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

Claude Sonnet 4.5

anthropic.com

Post reply on HN
Print view View original

421–430 of 819 posts

Re: Claude Sonnet 4.5

#421

Post by ewoodrich »

enraged_camel wrote: post #332
stefs wrote: post #272

Earlier quoted context omitted.

How can I trust it to do the complicated task well when it fails to do the simple thing?

It's not about simple vs. complex. It's about the types of tasks the AI has been trained on: pattern-matching, thinking, reasoning, research. Tasks like linting and formatting a block of code are pretty simple, but also very specialized. You're much better off using formatters/linters than an AI.

An agent being tasked to resolve simple issues from a compiler/test suite/linter/etc is pretty typical use case. It's not clear in this example if the linter was capable of auto fixing the problem, so ordinarily this would be a case where you'd hope an LLM would shine given specific, accurate context and known solution.

Re: Claude Sonnet 4.5

#422

Post by gloosx »

simonw wrote: post #403
gloosx wrote: post #394

Earlier quoted context omitted.

I am curious how the sandbox handles potentially malicious code. For example, what would happen if someone tried to run something like a crypto miner or a DDoS script?

Code running in the Claude environment can't make outbound network connections to anything more than a tiny list of domains (GitHub, PyPI, NPM) so DDoS targets are limited. If I was Anthropic I'd have measures in place to rate limit network traffic from individual containers.

This would be interesting to test in a controlled way, curious what the CPU/time/memory quotas per run, isolation method, and syscall filtering look like.

Would appreciate pointers to any documentation or tests.

Re: Claude Sonnet 4.5

#423

Post by mchusma »

For me, Opus 4.1 was so much better than Sonnet 4.0 that I used it exclusively in Claude Code and cancelled Cursor. I'm a bit skeptical that Sonnet 4.5 will be in practice better, but will test with it and see! Hopefully we get Opus 4.5 soon.

Re: Claude Sonnet 4.5

#424

Post by minimaxir »

landl0rd wrote: post #419
nchmy wrote: post #341

Earlier quoted context omitted.

are you aware of any "ai journalists"? Because simonw does great work, so perhaps blogger is what people should aspire towards?

Kylie Robison recently moved to Wired and is a solid "AI journalist".

Although she is indeed solid as an AI journalist, unfortunately she was recently let go for unknown reasons: https://www.kyliebytes.com/thank-god-i-got-fired/

Re: Claude Sonnet 4.5

#425

Post by mvdtnz »

kurtis_reed wrote: post #250
simonw wrote: post #204

I had access to a preview over the weekend, I published some notes here: https://simonwillison.net/2025/Sep/29/claude-sonnet-4-5/ It's very good - I think probably a tiny bit better than GPT-5-Codex, based on vibes more than a comprehensive comparison (there are plenty of benchmarks out there that attempt to be more methodical than vibes). It particularly shines when you try it on https://claude.ai/ using its brand n…

Why did you have access to a preview?

[flagged]

Re: Claude Sonnet 4.5

#426

Post by ojosilva »

To @simonw and all the coding agent and LLM benchmarkers out there: please, always publish the elapsed time for the task to complete successfully! I know this was just a "it works straight in claude.ai" post, but still, nowhere in the transcript there's a timestamp of any kind. Durations seem to be COMPLETELY missing from the LLM coding leaderboards everywhere [1] [2] [3]

There's a huge difference in time-to-completion from model to model, platform to platform, and if, like me, you are into trial-and-error, rebooting the session over and over to get the prompt right or "one-shot", it's important how reasoning efforts, provider's tokens/s, coding agent tooling efficiency, costs and overall model intelligence play together to get the task done. Same thing applies to the coding agent, when applicable.

Grok Code Fast and Cerebras Code (qwen) are 2 examples of how models can be very competitive without being the top-notch intelligence. Running inference at 10x speed really allows for a leaner experience in AI-assisted coding and more task completion per day than a sluggish, but more correct AI. Darn, I feel like a corporate butt-head right now.

1. https://www.swebench.com/

2. https://www.tbench.ai/leaderboard

3. https://gosuevals.com/agents.html

Re: Claude Sonnet 4.5

#427

Post by bigwheels »

Aflynn50 wrote: post #374

When I see how much the latest models are capable of it makes me feel depressed. As well as potentially ruining my career in the next few years, its turning all the minutiae and specifics of writing clean code, that I've worked hard to learn over the past years, into irrelivent details. All the specifics I thought were so important are just implementation details of the prompt. Maybe I've got a fairly backwards view…

I've been deep into AI full-time professionally for some months now, and for the first 4+ weeks I felt the exact same way as you describe - it is a form of existential crisis, especially after spending the bulk of the past 25 years honing my coding-fu algo ninja skills, my identity was totally wrapped up in it.

Keep at it and keep leaning in to embrace it, I promise it gets better! It's just a big adjustment.

You are FAR more than your code!

Re: Claude Sonnet 4.5

#428

Post by artursapek »

iagooar wrote: post #258

Anecdotal evidence. I have a fairly large web application with ~200k LoC. Gave the same prompt to Sonnet 4.5 (Claude Code) and GPT-5-Codex (Codex CLI). "implement a fuzzy search for conversations and reports either when selecting "Go to Conversation" or "Go to Report" and typing the title or when the user types in the title in the main input field, and none of the standard elements match, a search starts with a 2s de…

I think Codex working for 20 mins uninterrupted is actually a strength. It’s not “slow” as critics sometimes say - it’s thorough and autonomous. I can actually walk away and get something else done around the house while it does my work for me.

Re: Claude Sonnet 4.5

#429

Post by epolanski »

andrewstuart wrote: post #281
trevin wrote: post #69

I’m always fascinated by the fine-tuning of LLM personalities. Might we finally get less of the reflexive “You’re absolutely right” with this one? Maybe we’re entering the Emo Claude era. Per the system card: In 250k real conversations, Claude Sonnet 4.5 expressed happiness about half as often as Claude 4, though distress remained steady.

I like being lightly flattered.

I don't, I need someone telling me the flaws of my ideas, not to confirm them for the sake of it.

Re: Claude Sonnet 4.5

#430

Post by Workaccount2 »

epolanski wrote: post #416
Aflynn50 wrote: post #374

When I see how much the latest models are capable of it makes me feel depressed. As well as potentially ruining my career in the next few years, its turning all the minutiae and specifics of writing clean code, that I've worked hard to learn over the past years, into irrelivent details. All the specifics I thought were so important are just implementation details of the prompt. Maybe I've got a fairly backwards view…

That minutiae was always borderline irrelevant, the skill was always making somebody money, possibly with software. The reality is that more software will be pushed than before, and more of it will need to be overseen by a professional.

The real question is what kind of pay that work will demand.

It's will be great to still be employed as a senior dev. It will be a little less great with a $110k salary, 5 day commute, and mediocre benefits being the norm.

Post reply on HN