Rendered at 20:11:43 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
superasn 23 hours ago [-]
Recently I came across the /handoff skill, which I've been using a lot. I find it much better than /compact.
Basically:
- /handoff file creates a short document with the important context from your current session and maybe next steps as checklist.
- You can then start a fresh session with /continue file
- You can also hand the work from Claude to ChatGPT, or the other way around. Very useful at time of session limits.
- Plus your handoff files becomes a useful piece of project memory that you can reference later.
I find this much more useful than /compact or /clear because the context is saved in something portable instead of being tied to one session and i've seen better results doing this every 20 messages or so than running long sessions.
xpct 8 hours ago [-]
Ah I've been doing handoff manually by asking the model to do it. If Claude's handoff includes better context than its summaries, it's probably worth checking out.
dboreham 7 hours ago [-]
I prefer to avoid its internal magic mechanisms. The LLM can read documents like a human (actually better since...no ADD) so I ask it to create documents. This also has the benefit that you can change to a different AI tool and the information is still usable.
3371 16 hours ago [-]
> Plus your handoff files becomes a useful piece of project memory that you can reference later.
I've been doing this since I started agentic development, and have a whole framework based on this; Simply put I define workflow s that output templated files for everytype of tasks that happens in development.
It's a powerful pattern I'd recommend everyone.
subscribed 3 hours ago [-]
Oh, so there's whole _skill_ for that!
I now feel a bit silly but I reinvented the wheel during my last lil project and indeed found it very powerful. A variant of it is the "implementation (handover) prompt" when I conclude the planning session with plans and design documents, updated handoff, clean tree and a file for a new Opus implementing orchestrator (unusually do a single, highly specific implementer and a single, highly specific tester).
After implementation and task-level tests I end up with a long and very detailed implementation progress ledger and a summary findings from the orchestrator.
Then in the new session I do the whole branch tests.
Works really well, uses much less tokens than any other approach with more of Opus and very little repetitions/corrections.
cityofdelusion 3 hours ago [-]
This is where I landed too. Basic skill for a Jira-like task board and another for keeping an orchestration ledger. Let’s me arbitrarily clear context without losing big picture knowledge.
ulimn 13 hours ago [-]
I started doing this and then a colleague of mine recommended me OpenSpec. It might be overkill for small projects but it’s pretty clean and if you have goldfish memory like me, it helps to track changes.
3371 4 hours ago [-]
Yes it's just another (but my own) impl of SDD
poseva 11 hours ago [-]
Can you please add more details? Thank you in advance!
pixelsort 8 hours ago [-]
I've been using my own variant docset driven development since late last year. It emerged organically for me by adapting to the limits of the agents.
- Prefix-numbered sets of documents, keeping them clustered visually and easily referenced by humans and agents. (124.5 = docset 125, phase 5)
- Each doc gets a suffix. Most of the time STRATEGY is the first doc. Sometimes NOTES, DISCUSSION, or HANDOFF though.
- Once the strategy doc is comprehensive, the multi-phase checklist PLAN doc can emerge.
- During execution, the agent drives the PLAN while purposefully expanding lightly scaffolded phases before entry.
- As the user, I drive a PROGRESS doc during the PLAN execution. This tracks the progression of my own questions and the important work summaries that I need for tracking the current work trajectory and for historic analysis over past decision-making.
- Any concern that exceed the scope of the active docset numbers calls for a new one.
- A "docs/archive" folder is kept where I sweep docsets into numbered eras.
I have aliases and templates too, but DDD is so simple that it's overkill.
Any modern reasoning agent can take the plain explanation above and understand you effortlessly when you say "open a new docset strategy" or "read docset 125 and proceed".
3371 4 hours ago [-]
It's just good old prompting in skill format. You can check out the skill repo: no3371/projex
skybrian 7 hours ago [-]
I do something similar but without using a skill.
It might be even better if the harness were to automatically write a handoff note under some circumstances? If the user is away and the cache is going to expire, that would be a good time to do it.
bombcar 8 hours ago [-]
I've been doing handoffs manually, but maybe the command is a better way. It's a great way to keep track of work, remove useless context, and more importantly, it's an enforced "what is the agent doing" line - you immediately notice if the handoff writing is missing something or confused. Commit each and every one (git is the unsung hero of AI).
devoidfury 14 hours ago [-]
I like the handoff flow so much I built it into my agent as a native tool; and going to add it as a auto compaction strategy as well.
kxrm 8 hours ago [-]
You can also do this with memories.
I've been asking Claude to remember important points from our session, or future tasks. Then /clear and continue.
sjanes 21 hours ago [-]
/handoff is the skill I use when I plainly are about to run out of token budget to give to a different harness system. Using this to avoid long sessions is something I will need to try, plus maybe "caveman" mode to cull the verbosity.
ls612 22 hours ago [-]
I have been doing this a lot even without a skill, having Fable write a planning document, then spawning an Opus subagent with instructions to strictly follow the plan and report any deviance at the end. It also helps that then the plan is always saved in an md file so any future agent can look at it and see what happened.
CBLT 20 hours ago [-]
Interestingly, this was tackled in this blog post[0] a month ago. They claim that plan files aren't token-efficient, because after reading the plan the workhorse model then reads all the relevant files anyways.
My plan and implementation files are task specific (so specific workhorse reads only its own slice), and the workhorse itself is spawned from an orchestrator with a very specific small prompt.
I managed even the orchestrator to NOT read the plan whole, at once, but in sections.
The most useful thing is the task ledger the task agent leaves behind, which alongside its structured status message makes a very resilient handoff between all stages.
hombre_fatal 19 hours ago [-]
That link just says the planning stage should vet the idea concretely so that the plan focuses on a solution that won’t immediately have to pivot.
And I think plan files should focus on general ideas and invariants, not do “implementation as prose”. That way they perform as mini-ADRs that are useful historically, especially to mine why the system is the way it is.
inopinatus 19 hours ago [-]
The trick is to delete all the relevant files after the plan is written.
tarekabi 1 hours ago [-]
[flagged]
alfonsobuilds 4 hours ago [-]
[flagged]
rhaksw 1 days ago [-]
> @-mention files instead of naming them
Love Claude, but the @ mention is broken in the desktop app. For the same project if I type the same query "@ephem" I get:
I could have written the issue better (using CLI as comparison instead of VS Code). But, no doubt in my mind Claude could fix this itself in a minute.
guessmyname 1 days ago [-]
> Opened issue for this and it was automatically closed: […]
Clarification: It wasn’t closed on submission though. It sat open ~17 days, a bot marked it stale, and it closed when nobody responded to the stale label.
The two-phase thing is the part I didn’t know until recently: the stale label is basically asking “is this still relevant?”, and answering it makes the bot back off next time around. nixpkgs does the same. Bumping feels wrong on most trackers, agreed, but at this issue volume I don’t know what else works.
Anyway a comment should reopen it. Your CLI vs desktop screenshots are a better repro than most things in that tracker.
DangitBobby 1 days ago [-]
By what definition of automatically does a bot closing an issue with low activity not count as "automatically closed"?
Banditoz 1 days ago [-]
Just because a Github issue doesn't have activity doesn't mean it's not an issue anymore.
1 days ago [-]
kristjansson 1 days ago [-]
the norms of issue trackers are strongly opposed to “bump”. these autoclose bots may yet change that
Doohickey-d 23 hours ago [-]
Yes, now the poster of the issue just needs to (tell claude to) write a bot which posts a "bump" comment every 17 days, and the bots will cancel each other out.
hackernudes 1 days ago [-]
In this case there is a two phase staleness check. The bot marks it as stale and it is correct to "bump" it if you think it is still important. The nixpkgs repo works the same way. I think it backs off from adding stale again as it gets bumped.
I'm not saying it's a perfect solution but for projects that deal with large amounts of issues it's workable.
jaggederest 17 hours ago [-]
They should just utilize the old joke about hiring: immediately throw away half the tickets, because you don't want to work on an unlucky ticket.
It'd save the run around and have the same ultimate effect. Or, we could properly work on tickets instead of making the gate "has enough time to follow up on this 14 days later"
fireflash38 8 hours ago [-]
Stale closing bugs is the worst.
Just because someone moved on from your broke ass product doesn't mean the bug was fixed. I can't count how many times I'll find an old bug still there years later closed with 5+ duplicate issues all linked back to the same closed github issue that was closed as stale.
It's just bullshit. Having a lower count on your issue tracker doesn't make the actual bugs disappear folks
watwut 11 hours ago [-]
The stale bots are such a crap.
Glyptodon 1 days ago [-]
What I see is that I have to read a bunch of stuff and go through a bunch of hassle to save money when the root of it is that if I tell an AI to do work on a task while I'm busy with something else and come back later I've doubled my cost because the cache expires too quickly?
chamsom 1 days ago [-]
As a driver I want to spend the majority of the day optimizing my truck's gasoline usage so that I can focus on optimizing my productivity for an outcome I am too far disconnected from to care about anymore.
datakan 9 hours ago [-]
The AI is so super intelligent that it can't optimize itself and instead burns through tokens needlessly. The people programming the AI are so good that they can write long winded articles about how to optimize their AI but can't get the AI to do these things by default.
It's all so tiring. I care less and less about Claude every single day because of the usage caps and the constant optimizing that has to be done. The whole point of AI was to get past this type of bullshit. They've failed miserably at their jobs.
StilesCrisis 7 hours ago [-]
Three years ago, a computer that can write working C++ on the first try would have been considered to be a miracle. It's amazing how quickly the goalposts move.
qaq 6 hours ago [-]
Dario with his "country of geniuses in a datacenter" talk is pretty instrumental in accelerating that
rambojohnson 4 hours ago [-]
AI has created a new tedious category of work: managing the AI itself.
HappMacDonald 6 hours ago [-]
I mean `printf("printf(\"Hello World!\\n\")");`, so...
hardbass 59 minutes ago [-]
Fail: Bad/legacy C library used. An LLM won't have generated C code if asked for C++.
agaj-nimm 3 hours ago [-]
[dead]
jnwatson 1 days ago [-]
Can anyone explain why the prefix cache is tied to effort?
I frequently run Fable at xhigh effort to run statistical modeling way above my undergraduate understanding. Claude Fable produces Masters-degree level output, and then I spend lots of round trips asking it to explain different parts to me.
The first part absolutely uses the extra effort, but the interrogation exercise is something a much simpler model, or the same model with much less effort, could answer.
hellohello2 1 days ago [-]
One trick is to simply as for a fast answer when talking to a high effort model, when working interactively. Sounds stupid but I do this all the time and it works. Just tell it you are working interactively now and need ultrafast answers with no thinking.
I would be really curious to know as well, why effort is linked to cache as its quite inconveniant. Is it possible the token used to indicate effort is only passed once at the start, not per thinking trace, or quite simply that different efforts have different model weights?
janalsncm 1 days ago [-]
I’m guessing that there’s a system prompt at the top telling the model about its reasoning budget. So when you switch reasoning effort it busts the cache.
cellularmitosis 11 hours ago [-]
Maybe switching effort routes you to a different rack of gpu’s which don’t have the cache
foota 22 hours ago [-]
Hmmm... Why wouldn't this be handled like other end of prompt things like the current mode?
SoMomentary 18 hours ago [-]
That makes sense to me. The output styles work the same way.
mojuba 10 hours ago [-]
This is all good to know, but funny how we are suddenly back to formal languages and commands. Aren't these things intelligent enough to figure these things out for us?
cactusplant7374 7 hours ago [-]
Codex is at least. The length of my prompts have decreased over time. Mostly I point it to relevant examples that already exist. It knows the drill.
nathanyz 1 days ago [-]
This feels like the Anthropic version of "You're holding it wrong" (1)
It feels worse. It's all noob level suggestions that any decent system would have optimized away already.
onlyrealcuzzo 1 days ago [-]
Step 1: use a different LLM that isn't 10x slower and 2-5x more expensive for the same level of quality.
BeetleB 1 days ago [-]
> @-mention files instead of naming them. The file gets attached to your message directly, which saves a Read call, or a search if Claude has to go find it.
I've heard it argued that this is an antipattern. If the file is large, it will read the whole file. With Read or something similar, it can do a targeted search and read only the relevant portion.
Is this still not the case?
Also, since they mention /context: Can anyone explain why /context takes so long to run? It usually takes several seconds, and I've had cases of it taking over a minute.
And why don't they just show the basics in a status line somewhere? Just a plain: "120K/200K tokens" I hate having to type /context just to get this. And I shouldn't need to install an extension.
sadfgknerknksdf 1 days ago [-]
You can make your own status line with something like "120K/200K Fable 5", it's nice.
9dev 1 days ago [-]
You can do tons more, really. There’s even a built-in /statusline command to modify it; mine shows both context window usage alongside session and weekly limit, all of them as progress bars. Just ask Claude to do it.
user43928 14 hours ago [-]
That's what is shown by default in the app.
I'll never understand why anyone would want to restrict themselves to a terminal interface instead, and I say this as a Vim user.
9dev 7 hours ago [-]
> That's what is shown by default in the app.
It is shown in Claude Desktop if you care to check in Settings > Usage, sure - but not in Claude Code, updated as you work with it.
> I'll never understand why anyone would want to restrict themselves to a terminal interface instead, and I say this as a Vim user.
I usually work with Claude in tandem, that is: The agent is actively working while I am either reviewing code or making changes myself in other places. So this means I want to work within my IDE. If you use Claude Code to vibe-code without interacting with the codebase at all, the Desktop app is probably fine, but for all other purposes, you'll need to run it either in the CLI or integrated into your editor.
And since I use different editors and don't like using either a sub-par U integration or locking myself into the harness of my IDE's vendor, I prefer the CLI as a universal way of running Claude Code.
user43928 5 hours ago [-]
Claude Desktop also contains Claude Code that you need to select specifically over the default or Cowork. I would refer to it as Claude app vs Claude (Code) CLI.
The app has a context wheel on the right of the chat box, showing both current conversation context breakdown and 5h as well as weekly limits.
bix6 6 hours ago [-]
> It is shown in Claude Desktop if you care to check in Settings > Usage, sure - but not in Claude Code, updated as you work with it.
What? Claude Code shows you context window and session limits right next to the text input box.
9dev 6 hours ago [-]
The default status line doesn't do that, but admittedly it has been a while since I used vanilla Claude Code. Either way, it definitely does not display weekly limits out of the box
> I've heard it argued that this is an antipattern. If the file is large, it will read the whole file. With Read or something similar, it can do a targeted search and read only the relevant portion.
I suspect you're right and that's why they haven't fixed @-search in the desktop app.
I actually don't find myself using it anymore since moving to the desktop app. I went from using various AI extensions in the IDE to Claude Code desktop.
But if that's accurate, why mention it in this post? Maybe because that's the first thing developers will try when moving away from a code editor?
19 hours ago [-]
zmmmmm 21 hours ago [-]
so i actually read beyond the TLDR and the reasons are:
- it gets attached early so fully cached, even if later cache is busted
- it gets included in every request automatically, so if your following requests are going to keep triggering File Read requests it will be much cheaper and faster to keep sending it
My main question is how this works if Claude itself keeps editing the file. Surely then you are sabotaging your own cache rather than helping it.
The things to add this way would be static files that you don't expect to change and to be highly relevant to the following requests. Especially if you want them to be mandatory reading and not just hope the agent will read it.
flonle 1 days ago [-]
[dead]
zmmmmm 19 hours ago [-]
What I want is a version of `/clear` that keeps the conversation but drops out things like bloated logs, error traces, etc that were only relevant in the immediate local context.
I guess compacting somewhat does that but I want something more explicitly that trims out these extremely bloated artefacts while maintaining in full the actual conversation history.
hetspookjee 16 hours ago [-]
Create a skill for this that you can invoke a new session in referencing your previous session id. Your instructions here read clear enough it seems to create it. Though a handover skill with this kind of behaviour in the same session might be more economical given the cache materials is already there.
StilesCrisis 6 hours ago [-]
/handoff
dizhn 12 hours ago [-]
Wouldn't that context with gaps where the output should be confuse the agent too much? At the very least they should be replaced with an explanation that sections were redacted. Otherwise I am imagining the agent will think the commando failed or it won't know how it fixed something.
apt-apt-apt-apt 1 days ago [-]
I'm finding that unexpected cache rewrites cost me huge.
I have 1h cache TTL set, and do nothing to cause rewrite (response in time, no model/effort/tool changes).
At 400K tokens in, I'll write a message, and /usage shows only a small increase in cache write. On the next message, cache writes shows 800K, and by the end, I often hit 2M cache writes with no explanation.
This seems to happen when: using /btw, asking it to review code, other random times. Anyone know what's going on?
AlexErrant 1 days ago [-]
To not answer your question, I have a claude stop hook that my status line uses to tell me how close my cache is to expiring https://pastebin.com/JLXUG16Q
It seems to be regular file edits rather than parallel tool calls.
I'm sitting on 1.6m cache write even now with 468k in /context. It drives my session costs above $100 regularly.
Can someone from Anthropic look into this?
Phemist 7 hours ago [-]
Incentives to optimize cache-usage are only aligned between anthropic and you, dear user, when there is not enough compute available to serve the tokens fresh. Apparently enough compute is now available that frugal usage is no longer a requirement and user cost/profit maximization is now the name of the game. I would expect to see more and more "cache-busting" strategies implemented in order to both eek out marginal performance increases and to heavily increase profit.
radlad 1 days ago [-]
I have long suspected but not confirmed that /btw uses a lower tier model like Haiku.
Depending on how you're triggering reviews, you may be using a sub-agent?
apt-apt-apt-apt 24 hours ago [-]
I don't use sub-agents, only the normal linear message-response flow.
crthpl 18 hours ago [-]
it does not use haiku. you can just ask Claude (possibly not fable BC of cyber) to reverse engineer the obfuscated JS.
StilesCrisis 5 hours ago [-]
I've had Opus refuse to do work due to cybersecurity concerns as well. Quite frustrating when you're trying to fix an exploit--as soon as it reads the bad code, it just shuts down! I ended up cutting out more and more bits of the code until eventually Opus relented.
alekstret 9 hours ago [-]
Here is my working flow, confirmed by more than 400 pr merged over the last 4 months. More than half of them were following my current strategy:
1. My agent writes code.
2. Then it creates tests and verifies that all of them actually work, not just pass. To do this, my agent writes the test, then it deletes the code it covers, reruns the test, confirms it goes red, and finally puts the code back.
3. I receive the ready-to-test code and environment setup.
4. I check that the business logic works as I expected it to be on a working product. Here we usually do several iterations of coding and bug fixing.
5. When the manual part is finished, the agent starts an external review using /code review skill. At that stage, it makes some additional fixes and corrections to the tests.
6. Finally, a branch is ready to be merged. We start CI/CD and wait until the run finishes successfully.
That's what I actually use because it generally works.
Note about only docs PRs:
I just ask the agent to make the changes, then it runs the / code review skill, and then we merge the branch into main without CI running.
ultrasandwich 7 hours ago [-]
> To do this, my agent writes the test, then it deletes the code it covers, reruns the test, confirms it goes red, and finally puts the code back.
Why? Are you aware of red/green/refactor?
AlfeG 8 hours ago [-]
I tend to do code review in separate session. So that context is not affecting judgments.
tarekabi 1 hours ago [-]
[flagged]
mccoyb 1 days ago [-]
I mean, it feels hard not to laugh at this type of blog post. My cynical interpretation is that this is a type of passing the buck to engineers in enterprise settings ("Stop spending tokens. Did you read the value maximization blog post? It is your fault.")
Oh yes, Claude will do all sorts of different things -- it depends on how you use it! You should totally learn all of these little finicky things ... because now completing your tasks cost money. It's not "free" anymore haha like when you used your old text editor, what are you a grandpa?
Oh, and those things will definitely change, as we (the priests of Claude) are vibe coding the system you use to do your little "tasks" ... right, you can't see how it works ... the code is not available. It's all good, just trust us -- we're totally looking out for you.
I mean it is utterly ridiculous to talk around this model of development. There are so many walls between you and doing the thing you want to do.
Agents are great, but the notion of "best tricks" for how to best use an opaque costful tool which will, by all odds, be completely different in a few months time is quite funny.
You know what won't change? A fucking text editor. Or your pi config, or a local model you run and trust.
DangitBobby 1 days ago [-]
Not sure how this level of cynicism is even remotely warranted. The post helps people who don't understand LLMs very well get the most out of Claude. Your incentives here are actually aligned with Anthropics since both of you want fewer tokens inputted and outputted per task completed.
mccoyb 1 days ago [-]
Perhaps my enterprise cynicism is not warranted, but my other comments refer to accurate descriptions of reality: Anthropic wants to place their opaque system between you and any computational task that you wish to perform. Do you contest this or think it is not accurate?
Why do you think that Anthropic wants fewer tokens inputted and outputted?
DangitBobby 23 hours ago [-]
Because they sell subscriptions and tokens cost them compute, and their margin lives in the difference between what your subscription pays in and what you cost them in compute.
They have also been supply constrained on compute and if users cost them less in compute they can more subscriptions and less customer frustration.
I agree they want you to have a subscription. That doesn't mean they aren't aligned with their subscribers.
fcarraldo 21 hours ago [-]
Subscriptions are a very small part of their overall revenue (estimates have been between 5% and 20% based on financial reporting). Enterprise users are charged per-token, and maximal input/output tokens nets them maximal revenue.
DangitBobby 20 hours ago [-]
They still want you to hit the cache because their margin is higher on cache hits. That's actual compute they don't have to pay for and they don't have to have capacity for because they are supply limited on the compute side.
And the unit economics need to be there because there are competitors in the space. They can't just skin you on tokens or you'll jump ship.
NamlchakKhandro 16 hours ago [-]
1000% warranted. not sure what level of peak echo bubble you live in where this level of critiscism feels like you need to defend a ONE TRILLION DOLLAH company.
seriously... priorities yeah?
DangitBobby 7 hours ago [-]
I don't typically decide things on the basis of whether it's a trillion dollar company. If it were a blog post about puppy slaughter I wouldn't be defending it.
mnahkies 1 days ago [-]
Through my weekend experiments, I've found I can get way better outcomes, and an order of magnitude less cost with my slapped together sandboxed omp setup plus ZDR openrouter models (DeepSeek, Kimi, etc) than I've ever seen from Claude Code at work.
Everything is version pinned and a deliberate choice to change, and a git revert away from changing back.
TBF the models may change underneath me to some extent still, but the cost benefit of running them myself doesn't pan out yet (for agentic coding at least, don't have enough local vram to get a usable context window and generation speed, self hosting on runpod or similar isn't economically sensible for my current consumption though I have tinkered with it)
bmitc 17 hours ago [-]
I fully agree. All these blog posts are basically features they should be implementing. They advertise they are replacing software development, but then these tools require a massive amount of overhead akin to having to train new hires. But these tools never actually learn and are not trainable, and Anthropic releases a blog post every six months about how to re-invent your workflow. Even the author of Claude Code just told everyone they should delete all their `CLAUDE.md` and skills every six months.
It's wildly lazy.
runeblaze 1 days ago [-]
dude, if you try to do harness development yourself you will realize that most things said in this blogpost is shared with any ${sufficiently_advanced_harness}. this is not really claude-specific, this is just how this class of tools, OSS or not, works
mccoyb 1 days ago [-]
That's not my complaint. I know well the concerns of agent harnesses.
My complaint is that this is a low-dimensional projection of a system which I have no insight into, and therefore, I cannot evaluate the tips myself against their source.
Am I to believe the creators, knowing full well that the source will, as Boris Cherny put it in a recent interview, be deleted and rewritten from scratch at the release of the next big model?
Further: I'm responding to content in the blog post itself:
> Until pretty recently, the tools you wrote code with were a flat fee (or free). Your editor cost the same whether you fixed one test or fifty that afternoon, so an individual task didn't really have a price of its own.
I find this type of prose ridiculous. It conveys "this is the way things are now, get used to it".
Does that make sense?
runeblaze 24 hours ago [-]
> which I have no insight into
i guess you do? claude code is the commercial closed sourced version provides by ant. reading a mini version of vllm or sglang and then read codex source code or grok build source code will teach you all things taught by this article, fully in the open
it is like saying that you have no insights into some $commercial_db_system which is kinda true but imagine if the article is to teach you indices, query normalization, etc..
bluefirebrand 1 days ago [-]
> I find this type of prose ridiculous. It conveys "this is the way things are now, get used to it".
This is absolutely what AI companies and AI lovers want you to believe
csallen 1 days ago [-]
I'm trying to understand your point of view, but it kind of just sounds like you're against learning how to use tools efficiently?
I mean, agentic coding software is hardly the first tool to exist where learning some idiosyncrasies of how to use it well can result in more efficiency and cost savings.
and as is normal for hosted models, almost everything... based on load flucation they may even send your prompt to a quantised model
csallen 1 days ago [-]
I think I'm happy about the first two. The third I suppose I care less about, just because I've kind of become used to it from decades working on the internet where many businesses/tools/apps are more like services and less like physical tools that never change.
RossBencina 1 days ago [-]
> The third I suppose I care less about, just because I've kind of become used to it
I've never become used to it. My impression is that the constant churn has accelerated. Plausible drivers are (1) normalize novelty as desirable (like fast fashion), (2) product developer/designer incentive structures that reward revolutionary change over progressive refinement. The global switch to subscription models and continuous deployment didn't help.
> more like services and less like physical tools that never change.
I'm not sure that constant change is a characteristic feature of services, especially not professional services.
It used to be that you bought a piece of software and used that version until you decided it was worth upgrading, like a particular physical tool. The software still evolved, just like the design of physical tools can, in principle, evolve.
All that said, agentic AI tooling is evolving so rapidly I'm not sure an expectation of stability is realistic.
Wowfunhappy 1 days ago [-]
> It used to be that you bought a piece of software and used that version until you decided it was worth upgrading, like a particular physical tool.
But Claude is running on someone else's computer, not yours, so it's not Photoshop so much as AWS. Or a rented server farm, if AWS is too new school for you. Of course there's an ongoing cost! And if you configure the server to use more electricity, you get billed more.
If you want to do agentic tooling locally, you can do that—the models aren't quite as good, but they're not bad either. But be warned, for the large models you're going to have to acquire some serious hardware, to the point where you may wish you'd chosen to just rent it instead!
mccoyb 1 days ago [-]
I agree that the third seems to be implied by industry, but I'd argue that it's not clear that it is necessary -- and it is subtle whether or not it is beneficial?
My contention is that we should be building towards less churn, not more. I'm aware that some churn is the cost of engaging in any sort of enterprise, but I'm deeply suspicious of an AI company inserting themselves between me, and the tasks I wish to do with my device -- with a completely opaque system that I can't really "learn".
dist-epoch 1 days ago [-]
I'm curious if you also laugh at articles about how to reduce your AWS bill, or how to add indices to Postgres such that you can run it on cheaper hardware.
mhitza 21 hours ago [-]
You can generally understand your AWS cloud usage, and waste can be self evident with their existing tools. Not at all with llms.
A postgres index post is unlikely to reach front page. It's already part if the docs, and should include more context to be read worthy.
They are not equal comparison.
This before the fact that there is no guarantee that a model follows your agent instructions (plenty of easy to reach for research on it), and you also get suggestions by devs at these companies to wipe parts of your model's instructions because the model is better now tm.
If cloud providers change their billing quasi monthly, and if you'd need to fiddle with your indexes every couple of days. I'm not sure we'd be using them as much.
There is interesting information about the inference pipeline, but almost too late to the party (by at least a year), and for which audience? Techies understand in broad strokes the tech if they are interested, normies will definitely not read it.
All that to say, that yes, it's worth having a laugh. If for nothing else, as a release valve for all the problems they create in the real non-VC world.
Anthropic is IPOing in October according to news, you might be interested in investing.
dist-epoch 12 hours ago [-]
I thought that software engineers were supposed to do, you know, engineering - solving hard problems, dealing with uncertainty.
But you might be right, engineering around the difficult LLM primitive might be a task which is just too hard for your typical software engineer, as you said, they want predictability, hand holding, determinism, most are unable to deal with the real world which is not a spherical cow in a vacuum. So I guess they can stick to simple very well understood primitives like EC2 or Postgres and leave dealing with LLMs for others.
mhitza 9 hours ago [-]
I'm a pleb developer, don't have the smarts, the prestige or the salary of those working at BigTech.
The most well paid people in the industry brought us here. And "here" is very much as fuzzy as last year with better harnessing towards the local optima. And I say local optima because even the perceived capabilities have slowed down, nevermind the benchmark numbers which are in aggrement.
The best paid engineers in the world, with almost no practical budget limit, still deliver shoddy quality software with AI. Is that not fact? And if it is what does that say for the rest of us.
You are allowed to believe. I'm still waiting for the beneficial results, not only those that benefit griefters, hackers and scammers. AI has been a huge boon there.
Reality will materialize and markets will redress hopefully once they go public. Which they very much seem to be hesitant to do right now.
dist-epoch 8 hours ago [-]
> I would expect better results than what we have today.
Not sure what your baseline was, if you said 10 years ago "in 2026 you'll be able to describe an app into the microphone, and the computer will write by itself in one day 50k lines of code to implement it, in a language and tech stack of your choosing, costing $200, and it will sort-of-work, and it will be at least as good as a junior-level programmer writing it from the same requirements in 3 months", most people would have said "implausible, that's at least 50 years away"
mhitza 50 minutes ago [-]
Yes, it is very impressive what they can do by recycling copyrighted material. They are more impressive when they are not used in agentic contexts. Though that doesn't sell hype anymore to inflate valuation.
For more than a year now I was renting a limited GPU server for ~300$/month to learn, experiment, research and build internal tooling around open weight models. Thinking they are tools with potential and buying the exaggerated marketing are different things.
My history of comments on HN lands often on both providing what I believe to be my insights working with LLMs and calling out exaggerations, stupid terms of service, and the other mishaps in the field. You are free to browse them if you'd like to see my broader opinion.
skydhash 4 hours ago [-]
So where’s the value in that description? Something that can justify the mania and size of investment we have currently?
Banditoz 23 hours ago [-]
I think you're conflating two different things here. I am not aware of any DB optimization articles that say "trust me bro, throw your data, don't build indices, it'll Just Work™!"
NoDodgeQuestion 1 days ago [-]
Bro: superintelligent machine line go up AI AGI software solved automate everything
Also bro: Run /clearbetween tasks. This prevents prior irrelevant context from being sent back to the model, which can reduce token usage.
Set your model and effort level before you start. Changing either one mid-conversation can bust your prompt cache, which can increase token cost.
@-mention files instead of naming them. The file gets attached to your message directly, which saves a Read call, or a search if Claude has to go find it.
Add quiet flags to noisy commands, or run them in a subagent. Command output is added to the conversation just like a file, and stays there for the rest of the session.
Run /context once in a fresh session. It shows what's loaded (CLAUDE.md, MCP tool definitions), so you can cut out anything unnecessary.
/compact before you take a break from your keyboard. The prompt cache expires after an hour, and summarizing a conversation is much cheaper while it's still cached.
NoDodgeQuestion 1 days ago [-]
Author not bro, sorry misgender
cyber_kinetist 14 hours ago [-]
It has been a while since bro has become a gender-neutral term, particularly in younger circles...
pdpi 1 days ago [-]
I'd argue that women can be bros too, especially when using the word in this sense.
NamlchakKhandro 16 hours ago [-]
There are no women on the internet.
I remember when the internet was an exchange of ideas instead of using gender to justify value of bad ideas
runeblaze 22 hours ago [-]
puts on my etiquette hat
don’t do that, it is weird, use “bruh” or “dude”
DangitBobby 1 days ago [-]
What do you want? Lacking omniscience, even the smartest superintelligence imaginable has to do more thinking to deal with worse inputs.
DangitBobby 20 hours ago [-]
Still waiting to hear what you want.
lovasoa 13 hours ago [-]
I don't understand why changing effort levels busts the cache. Couldn't effort levels be a decoding-only thing where they just change the probability of the <end of thought> token? Are they literally adding a hidden system prompt that says "effort level: $level" ?
namjh 7 hours ago [-]
In that way the autoregressive nature of LLM won't let itself "plan" to reason with the intended budget. It doesn't "look ahead".
Philpax 13 hours ago [-]
> Are they literally adding a hidden system prompt that says "effort level: $level" ?
Obviously it‘s something being put in the context that can not be taken out of it anymore.
fwlr 16 hours ago [-]
Until pretty recently, the tools you wrote code with were a flat fee (or free) … [so] an individual task didn't really have a price of its own … [but] with agentic coding tools like Claude Code, it does.
I’ve heard this anti-AI thesis before, but it’s certainly novel to read it on “claude.com”.
ahurmazda 1 days ago [-]
What’s the point of running /clear vs starting a brand new session. At least with the latter I have session history, no? Pardon my ignorance since Claude isn’t my primary driver
As far as I've seen /clear is the same thing as starting a new session.
If you type /resume right after clear, the first thing in the list is the session you just cleared.
andai 1 days ago [-]
Not 100% sure what clear does, but starting a new session invalidates the cache*, whereas I assume clear only removes part of the context, so it should be cheaper and faster.
* In theory the system prompt is always the same and should therefore be cached, but in practice there's some dynamic strings in there so it doesn't work that way. (Unless they changed this recently.)
olsondv 1 days ago [-]
After having used Codex for a promotional month, and now using Claude, Claude is not as efficient with finding relevant information. I can give it the one file it should be using and then it goes off and greps parent directories for more context. It’s also incredibly slow at producing results because of this side work. In this article, it seems like they are catching up to what GitHub copilot users had already been doing since the cost restructuring in June.
brachkow 6 hours ago [-]
Considering that I'm mostly unable to reach the limit of my X5 subscription and we have 1M context, it is a guide to maximize Anthropic PnL pre-IPO
ramraj07 6 hours ago [-]
This very much applies to most enterprises. These unlimited subscriptions are only available to individuals and teams less than 150 folks..
brachkow 33 minutes ago [-]
API is highly overpriced and keeps Anthropic profitable or at least close to break even, as was reported recently. Why would you think they will be interested in advising enterprises in cutting their bills pre-IPO?
wjakob 21 hours ago [-]
When rewinding to an earlier turn, what if that turn is more than 1hr old? Can this cause KV-cache misses compared to continuing the conversation?
docheinestages 20 hours ago [-]
Anthropic should build a harness (and model) that smartly takes care of all these points. Not requiring the user to do the manual work. All I see are excuses because they cannot handle the load and enforce strict quotas on users, all while OpenAI constantly resets their quotas.
With Qwen 3.8 27B, we're one step closer to on-device LLMs that can replace subscriptions.
Petersipoi 17 hours ago [-]
The amount that OpenAI resets their quotas is nuts. It's like, every 2 days I swear. Feels so fucking good. Whenever I think about switching my $200 plan back to Claude for a month I'm reminded that they still have a 5 hour usage limit, which feels so absurd now that I've used Codex for a couple of months.
maCDzP 11 hours ago [-]
I have had some success with using Ollama cloud and just instructing Claude Code to hand off tasks to Ollama because of tokens economics.
pzo 1 days ago [-]
> Set your model and effort level before you start. Changing either one mid-conversation can bust your prompt cache, which can increase token cost.
I know we supposed to do this but is there any particular reason why such things cannot be supported? I thought its running on same model just different settings like reasoning. This would be super useful.
22 hours ago [-]
22 hours ago [-]
8note 20 hours ago [-]
how many times it stays there i think doesnt give the best comparison
if youre working on the same codebase, that cache stays quite relevant, and i dont think they make the case that clearing and reading the same couple files over and over again is cheaper that relying on it already being cached. same with doing some of the same teaching claude the right way to approach changes in that codebase again and again.
what would be nice is pulling back and reusing an earlier part of the cache for the later two tasks, but claude code doesnt make that particularly easy, and using an LLM to pick where to go back to isnt really gonna save much when it reads all the same text again.
tosh 8 hours ago [-]
high leverage for most agents:
- review system prompt + cut it down or remove completely
- review agents.md file(s), check which ones are loaded, remove or improve them
- review context spam from tools, skills etc, de-activate all, see what needs re-adding
- review past sessions to see where tokens get wasted
more advanced:
- keep sessions short (be conscious about compaction)
- form a habit of starting new sessions
- deliberately practice how to effectively get the right context into a new session (vs hanging on to a 'good' session)
- you can ask the agent to write the essential context into a .md file and have the new session read that
- learn about forking sessions
- experiment with starting sessions from a custom-built history/context
a good agents.md file can be small and still effective re helping the agent navigate the code base
that said: you will surprised by how well current models can navigate (way better than last year!)
dpkirchner 19 hours ago [-]
TIL the prompt cache lasts 1 hour. I thought it was reduced to 5 minutes.
moebrowne 11 hours ago [-]
It depends. If you have a Claude Code subscription then it defaults to an hour, if you use Claude via an API or third party then it defaults to 5 minutes. You can opt-in to the 1 hour TTL but it obviously costs more.
> 5-minute cache write tokens are 1.25 times the base input tokens price
> 1-hour cache write tokens are 2 times the base input tokens price
I think it depends on your subscription? I just have the 20$ and the cache is only 5 minutes. I’ve got the timer in my status line via ccstatusline and unless it is wrong, it says 5 minutes.
aleksiy123 1 days ago [-]
Is it possible to have some kind of script to keep your cache warm, or auto compact or something.
I sometimes just leave some goals or something running before I go to bed or out and I don’t want to pay the cache text when I come back.
cube00 24 hours ago [-]
If everyone does it they lose the memory savings they're getting by expiring the cache.
8note 20 hours ago [-]
which is to say that they set their ttl too short, it should be longer than people spend at lunch
swingboy 18 hours ago [-]
How about Anthropic just be more generous with their usage limits instead?
tolugenius 1 days ago [-]
Part of the cynic in me just wants to ask "why not make a better harness by default?" The other cynic in me knows I'm about to see a hundred post on 'HOW TO 10X CLAUDE" from the ai bros and I'm already tired.
I guess if I had to ask something (as someone who doesn't use CC as their daily driver), how much control do you have on subagents and roughly how do define or know when a session is getting too long? I know the answer is "when the model is getting worse" but worse is doing a lot of lifting in that sentence.
nathanyz 1 days ago [-]
Yeah, I sort of feel like they could just do this in Claude Code for us in some way. I mean they already run some mini classifier on whether a given prompt is waiting on input, so they could do the same to detect some of these cases, and just handle it.
If you have to explain that someone is "holding it wrong"(1), that is product error, not a user error.
I do all that, but an 'AS-BUILT' full review of my project still eats 3x my 5 hour budget on max 100€. Meanwhile, my 20€ GPT never hit a limit. Different, but just saying.
andai 1 days ago [-]
Austerity on tap!
apercu 1 days ago [-]
I didn't read the whole thing, but I got my back up at the headline and my first reaction is now even the "AI" companies are telling you that "you're holding it wrong".
I mean, they told us "just talk naturally to the AI because it's so much smarter than all you meatbags" and now it's “for best results, please learn to manage context windows, prompt caching, cache invalidation, model switching, output verbosity and when to manually clear or compact your session.”
I get it, but it seems like the "PRODUCT" should be doing this shit. I.e., the PRODUCT is getting less efficient because I didn't manually manage its context correctly and now it's MY fault.
Edit: i.e., for e.g. Doh. Even the robots get that right. Sigh.
QuercusMax 1 days ago [-]
"It is difficult to get a man to understand something, when his salary depends on his not understanding it." -Upton Sinclair
Anthropic has no incentive to make their products more efficient as long as they're selling them by the token.
DangitBobby 1 days ago [-]
They are selling to many customers by subscription.
DangitBobby 1 days ago [-]
There is no shortage of literature on how to communicate effectively with real people.
apercu 24 hours ago [-]
I'm doing my best to interpret your comment but if I parse it it seems like you think my complaint is "I shouldn't have to express myself clearly" when really its "Why am I being asked to understand and manually manage the implementation details of the product in order to keep it working efficiently?".
Unless I am fully not understanding your comment and you don't actually mean “humans require communication skills too” which in honesty feels orthogonal to my complaint.
DangitBobby 23 hours ago [-]
> I mean, they told us "just talk naturally to the AI because it's so much smarter than all you meatbags" and now it's “for best results, please learn to manage context windows, prompt caching, cache invalidation, model switching, output verbosity and when to manually clear or compact your session.”
It's true both that it can be smarter than all us meat bags and that talking to it a certain way gets better results. On some of the things it's a limitation of the technology and on some of the others it's just how information and effort work in any context. I don't see it as orthogonal to your complaint, I see your complaint as misplaced frustration, like Anthropic invented GIGO and compute so they'd have an excuse to write a blog post.
apercu 22 hours ago [-]
Huh. And to me it’s like “I can’t believe Anthropic has the audacity to post that and I can’t believe my peers defend it”
Oh well.
DangitBobby 20 hours ago [-]
And I in turn can't believe your response to the post. Different strokes for different folks. I understand that HN fucking hates everything and everyone, especially big model providers at the moment, but this shit doesn't make any sense to me and you people are starting to actually make me crazy.
apercu 7 hours ago [-]
"Dang it, Bobby" :)
dude250711 22 hours ago [-]
"TL;DR ..."
If only they had some kind of technology that could make a judgement and automate those actions...
grey-area 14 hours ago [-]
Well, they claim it is capable of such tasks, but in reality, it can’t reliably do so, or they would address them.
NamlchakKhandro 16 hours ago [-]
meah Claude is peak trash. Really fucking upset that they ban you for using a superior harness
Basically:
- /handoff file creates a short document with the important context from your current session and maybe next steps as checklist.
- You can then start a fresh session with /continue file
- You can also hand the work from Claude to ChatGPT, or the other way around. Very useful at time of session limits.
- Plus your handoff files becomes a useful piece of project memory that you can reference later.
I find this much more useful than /compact or /clear because the context is saved in something portable instead of being tied to one session and i've seen better results doing this every 20 messages or so than running long sessions.
I've been doing this since I started agentic development, and have a whole framework based on this; Simply put I define workflow s that output templated files for everytype of tasks that happens in development.
It's a powerful pattern I'd recommend everyone.
I now feel a bit silly but I reinvented the wheel during my last lil project and indeed found it very powerful. A variant of it is the "implementation (handover) prompt" when I conclude the planning session with plans and design documents, updated handoff, clean tree and a file for a new Opus implementing orchestrator (unusually do a single, highly specific implementer and a single, highly specific tester).
After implementation and task-level tests I end up with a long and very detailed implementation progress ledger and a summary findings from the orchestrator.
Then in the new session I do the whole branch tests.
Works really well, uses much less tokens than any other approach with more of Opus and very little repetitions/corrections.
- Prefix-numbered sets of documents, keeping them clustered visually and easily referenced by humans and agents. (124.5 = docset 125, phase 5) - Each doc gets a suffix. Most of the time STRATEGY is the first doc. Sometimes NOTES, DISCUSSION, or HANDOFF though. - Once the strategy doc is comprehensive, the multi-phase checklist PLAN doc can emerge. - During execution, the agent drives the PLAN while purposefully expanding lightly scaffolded phases before entry. - As the user, I drive a PROGRESS doc during the PLAN execution. This tracks the progression of my own questions and the important work summaries that I need for tracking the current work trajectory and for historic analysis over past decision-making. - Any concern that exceed the scope of the active docset numbers calls for a new one. - A "docs/archive" folder is kept where I sweep docsets into numbered eras.
I have aliases and templates too, but DDD is so simple that it's overkill.
Any modern reasoning agent can take the plain explanation above and understand you effortlessly when you say "open a new docset strategy" or "read docset 125 and proceed".
It might be even better if the harness were to automatically write a handoff note under some circumstances? If the user is away and the cache is going to expire, that would be a good time to do it.
I've been asking Claude to remember important points from our session, or future tasks. Then /clear and continue.
[0] https://news.ycombinator.com/item?id=48916512
I managed even the orchestrator to NOT read the plan whole, at once, but in sections.
The most useful thing is the task ledger the task agent leaves behind, which alongside its structured status message makes a very resilient handoff between all stages.
And I think plan files should focus on general ideas and invariants, not do “implementation as prose”. That way they perform as mini-ADRs that are useful historically, especially to mine why the system is the way it is.
Love Claude, but the @ mention is broken in the desktop app. For the same project if I type the same query "@ephem" I get:
CLI: https://imgur.com/a/VZMUCOa (good, relevant results)
Desktop: https://imgur.com/a/QLSo4Ms (bad, irrelevant)
Opened issue for this and it was automatically closed:
https://github.com/anthropics/claude-code/issues/71421
I could have written the issue better (using CLI as comparison instead of VS Code). But, no doubt in my mind Claude could fix this itself in a minute.
Clarification: It wasn’t closed on submission though. It sat open ~17 days, a bot marked it stale, and it closed when nobody responded to the stale label.
The two-phase thing is the part I didn’t know until recently: the stale label is basically asking “is this still relevant?”, and answering it makes the bot back off next time around. nixpkgs does the same. Bumping feels wrong on most trackers, agreed, but at this issue volume I don’t know what else works.
Anyway a comment should reopen it. Your CLI vs desktop screenshots are a better repro than most things in that tracker.
I'm not saying it's a perfect solution but for projects that deal with large amounts of issues it's workable.
It'd save the run around and have the same ultimate effect. Or, we could properly work on tickets instead of making the gate "has enough time to follow up on this 14 days later"
Just because someone moved on from your broke ass product doesn't mean the bug was fixed. I can't count how many times I'll find an old bug still there years later closed with 5+ duplicate issues all linked back to the same closed github issue that was closed as stale.
It's just bullshit. Having a lower count on your issue tracker doesn't make the actual bugs disappear folks
It's all so tiring. I care less and less about Claude every single day because of the usage caps and the constant optimizing that has to be done. The whole point of AI was to get past this type of bullshit. They've failed miserably at their jobs.
I frequently run Fable at xhigh effort to run statistical modeling way above my undergraduate understanding. Claude Fable produces Masters-degree level output, and then I spend lots of round trips asking it to explain different parts to me.
The first part absolutely uses the extra effort, but the interrogation exercise is something a much simpler model, or the same model with much less effort, could answer.
I would be really curious to know as well, why effort is linked to cache as its quite inconveniant. Is it possible the token used to indicate effort is only passed once at the start, not per thinking trace, or quite simply that different efforts have different model weights?
1) https://www.wired.com/2010/06/iphone-4-holding-it-wrong/
I've heard it argued that this is an antipattern. If the file is large, it will read the whole file. With Read or something similar, it can do a targeted search and read only the relevant portion.
Is this still not the case?
Also, since they mention /context: Can anyone explain why /context takes so long to run? It usually takes several seconds, and I've had cases of it taking over a minute.
And why don't they just show the basics in a status line somewhere? Just a plain: "120K/200K tokens" I hate having to type /context just to get this. And I shouldn't need to install an extension.
I'll never understand why anyone would want to restrict themselves to a terminal interface instead, and I say this as a Vim user.
It is shown in Claude Desktop if you care to check in Settings > Usage, sure - but not in Claude Code, updated as you work with it.
> I'll never understand why anyone would want to restrict themselves to a terminal interface instead, and I say this as a Vim user.
I usually work with Claude in tandem, that is: The agent is actively working while I am either reviewing code or making changes myself in other places. So this means I want to work within my IDE. If you use Claude Code to vibe-code without interacting with the codebase at all, the Desktop app is probably fine, but for all other purposes, you'll need to run it either in the CLI or integrated into your editor.
And since I use different editors and don't like using either a sub-par U integration or locking myself into the harness of my IDE's vendor, I prefer the CLI as a universal way of running Claude Code.
The app has a context wheel on the right of the chat box, showing both current conversation context breakdown and 5h as well as weekly limits.
What? Claude Code shows you context window and session limits right next to the text input box.
I suspect you're right and that's why they haven't fixed @-search in the desktop app.
I actually don't find myself using it anymore since moving to the desktop app. I went from using various AI extensions in the IDE to Claude Code desktop.
But if that's accurate, why mention it in this post? Maybe because that's the first thing developers will try when moving away from a code editor?
- it gets attached early so fully cached, even if later cache is busted
- it gets included in every request automatically, so if your following requests are going to keep triggering File Read requests it will be much cheaper and faster to keep sending it
My main question is how this works if Claude itself keeps editing the file. Surely then you are sabotaging your own cache rather than helping it.
The things to add this way would be static files that you don't expect to change and to be highly relevant to the following requests. Especially if you want them to be mandatory reading and not just hope the agent will read it.
I guess compacting somewhat does that but I want something more explicitly that trims out these extremely bloated artefacts while maintaining in full the actual conversation history.
I have 1h cache TTL set, and do nothing to cause rewrite (response in time, no model/effort/tool changes).
At 400K tokens in, I'll write a message, and /usage shows only a small increase in cache write. On the next message, cache writes shows 800K, and by the end, I often hit 2M cache writes with no explanation.
This seems to happen when: using /btw, asking it to review code, other random times. Anyone know what's going on?
_entirely_ vibecoded don't @ me.
It seems to be regular file edits rather than parallel tool calls.
I'm sitting on 1.6m cache write even now with 468k in /context. It drives my session costs above $100 regularly.
Can someone from Anthropic look into this?
Depending on how you're triggering reviews, you may be using a sub-agent?
1. My agent writes code.
2. Then it creates tests and verifies that all of them actually work, not just pass. To do this, my agent writes the test, then it deletes the code it covers, reruns the test, confirms it goes red, and finally puts the code back.
3. I receive the ready-to-test code and environment setup.
4. I check that the business logic works as I expected it to be on a working product. Here we usually do several iterations of coding and bug fixing.
5. When the manual part is finished, the agent starts an external review using /code review skill. At that stage, it makes some additional fixes and corrections to the tests.
6. Finally, a branch is ready to be merged. We start CI/CD and wait until the run finishes successfully.
That's what I actually use because it generally works.
Note about only docs PRs: I just ask the agent to make the changes, then it runs the / code review skill, and then we merge the branch into main without CI running.
Why? Are you aware of red/green/refactor?
Oh yes, Claude will do all sorts of different things -- it depends on how you use it! You should totally learn all of these little finicky things ... because now completing your tasks cost money. It's not "free" anymore haha like when you used your old text editor, what are you a grandpa?
Oh, and those things will definitely change, as we (the priests of Claude) are vibe coding the system you use to do your little "tasks" ... right, you can't see how it works ... the code is not available. It's all good, just trust us -- we're totally looking out for you.
I mean it is utterly ridiculous to talk around this model of development. There are so many walls between you and doing the thing you want to do.
Agents are great, but the notion of "best tricks" for how to best use an opaque costful tool which will, by all odds, be completely different in a few months time is quite funny.
You know what won't change? A fucking text editor. Or your pi config, or a local model you run and trust.
Why do you think that Anthropic wants fewer tokens inputted and outputted?
They have also been supply constrained on compute and if users cost them less in compute they can more subscriptions and less customer frustration.
I agree they want you to have a subscription. That doesn't mean they aren't aligned with their subscribers.
And the unit economics need to be there because there are competitors in the space. They can't just skin you on tokens or you'll jump ship.
seriously... priorities yeah?
Everything is version pinned and a deliberate choice to change, and a git revert away from changing back.
TBF the models may change underneath me to some extent still, but the cost benefit of running them myself doesn't pan out yet (for agentic coding at least, don't have enough local vram to get a usable context window and generation speed, self hosting on runpod or similar isn't economically sensible for my current consumption though I have tinkered with it)
It's wildly lazy.
Am I to believe the creators, knowing full well that the source will, as Boris Cherny put it in a recent interview, be deleted and rewritten from scratch at the release of the next big model?
Further: I'm responding to content in the blog post itself:
> Until pretty recently, the tools you wrote code with were a flat fee (or free). Your editor cost the same whether you fixed one test or fifty that afternoon, so an individual task didn't really have a price of its own.
I find this type of prose ridiculous. It conveys "this is the way things are now, get used to it".
Does that make sense?
i guess you do? claude code is the commercial closed sourced version provides by ant. reading a mini version of vllm or sglang and then read codex source code or grok build source code will teach you all things taught by this article, fully in the open
it is like saying that you have no insights into some $commercial_db_system which is kinda true but imagine if the article is to teach you indices, query normalization, etc..
This is absolutely what AI companies and AI lovers want you to believe
I mean, agentic coding software is hardly the first tool to exist where learning some idiosyncrasies of how to use it well can result in more efficiency and cost savings.
- I'm happy to learn how to use tools efficiently
- I like to be able to inspect my tools
- I'm against tools changing underneath me
Are you against any of these points?
and as is normal for hosted models, almost everything... based on load flucation they may even send your prompt to a quantised model
I've never become used to it. My impression is that the constant churn has accelerated. Plausible drivers are (1) normalize novelty as desirable (like fast fashion), (2) product developer/designer incentive structures that reward revolutionary change over progressive refinement. The global switch to subscription models and continuous deployment didn't help.
> more like services and less like physical tools that never change.
I'm not sure that constant change is a characteristic feature of services, especially not professional services.
It used to be that you bought a piece of software and used that version until you decided it was worth upgrading, like a particular physical tool. The software still evolved, just like the design of physical tools can, in principle, evolve.
All that said, agentic AI tooling is evolving so rapidly I'm not sure an expectation of stability is realistic.
But Claude is running on someone else's computer, not yours, so it's not Photoshop so much as AWS. Or a rented server farm, if AWS is too new school for you. Of course there's an ongoing cost! And if you configure the server to use more electricity, you get billed more.
If you want to do agentic tooling locally, you can do that—the models aren't quite as good, but they're not bad either. But be warned, for the large models you're going to have to acquire some serious hardware, to the point where you may wish you'd chosen to just rent it instead!
My contention is that we should be building towards less churn, not more. I'm aware that some churn is the cost of engaging in any sort of enterprise, but I'm deeply suspicious of an AI company inserting themselves between me, and the tasks I wish to do with my device -- with a completely opaque system that I can't really "learn".
A postgres index post is unlikely to reach front page. It's already part if the docs, and should include more context to be read worthy.
They are not equal comparison.
This before the fact that there is no guarantee that a model follows your agent instructions (plenty of easy to reach for research on it), and you also get suggestions by devs at these companies to wipe parts of your model's instructions because the model is better now tm.
If cloud providers change their billing quasi monthly, and if you'd need to fiddle with your indexes every couple of days. I'm not sure we'd be using them as much.
There is interesting information about the inference pipeline, but almost too late to the party (by at least a year), and for which audience? Techies understand in broad strokes the tech if they are interested, normies will definitely not read it.
All that to say, that yes, it's worth having a laugh. If for nothing else, as a release valve for all the problems they create in the real non-VC world.
Anthropic is IPOing in October according to news, you might be interested in investing.
But you might be right, engineering around the difficult LLM primitive might be a task which is just too hard for your typical software engineer, as you said, they want predictability, hand holding, determinism, most are unable to deal with the real world which is not a spherical cow in a vacuum. So I guess they can stick to simple very well understood primitives like EC2 or Postgres and leave dealing with LLMs for others.
When AI firms ate more than half of global VC private investment in 2025 https://www.oecd.org/en/about/news/announcements/2026/02/ai-... I would expect better results than what we have today.
The most well paid people in the industry brought us here. And "here" is very much as fuzzy as last year with better harnessing towards the local optima. And I say local optima because even the perceived capabilities have slowed down, nevermind the benchmark numbers which are in aggrement.
The best paid engineers in the world, with almost no practical budget limit, still deliver shoddy quality software with AI. Is that not fact? And if it is what does that say for the rest of us.
You are allowed to believe. I'm still waiting for the beneficial results, not only those that benefit griefters, hackers and scammers. AI has been a huge boon there.
Reality will materialize and markets will redress hopefully once they go public. Which they very much seem to be hesitant to do right now.
Not sure what your baseline was, if you said 10 years ago "in 2026 you'll be able to describe an app into the microphone, and the computer will write by itself in one day 50k lines of code to implement it, in a language and tech stack of your choosing, costing $200, and it will sort-of-work, and it will be at least as good as a junior-level programmer writing it from the same requirements in 3 months", most people would have said "implausible, that's at least 50 years away"
For more than a year now I was renting a limited GPU server for ~300$/month to learn, experiment, research and build internal tooling around open weight models. Thinking they are tools with potential and buying the exaggerated marketing are different things.
My history of comments on HN lands often on both providing what I believe to be my insights working with LLMs and calling out exaggerations, stupid terms of service, and the other mishaps in the field. You are free to browse them if you'd like to see my broader opinion.
Also bro: Run /clearbetween tasks. This prevents prior irrelevant context from being sent back to the model, which can reduce token usage. Set your model and effort level before you start. Changing either one mid-conversation can bust your prompt cache, which can increase token cost. @-mention files instead of naming them. The file gets attached to your message directly, which saves a Read call, or a search if Claude has to go find it. Add quiet flags to noisy commands, or run them in a subagent. Command output is added to the conversation just like a file, and stays there for the rest of the session. Run /context once in a fresh session. It shows what's loaded (CLAUDE.md, MCP tool definitions), so you can cut out anything unnecessary. /compact before you take a break from your keyboard. The prompt cache expires after an hour, and summarizing a conversation is much cheaper while it's still cached.
I remember when the internet was an exchange of ideas instead of using gender to justify value of bad ideas
don’t do that, it is weird, use “bruh” or “dude”
Yes. https://magazine.sebastianraschka.com/p/controlling-reasonin...
- https://github.com/anthropics/claude-code/issues/47756 > [BUG] /clear bleeds into the next session (what also breaks cache)
- https://github.com/anthropics/claude-code/issues/47098 > [BUG] new sessions will *never* hit a (full)cache
If you type /resume right after clear, the first thing in the list is the session you just cleared.
* In theory the system prompt is always the same and should therefore be cached, but in practice there's some dynamic strings in there so it doesn't work that way. (Unless they changed this recently.)
With Qwen 3.8 27B, we're one step closer to on-device LLMs that can replace subscriptions.
I know we supposed to do this but is there any particular reason why such things cannot be supported? I thought its running on same model just different settings like reasoning. This would be super useful.
if youre working on the same codebase, that cache stays quite relevant, and i dont think they make the case that clearing and reading the same couple files over and over again is cheaper that relying on it already being cached. same with doing some of the same teaching claude the right way to approach changes in that codebase again and again.
what would be nice is pulling back and reusing an earlier part of the cache for the later two tasks, but claude code doesnt make that particularly easy, and using an LLM to pick where to go back to isnt really gonna save much when it reads all the same text again.
that said: you will surprised by how well current models can navigate (way better than last year!)
> 5-minute cache write tokens are 1.25 times the base input tokens price > 1-hour cache write tokens are 2 times the base input tokens price
https://code.claude.com/docs/en/prompt-caching#on-a-claude-s...
https://platform.claude.com/docs/en/build-with-claude/prompt...
I sometimes just leave some goals or something running before I go to bed or out and I don’t want to pay the cache text when I come back.
I guess if I had to ask something (as someone who doesn't use CC as their daily driver), how much control do you have on subagents and roughly how do define or know when a session is getting too long? I know the answer is "when the model is getting worse" but worse is doing a lot of lifting in that sentence.
If you have to explain that someone is "holding it wrong"(1), that is product error, not a user error.
(1) https://www.wired.com/2010/06/iphone-4-holding-it-wrong/
Especially the /compact part. Like, if a session has been idle for 55 min, why not just automatically run compact at that point?
• https://news.ycombinator.com/item?id=49080605 (JetBrains, Does Speaking to Agents Like Cavemen Save 65% of Tokens? We Test)
• https://news.ycombinator.com/item?id=48588755 (The Token Compression Illusion: Why I'm Skeptical of RTK )
https://regular-reviews.pages.dev/rtk
https://github.com/taobojlen/anthropic-rss-feed/tree/main
I mean, they told us "just talk naturally to the AI because it's so much smarter than all you meatbags" and now it's “for best results, please learn to manage context windows, prompt caching, cache invalidation, model switching, output verbosity and when to manually clear or compact your session.”
I get it, but it seems like the "PRODUCT" should be doing this shit. I.e., the PRODUCT is getting less efficient because I didn't manually manage its context correctly and now it's MY fault.
Edit: i.e., for e.g. Doh. Even the robots get that right. Sigh.
Anthropic has no incentive to make their products more efficient as long as they're selling them by the token.
Unless I am fully not understanding your comment and you don't actually mean “humans require communication skills too” which in honesty feels orthogonal to my complaint.
It's true both that it can be smarter than all us meat bags and that talking to it a certain way gets better results. On some of the things it's a limitation of the technology and on some of the others it's just how information and effort work in any context. I don't see it as orthogonal to your complaint, I see your complaint as misplaced frustration, like Anthropic invented GIGO and compute so they'd have an excuse to write a blog post.
Oh well.
If only they had some kind of technology that could make a judgement and automate those actions...