under its own sloppy weight?

Those seem more like the right questions to me, in the long run, and I suspect that’s what you really have in mind. Because if the answers are no, then why wouldn’t you be doing as many PRs as possible?

Re:How the Claude Code team thinks about verification by JambalayaJimbo on Tuesday November 17, 2026 @11:06AM (#50533380)
A PR to me represents a specific boundary of human ... review. It's a unit of work that is typically reviewed by at least 2 team members. If I'm working solo on some project, I don't use PRs.

So to me "not micromanaging PRs" doesn't make any sense at all. Bragging about how many PRs you can get merged is like bragging about how many WPM you can get when prompting an AI instead of manually typing.

Re:How the Claude Code team thinks about verification by ffsm8 on Tuesday November 17, 2026 @02:15PM (#50536676)
That's been the way I've worked in previous jobs too. Nonetheless some PRs need less review, others more. It's a gradient, and they're merely moving the range further towards "doesn't need in-depth" by introducing other quality gates.

That comes with trade-offs, but some may be willing to accept those. It's ultimately the decision of the developers and enterprise how they want to build their product, can't really say anything else about this

Re:How the Claude Code team thinks about verification by jwilliams on Tuesday November 17, 2026 @04:27PM (#50538438)
Are you arguing that the PR is unnecessary for AI/agents? Or that human review is mandatory?

All productivity metrics have flaws (usually deep ones). Doesn’t mean they’re not instructional though.

Re:How the Claude Code team thinks about verification by jimbokun on Tuesday November 17, 2026 @05:15PM (#50539024)
PRs are part of a software development process meant for collaborating humans. If no humans are reviewing anything, having PRs as part of the process is pretty pointless.
Re:How the Claude Code team thinks about verification by jwilliams on Tuesday November 17, 2026 @07:03PM (#50540051)
I mean, I agree conceptually. But the PR is just a useful unit of work.

My PRs get assessed for if a human is needed. I could probably do that at the branch level, but PRs just has all the metadata and CI/CD tooling etc etc.

Re:How the Claude Code team thinks about verification by senordevnyc on Tuesday November 17, 2026 @08:25PM (#50540693)
Exactly this. I have repos where agents can merge their own PRs, I literally never look at them. It’s still a useful primitive.
Re:How the Claude Code team thinks about verification by dmoy on Tuesday November 17, 2026 @05:35PM (#50539258)
> Or that human review is mandatory?

Depending on the domain, human review can be mandatory. (i.e. if you're a public company and the thing you're touching can materially affect the company finances)

Re:How the Claude Code team thinks about verification by wanderlust123 on Tuesday November 17, 2026 @08:49PM (#50540869)
Frontier LLMs can do reviews equal to if not better than the average dev once setup properly with guidance.
Re:How the Claude Code team thinks about verification by WeakSpeller on Tuesday November 17, 2026 @10:09PM (#50541377)
You need to work with better average devs then.
Re:How the Claude Code team thinks about verification by fg137 on Tuesday November 17, 2026 @08:03AM (#50530817)
The talk makes that sound something new, but every team I know has already adopted that before AI. I really don't get what value that talk provides -- I learned nothing after watching it.
Re:How the Claude Code team thinks about verification by xnorswap on Tuesday November 17, 2026 @08:09AM (#50530868)
You are fortunate to have always worked in highly effective teams and organisations.

This reminds me of when "The Joel Test" was first published and it was met by people unbelieving there was a world out there who weren't even using source control at the time, let alone a bug-tracker.

Re:How the Claude Code team thinks about verification by JambalayaJimbo on Tuesday November 17, 2026 @11:00AM (#50533264)
I have a friend who worked for a billion dollar company (a household name here in Canada) who conferred that they weren't using source control. As late as 2019.
Re:How the Claude Code team thinks about verification by vinnymac on Tuesday November 17, 2026 @02:03PM (#50536492)
Reminds me of a DevOps conference I attended in 2019 where everyone in the room hadn't heard of a feature flag before, and wanted me to share my experiences with them so they could better understand how to use them in their enterprise software. Meanwhile I had been using them since I was an intern, and was confused why all these graybeards were listening to me talk about anything from a db table to a full fledged company like LaunchDarkly, Optimizely, etc.
Re:How the Claude Code team thinks about verification by jimbokun on Tuesday November 17, 2026 @05:10PM (#50538965)
Good on those graybeards for wanting to learn new things.
Re:How the Claude Code team thinks about verification by solyin on Tuesday November 17, 2026 @06:38PM (#50539804)
We are our own bubbles of competence.
Re:How the Claude Code team thinks about verification by jamesfinlayson on Tuesday November 17, 2026 @11:28PM (#50541883)
I hope their beards really were grey - I work with a guy who worked somewhere in the mid 1990s and there were two servers (mainframes) - a development server and the production server. I could believe feature flags weren't a thing at that company, but we've also come a long way.
Re:How the Claude Code team thinks about verification by SatvikBeri on Tuesday November 17, 2026 @03:40PM (#50537860)
For some contrast, my first job where the team used source control was in 2015 – 15 years after the Joel test came out. That was after switching jobs in manufacturing or finance to tech.
Re:How the Claude Code team thinks about verification by stingraycharles on Tuesday November 17, 2026 @06:40PM (#50539829)
Believe it or not, there still are so many places that I’m working with that have hundreds of thousands of lines of code and no source control. They just pass zip files around and whatnot. There’s a whole parallel universe invisible to the average HN reader.
Re:How the Claude Code team thinks about verification by dominotw on Tuesday November 17, 2026 @09:34AM (#50531928)
ai engineer channel was good for a brief moment. Now its just slop.

Love to find some high quality youtubes to listen too to reduce my screen time.

Re:How the Claude Code team thinks about verification by ntcho on Tuesday November 17, 2026 @12:31PM (#50534982)
[delayed]
Re:How the Claude Code team thinks about verification by khasan222 on Tuesday November 17, 2026 @01:39PM (#50536117)
Absolutely teams have. But for some reason we expect the ai to write code better than us, so we don’t hold it to the same standards all the time.
Re:How the Claude Code team thinks about verification by verelo on Tuesday November 17, 2026 @10:19AM (#50532564)
I wish my reaction to seeing this post was "oh wow Anthropic never write stuff like this, I should make time to read it". My gut instinct was to say "this is going to be long, over complicated and not worth my time". I really think this is the influence of the 90's - 00's SEO. Write lots of stuff, hope Google cares...get more business.
Re:How the Claude Code team thinks about verification by Creamsicle47 on Tuesday November 17, 2026 @01:26PM (#50535911)
Lauren Tan is a professional larper
Re:How the Claude Code team thinks about verification by arbll on Tuesday November 17, 2026 @06:50AM (#50530199)
What makes you think they are not using Playwright ?

For me this article is more aimed at devs that are still stuck treating code generation and verification as two separate tasks. Giving agents the right tools (including playwright) to do both at once makes things more token efficient, not less.

Re:How the Claude Code team thinks about verification by pluc on Tuesday November 17, 2026 @06:50AM (#50530204)
Here's how to properly use XYZ, yes it'll cose you more but we manufacture it so to us it's no issue.

Do you not see the propaganda through the pseudo helpfulness?

Re:How the Claude Code team thinks about verification by hackyhacky on Tuesday November 17, 2026 @06:53AM (#50530220)
It's the same reason they give you a recipe for Chex Mix on the back of every box of Chex.
Re:How the Claude Code team thinks about verification by ben_w on Tuesday November 17, 2026 @07:14AM (#50530380)
> Do you not see the propaganda through the pseudo helpfulness?

I find it amusing that people phrase bog-standard "what can our product do" documents as "propaganda", even when it's a dogfooding exercise.

Like, I guess? In the sense that all marketing can be seen as capitalist propaganda?

Re:How the Claude Code team thinks about verification by pluc on Tuesday November 17, 2026 @07:21AM (#50530446)
Marketing is for products, propaganda is for agendas.
Re:How the Claude Code team thinks about verification by mhluongo on Tuesday November 17, 2026 @07:30AM (#50530511)
Tokens are their product.
Re:How the Claude Code team thinks about verification by ben_w on Tuesday November 17, 2026 @07:37AM (#50530581)
Most businesses have an agenda of "sell more product".
Re:How the Claude Code team thinks about verification by SecretDreams on Tuesday November 17, 2026 @06:51AM (#50530213)
You don't get a nice juicy IPO valuation by trying to save money or be efficient with your token usage!
Re:How the Claude Code team thinks about verification by baxtr on Tuesday November 17, 2026 @06:52AM (#50530214)
Wait, are you saying a company that makes money from selling tokens generated a content piece recommending people solve their problems by using a lot of tokens?

Scandalous.

Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @06:52AM (#50530219)
Its almost become a meme at this point: well actually you can achieve the same using X so why west tokens?!

The reason you should use Claude and not playwright always is because the the invariant you wanna capture is at the natural language level.

Like imagine I wanted to verify that Claude Code works and doesn't break. I can describe the invariant I want in natural language. Like: make sure that subagents don't block the original thread.

Doing this in playwright means that you have already bought into some previously existing architecture which means your tests are now correlated to your code. Having the invariants written in natural language allows you to decouple code vs the invariant.

I expect a lot of greybeard conservative programmers to keep opposing using AI here until eventually, not using AI will be seen as more risky.

Re:How the Claude Code team thinks about verification by LelouBil on Tuesday November 17, 2026 @07:02AM (#50530296)
[delayed]
Re:How the Claude Code team thinks about verification by fingerlocks on Tuesday November 17, 2026 @07:47AM (#50530684)
[delayed]
Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @12:05PM (#50534507)
If you re architecture threads to processes, you have to rewrite your tests.

That was kind of my point. Natural language almost always survives rearchitectures because your higher level interface to your end customer rarely has churn.

In a few years, agent driven testing will become normal and almost expected in some enterprise companies. Obviously we will still have people like you opposing it on grounds of “but just write tests in playwright”.

I also find the greybeard archetype fascinating. They’ve opposed almost everything new including Cloud, microservices, kubernetes. Now with AI. And here in this thread with AI driven tests. They seem to think programming should be done exactly how it was done when they were in the prime of their career. Anything before that is archaic and anything after is hype.

Re:How the Claude Code team thinks about verification by jimbokun on Tuesday November 17, 2026 @05:35PM (#50539248)
There were also a lot of massive cost overruns and technical debt accrued by people applying those things where they weren’t needed.
Re:How the Claude Code team thinks about verification by simianwords on Tuesday November 17, 2026 @06:06PM (#50539519)
Maybe. But these new concepts won and they weren’t just hype. They have their place. I’m obviously not saying AI testing has no wrong way of being used.

I’m saying that a default response to challenge AI testing by saying “just write normal tests” is the typical greybeard archetypical response. And they don’t really have a good track record considering they’d by default challenge cloud, kubernetes etc

Re:How the Claude Code team thinks about verification by jaapz on Tuesday November 17, 2026 @05:11PM (#50538971)
Don't know about you but my Claude uses playwright to verify everything works, playwright MCP is great!
Re:How the Claude Code team thinks about verification by simianwords on Wednesday November 18, 2026 @12:15AM (#50542143)
Does Claude create a new playwright each time? If not, how does it know when to change it when the underlying architecture changes?

If yes, that’s what I recommended.

Re:How the Claude Code team thinks about verification by LelouBil on Tuesday November 17, 2026 @07:01AM (#50530284)
[delayed]
Re:How the Claude Code team thinks about verification by abirch on Tuesday November 17, 2026 @07:15AM (#50530383)
It seems like a new stack where you try to push things down as low as possible for efficiency.

    [    AI    ]
    [ Software ]
    [ Kernel/OS]
    [ Hardware ]
Re:How the Claude Code team thinks about verification by saagarjha on Tuesday November 17, 2026 @08:22AM (#50531012)
Fun fact: Codex computer use is trained to not enter passwords. When it was being designed this was considered to be a very cheap way to ensure the model would not go poking around places it shouldn’t be, instead of constantly trying to check whether the action being performed was authorized or not. People have spent a lot of time marking out their UIs for sensitive actions and it’s worth using it :)
Re:How the Claude Code team thinks about verification by pu_pe on Tuesday November 17, 2026 @07:12AM (#50530359)
> It’s got some very expensive practices, where near-free practices would produce the same or better results.

I think tokens are near-free already, even for those of us outside Anthropic.

More importantly, your philosophy comes from a time where thinking had to be front-loaded: you try to predict everything that can go wrong, then design deterministic tests that can catch those things and a decision tree to figure out what to do in each case. This approach was the only viable way until now, but maybe that will change. Having the equivalent of a senior engineer thoroughly review each error with full context of what the application and feature is trying to achieve could perhaps uncover different insights, or cover edge cases more smoothly.

Re:How the Claude Code team thinks about verification by skydhash on Tuesday November 17, 2026 @07:27AM (#50530492)
> you try to predict everything that can go wrong, then design deterministic tests that can catch those things and a decision tree to figure out what to do in each case

No real software engineer think that way, as things that can go wrong are near infinite. Instead you design the right thing and write tests that verify that it’s indeed the right thing and add some more that are the wrong thing, especially at the edge of definition. Then you rely on the dev to not make very obvious errors like coding something that is not in the requirements or is completely opposite.

Re:How the Claude Code team thinks about verification by jgalt212 on Tuesday November 17, 2026 @07:46AM (#50530672)
As far as timelines go both property based tests and fuzzers predate this token maxing strategy.
Re:How the Claude Code team thinks about verification by bayindirh on Tuesday November 17, 2026 @07:51AM (#50530715)
But they are old, well tested, and not AI based, hence bland.

In this age of AI, anything well understood and effective while being efficient is out. These tokens must be used, atmosphere must be heated.

It's November. Winter is coming.

Re:How the Claude Code team thinks about verification by NewJazz on Tuesday November 17, 2026 @11:05AM (#50533358)
[delayed]
Re:How the Claude Code team thinks about verification by bayindirh on Tuesday November 17, 2026 @11:27AM (#50533807)
Normally it should snow here. I walk around with a t-shirt.

Climate change and global warming is really a scam, they have made it snow with chemtrails in the 80s, it seems.

Though I miss snow and cold. They are not making seasons as they used to anymore. They shouldn't have privatized weather.

(This comment is written by a human, and contains an implicit /s)

Re:How the Claude Code team thinks about verification by skohan on Tuesday November 17, 2026 @08:03AM (#50530816)
It's great to have that option available, but it's incredibly inefficient to just throw GPU hours at a problem that can be solved deterministically.
Re:How the Claude Code team thinks about verification by pydry on Tuesday November 17, 2026 @08:23AM (#50531017)
Good for anthropic margins though.
Re:How the Claude Code team thinks about verification by duskdozer on Tuesday November 17, 2026 @10:21AM (#50532594)
It really seems too lucrative for them to weight the models toward higher token usage and less-deterministic solutions for them not to do it.
Re:How the Claude Code team thinks about verification by castis on Tuesday November 17, 2026 @08:09AM (#50530871)
[delayed]
Re:How the Claude Code team thinks about verification by chrisweekly on Tuesday November 17, 2026 @08:18AM (#50530967)
"non-deterministic" -> "deterministic"

right?

Re:How the Claude Code team thinks about verification by castis on Tuesday November 17, 2026 @08:21AM (#50530999)
yuppp, thank you :D
Re:How the Claude Code team thinks about verification by pu_pe on Tuesday November 17, 2026 @08:35AM (#50531146)
It isn't the same work, though, so the question is whether the added cost is worth it. Yes it costs more, but the price of tokens has fallen 100x in the past two years and it will continue to drop.
Re:How the Claude Code team thinks about verification by Grimburger on Tuesday November 17, 2026 @08:28AM (#50531074)
> tokens are near-free already

How long do you think that will last?

Am over 11 billion tokens this month on a subscription. In API usage its close to $5000 and there's still 13 days left in the month.

Eventually the music in musical chairs stops. Already many anecdotal reports of the always opaque limits being lowered across the board. On a gemini pro account for personal use I can use a weeks worth of Argon in about 20 minutes on agy.

Re:How the Claude Code team thinks about verification by olkre on Tuesday November 17, 2026 @09:06AM (#50531572)
The cost halves every year so far. Current level of llm will get cheaper
Re:How the Claude Code team thinks about verification by curt15 on Tuesday November 17, 2026 @09:15AM (#50531686)
Cost or price? ANT's shareholders and creditors aren't going to wait forever for ROI.
Re:How the Claude Code team thinks about verification by rcxdude on Tuesday November 17, 2026 @09:53AM (#50532190)
Cost. Their margins are still pretty healthy from what they've disclosed.
Re:How the Claude Code team thinks about verification by AYoung010 on Tuesday November 17, 2026 @05:14PM (#50539003)
If you disregard the inflexible and increasing cost of training lol
Re:How the Claude Code team thinks about verification by Grimburger on Tuesday November 17, 2026 @10:02AM (#50532308)
Undoubtedly you are right, except I see it the complete opposite way, companies providing the best models will probably need even more from those willing to pay in the near future, and likely get it from those wanting to ride ahead of the wave.
Re:How the Claude Code team thinks about verification by svnt on Tuesday November 17, 2026 @10:48AM (#50533031)
For the majority of tasks, the best models are no longer required.

Many benchmarks are already saturated today on many tedious tasks still being done by humans across the world. Frontier labs are introducing their own budget options prior to IPO, undercutting their revenue, because otherwise they risk losing seats.

Choosing what to do and how to do it is the actual frontier now and is still not apparently well supported in any model. Current so-called decision models are not decision models, they are available-fact-evaluating models.

Re:How the Claude Code team thinks about verification by Grimburger on Tuesday November 17, 2026 @04:35PM (#50538521)
> Many benchmarks are already saturated today on many tedious tasks

Exactly my point, think the easy things won't even need 3rd party apis in the near future.

But frontier models only seem to need more and more VRAM to keep winning the race. Feel free to look up the cost of running Kimi K3 open weights today, probably the best estimate known publicly, it's hardly a casual affair.

Re:How the Claude Code team thinks about verification by mikeocool on Tuesday November 17, 2026 @04:01PM (#50538123)
Ubers got cheaper and cheaper for a while, until they didn’t.
Re:How the Claude Code team thinks about verification by solatic on Wednesday November 18, 2026 @04:41AM (#50544095)
> How long do you think that will last?

Don't forget that it's a hypothesis, not a foregone conclusion, that this will stop. Part of the question is whether demand will continue to rise. If volume increases sufficiently, you can cut prices and improve profits at the same time, as long as you continue to make a profit at the margin.

Moore's Law kept Intel and competitors pushing R&D budget that pushed perf/watt year after year because the overall computing market just kept growing and growing. The growing demand soaked up everything Intel and co. pushed into the market; the R&D spend helped players retain market share in the face of competition. Why do you think AI is different? R&D into better, faster, cheaper chips brings down the cost of both training and inference over time, and cheaper token cost unlocks additional demand, which justifies more R&D budget, in a virtuous cycle. It continues until there's really no more demand to be unlocked, no more improvements can be made to the chips (i.e. you hit the laws of physics), or there's no more money to be invested (i.e. even capital markets have limits, with only so much cash in circulation, or macro interference like war).

Do we really have any reason to believe that demand will taper off?

Re:How the Claude Code team thinks about verification by NewJazz on Tuesday November 17, 2026 @11:04AM (#50533331)
[delayed]
Re:How the Claude Code team thinks about verification by kuschku on Tuesday November 17, 2026 @02:09PM (#50536587)
Near-free? Even using Claude just for /code-review constantly leads to

> ● Agent "[REDACTED] audit" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "Verify [REDACTED] bug" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "Verify [REDACTED] bug" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "[REDACTED]" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "Verify [REDACTED] injection" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "Verify [REDACTED] regression" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "Verify [REDACTED] race" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "[REDACTED] root-cause" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5) > ● Agent "[REDACTED]" failed: Agent terminated early due to an API error: You've hit your 5-hour limit · resets 10:50pm (Europe/Berlin) (error type rate_limit, HTTP 429, request id [REDACTED], model sent to the API: claude-opus-5-5)

Re:How the Claude Code team thinks about verification by saghm on Tuesday November 17, 2026 @05:44PM (#50539334)
Even if you aren't paying currency, there's still resource consumption in terms of compute, which in turn requires energy, which is priced in data centers based on usage. You're almost literally taxing people in communities near data centers, I guess because it's cooler to do it this way?
Re:How the Claude Code team thinks about verification by hypendev on Tuesday November 17, 2026 @07:47AM (#50530677)
Exactly.

If you're doing this, the agents should create deterministic tests from its process rather than Agent tests or just writes deterministic tests for start. This is a bit of an absurd approach.

The issue I find with these articles is that they are basically the same boat as "Facebook Engineering" articles were 20 years ago.

People will be "oh because Anthropic does it it must be the way", then a bunch of influencers will retweet it as "anthropic kills writing tests", which some CEO's and team leads will parrot and get their team to absorb these practices even tho they are not the best practices.

Some of these things are just reinventing the wheel, like this is a basic script called "run_simulator.sh" and a pre-commit test hook, not a skill and "commit gate" that needs to check session transcript:

>To codify how to start the simulator or run your app, create a skill called verify. Put it in a .claude/skills/ folder in your repo, as close to the relevant code as you can, and add as many as you need. Claude knows to use the verify skill before committing changes.

>We've also created a commit gate that can block a commit if the session did not contain /code-review or /verify. We've set it up with off, observe, and enforce modes.

Re:How the Claude Code team thinks about verification by urbsgpw on Tuesday November 17, 2026 @08:40AM (#50531204)
I got this feeling a couple of times this year with their sharing of best practices. They seemed theoretically interesting, but not if you're cost-conscious. It seemed like thinking with an unimited budget. And dont get me wrong, maybe that's the way to go since you assume costs of compute will fall down eventually (and these systems will become more optimized), but it would be completely untenable to run if you don't have a similar budget constraint (or lack thereof).
Re:How the Claude Code team thinks about verification by dcss_gardener on Tuesday November 17, 2026 @10:05AM (#50532354)
It's really interesting to me to contrast the code and publications of deepseek and anthropic over the last year. Anthropic's stuff has some ideas that can be extended or applied but everything they do assumes basically unlimited frontier model inference and so is nearly unusable to me in the form they publish it.

Where deepseek prices inference the way systems programmers price latency. As a resource to be spent carefully towards specific goals, with systems in place to ensure you're within budget and getting the value you expect from it.

And then this also has second order effects on how each company treats human attention. Having fable xhigh do 3-4 passes over a change can legitimately reduce the need for human attention on it in a way that deepseek flash just can't really. So then they price operator attention in a similar way. With ant nearly considering human intervention a failure mode, while ds again prices it as a budget to be designed within.

And neither one is "right" speaking generally. I have my aesthetic preferences and my own constraints that make one a more appealing approach personally. But both are building towards the goals they have with consideration of the resources at their disposal, it's good engineering either way.

Re:How the Claude Code team thinks about verification by svnt on Tuesday November 17, 2026 @10:49AM (#50533069)
It feels like that because it is marketing to improve distant-customer results while also increasing their token use to grow revenue.
Re:How the Claude Code team thinks about verification by radlad on Tuesday November 17, 2026 @10:52AM (#50533129)
Yes. Anecdotally, I saved a ton of tokens (and bigger: time!) switching from having Claude drive Android emulator VMs (even with detailed skills and helpers) to using Roboelectric tests.
Re:How the Claude Code team thinks about verification by seiru on Tuesday November 17, 2026 @10:57AM (#50533213)
e2e tests (which you describe at the end of your post) are literally the first thing the article talks about though.
Re:How the Claude Code team thinks about verification by empath75 on Tuesday November 17, 2026 @11:24AM (#50533739)
> This article feels like the result of having unlimited tokens at your disposal

I think it's worth designing around that because tokens will only get cheaper over time.

Re:How the Claude Code team thinks about verification by bismanawaz on Tuesday November 17, 2026 @12:01PM (#50534447)
Full app testing by agents would mean that each agent needs its own running app and browser use. So machines would cost too, not just the tokens in my opinion.
Re:How the Claude Code team thinks about verification by oceanplexian on Tuesday November 17, 2026 @12:08PM (#50534565)
> This article feels like the result of having unlimited tokens at your disposal.

Gong to be honest where do you think this is going?

Do you think tokens are going to magically become more expensive or models are going to become less efficient?

Of course you should be tokenmaxxing, you can already tokenmax with a model that competes with the SOTA 1 year ago on a moderately powerful laptop you can walk out of an Apple store with.

Re:How the Claude Code team thinks about verification by throw10920 on Tuesday November 17, 2026 @12:21PM (#50534804)
To be a bit more grounded and expand on your point:

I, personally, have already experienced situations where I was employing cost-optimization optimization strategies for models 8 months ago, that are borderline-useless now because the Pareto frontier has advanced both more than I thought and also in different ways than I thought (cheap models got smarter, better at following instructions, and more clever about tool calls).

That is, the strategies decreased in effectiveness disproportionately to how much the model improved.

It's worth considering that strategies that seem insanely expensive now may be economically viable in a year or two. Or possibly significantly sooner than that.

It's also possible that, like how jet engines get more efficient as they get larger (leading to the 737 MAX incidents, which seems spiritually significant) corporations with budgets orders of magnitude greater than ours benefit from very different strategies than we do.

Re:How the Claude Code team thinks about verification by ar_lan on Tuesday November 17, 2026 @12:30PM (#50534975)
This seems like a very strange comment to come from someone I am assuming is a SWE. When have we, ever, been happy with non-deterministic, objectively more expensive, solutions?

You're assuming a lot of costs that should just be considered ok in your stance:

* Why pay anything multiple times when you can replicate deterministically for free? * Why require always-online models when you can run computation locally? * Inversely, why use power-hungry/PU-hungry models locally when you can have a deterministic software that can be optimized to not* be memory/PU/power hungry to do validation?

I think tokenmaxxing can be a great start* to get things going but when it comes to productizing I want determinism and cheap optimizations, and using frontier models to build these cheap optimizations is what we've been doing at our company over this past year to great success.

Re:How the Claude Code team thinks about verification by giancarlostoro on Tuesday November 17, 2026 @03:41PM (#50537884)
Yeah, people don't realize what terrifies me most about stuff like Mythos, the employees can throw "unlimited" compute at it, to the point where it could go off and do something off-rails that they cannot react quickly enough to prevent. At what point is any of this okay? Honestly, I hope I'm wrong and it never happens.
Re:How the Claude Code team thinks about verification by jonstewart on Tuesday November 17, 2026 @07:43PM (#50540368)
Unlimited tokens that can be used for generating turgid blog posts…
Re:How the Claude Code team thinks about verification by thefunnyman on Tuesday November 17, 2026 @08:44PM (#50540834)
Yeah I work at a place that has high token budget and I see this nonsense all the time. Even putting cost aside it's crazy to me to introduce nondeterministic checks into your process. Surely we've all dealt with the pain of flaky tests before.

AI is a huge force multiplier and makes it almost trivial to create things like lint rules you'd otherwise probably not bother to set up. Use it for that instead of trying to get it to be a shitty linter itself.

Re:How the Claude Code team thinks about verification by pixelready on Tuesday November 17, 2026 @09:35PM (#50541145)
Not only is their token budget effectively unlimited, but they have a strong incentive to both discover more uses for their models when dogfooding and to encourage readers to consume more tokens. If all you’ve got is a hammer…