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

Claude Sonnet 5

anthropic.com

Post reply on HN
Print view View original

311–320 of 822 posts

Re: Claude Sonnet 5

#311

Post by mark_l_watson »

microtonal wrote: post #31

Claude Sonnet 5 is built to be the most agentic Sonnet model yet. It can make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models. I have been using Sonnet 4.6 more than Opus, because I'm mostly doing agent-assisted development and not fully agent-driven development. This announcement does not make me positive, I have fou…

Good point, I also like to do the work myself, with an assistant under my control. I am usually really happy with DeepSeek v4 Flash that I feel just mostly does what I tell it to do, but I do switch to Pro for harder tasks.

There are so many models, and I personally ignore benchmarks so it takes some time to try different models on my use cases. Fortunately, it is ‘good enough’ to do the work to find a few models that work for me, and just use them for a month or two before re-investing time for my own evals to possibly change models.

People should evaluate what works for them and ignore other people and benchmarks. (Apologies if that sounds snarky.)

Re: Claude Sonnet 5

#312

Post by quaverquaver »

ricardobayes wrote: post #299
everforward wrote: post #159

Earlier quoted context omitted.

There’s no way to justify their valuations if they get downgraded to a pair programming tool. They need fully agentic stuff to work and replace human engineers to even come close. Offhand, I’m not even certain whether a model like that could justify the constant retraining we’re doing on the agentic models. It doesn’t make a lot of sense to spend millions or billions on training to reduce hallucinations by 0.3% if yo…

At some point it's going to plateau, maybe already has. Then they will switch to FPGA/ASIC-based model-specific hardware for lower consumption. I'm pretty sure the "space data centers" won't use GPUs, they are not radiation-tolerant whereas FPGAs can be. https://www.cerebras.ai/blog/gemma-4-on-cerebras-the-fastest...

I would not take "space data centers" as a given! from most to least likely these will be vaporware, vaprorized-ware, rubble-ware, loss leaders.

Re: Claude Sonnet 5

#313

Post by bloppe »

ricardobayes wrote: post #308
keeda wrote: post #253

Earlier quoted context omitted.

Some napkin math -- total global labor compensation is about 50% of the GDP, which puts it in the USD 50 - 60 Trillion range: https://ourworldindata.org/grapher/labor-share-of-gdp This source claims that knowledge workers alone (probably because they are paid much more) account for 35 - 50 Trillion of that: https://github.com/danielmiessler/Substrate/blob/main/Data/K... If LLMs can boost their productivity even by an…

I am deeply surprised by the silence of philosophers, sociologists, liberal arts majors, economists. Where are the think tanks who contemplate and debate the societal aspects? The tech is advancing full steam but the "other side" doesn't feel anywhere nearly ready.

Idk why you're perceiving silence. Feels to me like this is the main thing people talk about nowadays.

Re: Claude Sonnet 5

#314

Post by quaverquaver »

popalchemist wrote: post #135
Brendinooo wrote: post #129

Earlier quoted context omitted.

Yeah, there's a real opportunity for one of these companies to invest time in a model that's tuned for, to use your term, agent-assisted developement. Trouble is, everyone inside their buildings seems to believe that no one will be working like that in a year or two.

Whether they believe it or not is immaterial. It is the end-goal they want to achieve, because then they own the means of production entirely.

these are capital intensive commodity businesses. They can be plenty big - see railroads or airplanes... or refining... but that doesn't mean that most value won't be added elsewhere.

Re: Claude Sonnet 5

#315

Post by jbvlkt »

AquinasCoder wrote: post #138
doctoboggan wrote: post #29

The cost per task chart is telling me that I should _never_ use Sonnet 5 above medium effort level - Opus always performs better for a given cost. So I guess the takeaway is that if Sonnet 5 medium isn't good enough for you, switch models, not effort levels.

While I appreciate, they publish this information, it's increasingly hard to keep track of it all. I've lost the mental model of how different models at different effort levels perform and what tasks they are good at. In practice, I tend to just use the default on Claude Code that works well enough. But I wonder to what degree other users really play around with these settings to optimize for their project.

Exactly this is my problem with all AI tools. I want someone else to create working tools for me so I can focus on my product. It is the same with other tools. I do not want to spent huge amounts of energy and time to setup my IDE, operating system or desk layout. I guess it is too early to have that now.

Re: Claude Sonnet 5

#316

Post by brokencode »

827a wrote: post #295

Tbh we'll see what using it looks like, but the reasoning/cost charts do not look promising. It seems like the only useful reasoning level for Sonnet 5 is Low; medium might trade blows at price/performance with Opus, but anything beyond that Opus is Just Better. I struggle to understand where this model fits in. If I need a cheap model for simple stuff (like, summarizing an email); I'd go Haiku (actually, I'd go Deep…

Kind of crazy how bad this release actually is. I even dug around in the full system card, and every graph showed the same thing.

Low and maybe medium will save money on simpler tasks, but after that it just isn’t worth it compared to Opus.

I wish they would have explained in the blog post why they think anybody would ever want to use this above medium.

Maybe it works well on things that aren’t clear in the benchmarks.

Re: Claude Sonnet 5

#317

Post by user43928 »

overgard wrote: post #193
everforward wrote: post #159

Earlier quoted context omitted.

There’s no way to justify their valuations if they get downgraded to a pair programming tool. They need fully agentic stuff to work and replace human engineers to even come close. Offhand, I’m not even certain whether a model like that could justify the constant retraining we’re doing on the agentic models. It doesn’t make a lot of sense to spend millions or billions on training to reduce hallucinations by 0.3% if yo…

That's a really good point. I think if there wasn't the insane amount of money involved and these were treated as tools instead, they would probably be MORE productive. I think a person working hand in hand with an AI instead of delegating is the sweet spot of making things fast while also not losing understanding or control of the system. You are absolutely right that these companies can't justify their valuations i…

I am thinking the opposite. I've been having great results with handing more and more responsibilities to the agent.

Contrary to what some people suggest, I have not hit any maintenance or reliability dead ends. If something breaks, the agent fixes it.

If it cannot, I have the agent instrument the code and work through the logs to check hypotheses, until the source of the issue is found.

If even that would fail, which did not yet happen, I can still do some old fashioned digging and learning, like I always have.

This is for native mobile app development, and the code base is around 100k LOC.

Re: Claude Sonnet 5

#318

Post by bazhand »

vtail wrote: post #257
bazhand wrote: post #229

Earlier quoted context omitted.

Have you tried '/model opusplan' I've had strong results mixing opus for planning with sonnet implementing.

Fascinating! How did you learn about this?

It was something that was used for token efficiency. Most of the settings and use cases are quite poorly communicated but asking Claude to review the latest release changelog (https://github.com/anthropics/claude-code/blob/main/CHANGELO...) is quite useful. Combined with @"claude-code-guide (agent)" to read it's own docs for settings/configs is super helpful.

The quite useful tool is to use /opusplan along with /codex:rescue (https://github.com/openai/codex-plugin-cc) means you get quite a strongly reviewed plan using native claude + codex without having to implement the mostly useless trust-me-bro plugins and other bs.

Re: Claude Sonnet 5

#319

Post by adam_arthur »

theLiminator wrote: post #28

Seems like the way to go for any smaller models is to only use the low reasoning levels, and for anything where you'd want it to reason harder, to just use a larger model. In effect, high reasoning only makes sense when you're using the frontier model and need extra performance (higher levels of reasoning are never pareto optimal unless you're at the largest model size).

I've found disabling reasoning entirely but adding a "reason" to the JSON response from the LLM to work significantly faster and consume many fewer tokens for narrowly scoped prompts.

At least for Claude family models.

e.g. {

  "reason": "",

  "selection": ""
}

I'm sure native reasoning produces more accurate results, but for my use case the quality was about the same, and the model would reason for thousands of tokens in native reasoning vs just 1-200 with response level reasoning.

Again, to be clear, this is for deterministic/pipeline style workflows, not agentic/coding use.

Re: Claude Sonnet 5

#320

Post by parineum »

keeda wrote: post #253
everforward wrote: post #159

Earlier quoted context omitted.

There’s no way to justify their valuations if they get downgraded to a pair programming tool. They need fully agentic stuff to work and replace human engineers to even come close. Offhand, I’m not even certain whether a model like that could justify the constant retraining we’re doing on the agentic models. It doesn’t make a lot of sense to spend millions or billions on training to reduce hallucinations by 0.3% if yo…

Some napkin math -- total global labor compensation is about 50% of the GDP, which puts it in the USD 50 - 60 Trillion range: https://ourworldindata.org/grapher/labor-share-of-gdp This source claims that knowledge workers alone (probably because they are paid much more) account for 35 - 50 Trillion of that: https://github.com/danielmiessler/Substrate/blob/main/Data/K... If LLMs can boost their productivity even by an…

> If LLMs can boost their productivity even by an average of 5% (studies from ~2024 put it in the ~30% range depending on task) that is ~1.5 - 2.5T in value

Minus the cost of inference, that might not be the boon you're making it out to be. I hear what people around here are spending on their api and I'm skeptical that these tools are making me that much more productive.

Personally, for assisted development, I haven't seen much progress in a while.

Post reply on HN