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

I've been using Claude Code for a couple of days

twitter.com

Post reply on HN
Print view View original

481–490 of 507 posts

Re: I've been using Claude Code for a couple of days

#481

Post by pramodbiligiri »

jonwinstanley wrote: post #196

Do any of the current AI systems allow you to use voice? I’d love to sometimes chat to an agent and dictate what I want to happen. For one project there’s a lot of boilerplate and I imagine AI could be really fast for tasks like: “create a new controller called x”, “make a new migration to add a new field the to users table called x” etc

VS Code Insiders has an Agent mode now that can also take voice input. You should be able to see the microphone icon to the left of "Agent" in the demo video in their blog post: https://code.visualstudio.com/blogs/2025/02/24/introducing-c...

Re: I've been using Claude Code for a couple of days

#482

Post by fragmede »

bufferoverflow wrote: post #366
CGamesPlay wrote: post #255

I decided to check this out after seeing the discussion here. I had previously misunderstood that it required a Claude.ai plan, but it actually just uses your API keys. I did a comparison between Claude Code and Aider (my normal go-to): I asked it to do clone a minor feature in my existing app with some minor modifications (specifically, a new global keyboard shortcut in a Swift app). Claude Code spent about 60 secon…

Now, the billion dollar question is - how long would it take you to code that diff?

time spent is not the only question. how much thought it takes, however impossible that may be to measure, is another one. If an LLM assisted programmer is able to solve the problem without deep focus, while responding to emails and attending meetings, vs the programmer who can't, is time really the only metric we can have here?

Re: I've been using Claude Code for a couple of days

#483

Post by curiouser3 »

tipsytoad wrote: post #470

I usually am a huge fan of “copilot” tools (I use cursor, etc) and Claude has always been my go to. But Sonnet 3.7 actually seems dangerous to me, it seems it’s been RL’d _way_ too hard into producing code that won’t crash — to the point where it will go completely against the instructions to sneak in workarounds (e.g. returning random data when a function fails!). Claude Code just makes this even worse by giving ver…

this is a huge issue for me as well. It just kind of obfuscates errors and masks the original intent, rather than diagnosing and fixing the issue. 3.5 seemed to be more clear about what it's doing and when things broke at least it didn't seem to be trying to hide anything.

Re: I've been using Claude Code for a couple of days

#484

Post by TeMPOraL »

lanstin wrote: post #454
naasking wrote: post #450

Earlier quoted context omitted.

> For exemple I can’t remember seing a PR with an API that seems plausible but never ever existed A PR is code that has already been tested and refined, which is not comparable to the output of an LLM. The output of an LLM is comparable to the first, untested code that you wrote based off of your sometimes vague memory of how some API works. It's not at all uncommon to forget some details of how an API works, what ca…

It’s kind of uncommon to be aware that you have only a vague recall of the API and not go check the documentation or code to refresh your memory. That self knowledge that you knew something and aren’t sure of the details is indeed the thing that these tools lack. So far.

Human programmers have continuous assistance on every keystroke - autocomplete, syntax highlighting, and ultimately, also the compilation/build step itself.

For an LLM-equivalent experience, go open notepad.exe and make substantial changes there, and then rebuild - and let the compiler tell you what's your base rate of hallucinating function names and such.

Re: I've been using Claude Code for a couple of days

#485

Post by JohnKemeny »

kittikitti wrote: post #419
JohnKemeny wrote: post #397

Earlier quoted context omitted.

Exception: Be very careful not to confuse "deferring to an authority on the issue" with the appeal to authority fallacy . Remember, a fallacy is an error in reasoning . Dismissing the council of legitimate experts and authorities turns good skepticism into denialism. The appeal to authority is a fallacy in argumentation, but deferring to an authority is a reliable heuristic that we all use virtually every day on issu…

Where are you going with this? The authority provided no evidence so noting the logical fallacy is correct here. In this scenario, people are appealing to authority because there's nothing else to count on.

The tweet is a domain expert talking about his experiences using a product.

If (hyperbole) Picasso wrote a tweet saying certain canvases work great for certain tasks, and some random internet dude counters with A single tweet with lots of analogy, with no images/painting examples whatsoever. These are just words. Are we just discussing art based on vibe? ...

I will read that comment in the comic book guy voice.

Re: I've been using Claude Code for a couple of days

#487

Post by lanstin »

TeMPOraL wrote: post #484
lanstin wrote: post #454

Earlier quoted context omitted.

It’s kind of uncommon to be aware that you have only a vague recall of the API and not go check the documentation or code to refresh your memory. That self knowledge that you knew something and aren’t sure of the details is indeed the thing that these tools lack. So far.

Human programmers have continuous assistance on every keystroke - autocomplete, syntax highlighting, and ultimately, also the compilation/build step itself. For an LLM-equivalent experience, go open notepad.exe and make substantial changes there, and then rebuild - and let the compiler tell you what's your base rate of hallucinating function names and such.

In the 1990s, that is closer to what making software was like. There, one had an even more heightened awareness of how confident one was in what one was typing. We would then go to the manual (physical in many cases) and look it up.

And we never made up APIs, as there just weren't that many APIs. We would open the .h file for the API we were targeting as we typed into the other window. And the LLMs have ingested all the documentation and .h files (or the modern equivalent) so they don't have a real excuse.

But I use the LLMs all the time for math, and they do profusely hallucinate in a way people do not. I think it's a bit disingenuous to say that LLMs don't have that failure mode that people don't really have.

Re: I've been using Claude Code for a couple of days

#488

Post by rglover »

baq wrote: post #386
leptons wrote: post #382

Earlier quoted context omitted.

>Code is not an asset, it’s a liability. Then AI generated code is a time bomb. Maintainability? Who needs that if you can sell the company quickly enough, I guess.

Always has been! Not sure what AI changes here.

It encourages developers to not worry about understanding the code they write. At least now, there's someone around (typically) who can figure things out.

But as more developers rely on AI and their skills atrophy, if entire products are being built with AI, eventually no one will exist who can fix the software when it breaks. Not only that, but anything that requires knowledge outside the training of the AI will be impossible (or at best, extremely difficult/costly) to implement.

Re: I've been using Claude Code for a couple of days

#489

Post by branko_d »

smallerfish wrote: post #192
alextingle wrote: post #166

Earlier quoted context omitted.

Is the question actually difficult, though? If you ask for some standard task, then of course those who are leaning heavily on LLMs will do well, as that's exactly where they work best. That doesn't tell you anything about the performance of those candidates in situations where the LLM won't help them. I suppose, if you are specifically looking for coders to perform routine tasks, then you'll get what you need. Of co…

No, I'm not doing leetcode or algorithm questions - it's basically "build a [tiny] product to specs", in a series of steps. I'm evaluating candidates on their process, their effectiveness, their communication (I ask for narration), and their attention to detail. I do review code afterwards. And, bear in mind that this is only round 1 - once I talk with the ones who do well, I'll go deep on a number of topics to under…

> I ask for narration

That's a mistake. There are plenty of people who are not good multitaskers and cannot effectively think AND talk at the same time, myself included.

IMHO, for programming, monotaskers tend to do better than multitaskers.

Re: I've been using Claude Code for a couple of days

#490

Post by throw234234234 »

kthxb wrote: post #43

As a junior, this scares me. I don't think I'll be out of a job soon, but certainly the job market will change drastically and the times where SWEs are paid like doctors and lawyers will end?

Across most of the world SWE's aren't paid like this anyway. It really is only the US that has this level of pay for the SWE staff unless you are into a different domain specialty (e.g. trading/finance) in which case you aren't just paid for coding knowledge.

Coding is the real market for LLM's - it is the "killer app". I'm starting to think that a possibility is that for most jobs it ends up being a "what was all the hype about" but for SWE's it will be carnage globally w.r.t jobs. They aren't generally intelligent (lots of hallucinations in other domains), but with RL can be trained on the mountains of open source software out there to displace software.

Software developers automating their own jobs away. The promise of "free software" with open source was realised - just not as people envisioned it.

No other profession would do this or at least not as fast. As a honest answer, as much as I want to be wrong, I don't recommend new comers join the industry unless they really want to do it and are happy taking the risk. The uncertainty is high right now, and the anxiety these tools are giving quite a number of people I talk to is very high. Capitalism doesn't reward value or how much is built - it allocates resources and rewards people who produce into scarce markets. AI makes the product of code significantly less scarce.

Post reply on HN