Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How Anthropic teams use Claude Code

anthropic.com

Post reply on HN
Print view View original

211–220 of 248 posts

Re: How Anthropic teams use Claude Code

#211

Post by danielbln »

mvieira38 wrote: post #187

I'm surprised they let it generate the tests. Aren't the tests precisely what you want to have fine control over, moreso than the code itself? Letting it write them itself feels like yielding control over what is considered "correct" code

Unpopular take: I couldn't give a single damn about the code. If it passes the spec and tests (that I review) and the feature works, that is good enough for me. I'm in the business of automation, and in the business of solving problems. Writing perfect, beautiful code is not what I'm in the business of. Now of course, good enough doesn't mean "complete horse shit spaghetti code", but it also doesn't mean I give a damn if some function is maybe not pure enough or whatever.

Re: How Anthropic teams use Claude Code

#212

Post by amedviediev »

forgotmypw17 wrote: post #138

I’ve implemented and maintained an entire web app with CC, and also used many other tools (and took classes and taught workshops on using AI coding tools). The most effective way I’ve found to use CC so far is this workflow: Have a detailed and also compressed spec in an md file. It can be called anything, because you’re going to reference it explicitly in every prompt. (CC usually forgets about CLAUDE.md ime) Start…

This matches my experience as well. But what I also found is that I hate this workflow so much that I would almost always rather write the code by hand. Writing specs and user stories was always my least favorite task.

Re: How Anthropic teams use Claude Code

#213

Post by apwell23 »

danielbln wrote: post #209
mixdup wrote: post #54

The first example was helping debug k8s issues, which was diagnosed as IP pool exhaustion, and Claude helped them fix it without needing a network expert But, if they had an expert in networking build it in the first place, would they have not avoided the error entirely up front?

Experts make mistakes too. In fact, all humans do.

they don't make dumb mistakes like claude

Re: How Anthropic teams use Claude Code

#215

Post by oblio »

Quinner wrote: post #208
oblio wrote: post #78

Earlier quoted context omitted.

Again, those prices aren't stable. Nobody is investing half a trillion in a tech without expecting a 10x return. And fairly sure soon those $20/month subscriptions will sell your data, shove ads everywhere AND basically only allow you to get that junior dev for 30 minutes per day or 2 days a month. And the $200/month will probably be $500-1000 with more limitations. Still cheap, but AI can't run an entire project, ca…

The wages aren't stable either. There's going to be gradual convergence.

Oh, by the way, this entire discussion revolves around LLMs being an exponential tech. Real life only works with sigmoids.

Re: How Anthropic teams use Claude Code

#216

Post by const_cast »

godelski wrote: post #191
jvanderbot wrote: post #143

Earlier quoted context omitted.

One time I explained that I was afraid of tesla full self driving, because while using it my tesla accelerated to 45mph in a parking lot that was parallel to the road and only separated by a curb. The pushback I got was "Why would you use FSD in a parking lot". Well, "Full", right? Same here. It's either capable of working unsupervised or not. And if not, you have to start wondering what you're even doing if you're a…

> We're still working out the edge cases The difficult part is that like with FSD, it's mostly edge cases

Driving is just mostly edge cases. I've thought about it a lot, but I think automating driving is much harder than automating even air travel.

Sure the air is 3 dimensions, but driving is too dynamic and volatile. Every single road is different, and you have to rely on heuristics meant for humans.

It's stupid easy for humans to tell what is a yellow line and what a stop sign looks like, but it's not so easy for computers. These are human tools - physical things we look at with our eyes. Not easy to measure. Whereas measurements in the air are quite easy to measure.

On top of the visual heuristics, everthing changes all the time and very fast. You look away from the road and look back and you don't know what you're gonna see. It's why texting and driving is so dangerous.

Re: How Anthropic teams use Claude Code

#217

Post by DANmode »

tayo42 wrote: post #158
thunky wrote: post #146

Earlier quoted context omitted.

> Lack of iteration What makes you think that agents can't iterate? > I'm going to throw them out if they are just constantly showering me with praise and telling me how much of a genius I am You can tell the agent to have the persona of an arrogant ass if you prefer it.

Llms only work in one direction, they produce the next token only. It can't go back and edit. They would need to be able to back track and edit in place somehow

"Somehow", like caching multiple layers of context, like all the free tools are now doing?

Re: How Anthropic teams use Claude Code

#218

Post by maronato »

ants_everywhere wrote: post #4

I've been trying Claude Code for a few weeks after using Gemini Cli. There's something a little better the tool use loop, which is nice. But Claude seems a little dumber and is aggressive about "getting things done", often ignoring common sense or explicit instructions or design information. If I tell it to make a test pass, it will sometimes change my database structure to avoid having to debug the test. At least tw…

Claude trying to cheat its way through tests has been my experience as well. Often it’ll delete or skip them and proudly claim all issues have been fixed. This behavior seems to be intrinsic to it since it happens with both Claude Code and Cursor.

Interestingly, it’s the only LLM I’ve seen behave that way. Others simply acknowledge the failure and, after a few hints, eventually get everything working.

Claude just hopes I won’t notice its tricks. It makes me wonder what else it might try to hide when misalignment has more serious consequences.

Re: How Anthropic teams use Claude Code

#219

Post by nightshift1 »

TrainedMonkey wrote: post #87
minimaxir wrote: post #7

A repeated trend is that Claude Code only gets 70-80% of the way, which is fine and something I wish was emphasized more by people pushing agents. This bullet point is funny: > Treat it like a slot machine > Save your state before letting Claude work, let it run for 30 minutes, then either accept the result or start fresh rather than trying to wrestle with corrections. Starting over often has a higher success rate th…

$200 per month will get you roughly 4-5 hours of non-stop single-threaded usage per day. A bigger issue here is that the random process is not a good engineering pattern. It's not repeatable, does not drive coherent architecture, and struggles with complex problems. In my experience, problem size correlates inversely with generated code quality. Engineering is a process of divide-and-conquer and there is a good reaso…

I agree with that list. I would also add that you should explicitly ask the llm to read the whole files at least once before starting edits because they often have tunnel vision. The project map is auto generated with a script to avoid reading too many files but the files to be edited should be fresh in the context imo.

Re: How Anthropic teams use Claude Code

#220

Post by threecheese »

mwigdahl wrote: post #69
lovich wrote: post #19

Earlier quoted context omitted.

> I use it at home via the $20/m subscription and am piloting it at work via AWS Bedrock. When used with Bedrock APIs, at the end of every session it shows you the dollar amount spent which is a bit disconcerting. I hope the fine-grained metering of inference is a temporary situation otherwise I think it will have a chilling/discouraging effect on software developers, leading to less experimentation and fewer rewrite…

I’ve seen it firsthand at work, where my developers are shy about spending even a single digit number of dollars on Claude Code, even when it saves them 10 times that much in opportunity cost. It’s got to be some kind of psychological loss aversion effect.

Years of being hammered to reduce reduce reduce in monthly cloud spend certainly makes us cost-averse. In reality you are completely correct wrt the real value, and leadership will always see this (hopefully), but I still don’t want to hear my name spoken in our monthly budget review.
Post reply on HN