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

Claude Opus 4.8

anthropic.com

Post reply on HN
Print view View original

911–920 of 1001 posts

Re: Claude Opus 4.8

#911

Post by eknkc »

dudeinhawaii wrote: post #395

This is the first time I saw a model pop-up on HN and didn't really care. Model exhaustion? It looks interesting but not exciting. While I'd normally _love_ incremental improvements --- I think the recent ones are far too minor to get excited about or change up a workflow. Besides, benchmarks tend to exaggerate the gap between versions. At this point I'd almost rather Anthropic wait and really wow us with a 5.0 relea…

Yeah.. Also, after they fucked up the CC, released that 4.7 etc. I switched completely to Codex and honestly do not wish to go back to half assed harness of Claude. Codex somehow got extremely good in a couple of months.

Re: Claude Opus 4.8

#912

Post by onlyrealcuzzo »

pseudosavant wrote: post #863
onlyrealcuzzo wrote: post #136

Earlier quoted context omitted.

I won't be surprised if the next gen frontier models are the last. There's orders of magnitude of low hanging juice to squeeze out of smaller models. It is almost guaranteed that a 60-90B model can outperform current SOTA in coding tasks within 2-3 years (design not certain, probably unlikely). It is far less clear that a 1.2T model will be meaningfully better enough to justify training it. As far as reasoning is con…

It is fascinating to me to see a new product category that improves so vastly year-after-year, where people commonly state that this is now the peak already. I couldn’t even imagine having to go back to a model from 12 months ago, much less 24 months ago. GPT-5.5 is so much better than GPT-4o that it sure seems like they keep finding new juice to squeeze. This is like going from dialup internet to DSL and acting like…

> I couldn’t even imagine having to go back to a model from 12 months ago, much less 24 months ago. GPT-5.5 is so much better than GPT-4o that it sure seems like they keep finding new juice to squeeze

The difference in progress in smaller models is far more impressive.

Compare Gemini 3.5 Flash to a ~16B parameter model from 24 months ago.

Compare GPT-5.5 to a frontier model 24 months ago.

Yes, GPT-5.5 got better. At orders of magnitude smaller parameter sizes (when factoring in ACTIVE parameters) the increase is far more pronounced.

Re: Claude Opus 4.8

#914

Post by onlyrealcuzzo »

notrealyme123 wrote: post #857
onlyrealcuzzo wrote: post #136

Earlier quoted context omitted.

I won't be surprised if the next gen frontier models are the last. There's orders of magnitude of low hanging juice to squeeze out of smaller models. It is almost guaranteed that a 60-90B model can outperform current SOTA in coding tasks within 2-3 years (design not certain, probably unlikely). It is far less clear that a 1.2T model will be meaningfully better enough to justify training it. As far as reasoning is con…

The GRAM model is so much into my research direction, I love it. Thank you for posting it. Where do I find papers like this? Outside of hacker news comments. It's so hard to find the good stuff in all the noise IMO.

> Where do I find papers like this?

I got it from my Google News recs on my phone, because I've been watching a bunch of videos on YouTube about LeCun's ideas on World Models and JEPA (I think).

Re: Claude Opus 4.8

#915

Post by narrator »

nerevarthelame wrote: post #110
onlyrealcuzzo wrote: post #13

Does anyone troll these releases and cherry pick random metrics other companies would cherry pick to show how amazing their models are? There's like 8 million benchmarks. Every release, every model randomly picks 5-10 where they win in everything except 1, to make it look like they aren't randomly cherry picking benchmarks they probably benchmaxxed for.

It's interesting they only included 6 metrics this time. Opus 4.7 had 12, and 4.6 had 13. Of the metircs they reported for 4.7, for 4.8 they excluded BrowseComp, CharXiv Reasoning, CyberGym, GPQA Diamond, MCP Atlas, MMMLU, SWE-bench Verified. The last 4 were almost always mentioned in previous Opus releases.

Their Cybergym score is reportedly awful because of the cybersecurity nerfing. https://x.com/i/status/2060046843023630841

Re: Claude Opus 4.8

#916

Post by calpaterson »

epitrochoid413 wrote: post #906

Meanwhile Deepseek is cutting inference costs to mere cents. Thats the real AI revolution for you.

Yes I switched from claude code to opencode with deepseek recently.

It is basically indistinguishable from sonnet. At this point my own prompts, AGENTS.md, background docs and so on matter a great deal more than the differences between models.

And deepseek v4 flash (the sonnet comparable) costs 3% of what sonnet does.

Re: Claude Opus 4.8

#917

Post by epitrochoid413 »

calpaterson wrote: post #916
epitrochoid413 wrote: post #906

Meanwhile Deepseek is cutting inference costs to mere cents. Thats the real AI revolution for you.

Yes I switched from claude code to opencode with deepseek recently. It is basically indistinguishable from sonnet. At this point my own prompts, AGENTS.md, background docs and so on matter a great deal more than the differences between models. And deepseek v4 flash (the sonnet comparable) costs 3% of what sonnet does.

Same here. I use codex for planning and deepseek v4 flash for implementation. Its worked really well so far.

Re: Claude Opus 4.8

#918

Post by _heimdall »

northern-lights wrote: post #23

> Not only that, but we plan to release a new class of model with even higher intelligence than Opus. As part of Project Glasswing, a small number of organizations are currently using Claude Mythos Preview for cybersecurity work. Models of this capability level require stronger cyber safeguards before they can be generally released. We’re making swift progress on developing these safeguards and expect to be able to b…

I'm still not sure what safeguards they can be adding here. Unless they've suddenly solved alignment, at best isn't it a collection of system prompts saying what not to do and potentially some screening algorithms that try to catch key phrases in inputs/outputs?

Re: Claude Opus 4.8

#919

Post by Npovview »

arnorhs wrote: post #871
colonCapitalDee wrote: post #32

"Users will find Opus 4.8 to be a modest but tangible improvement on its predecessor." This is a refreshing attitude! I've also verified that you can now turn off adaptive thinking in the web UI, which is great. I've had a lot of problems with thinking not triggering and the model producing sub-par output. Glad we can finally turn it off. (I hope being able to turn off adaptive thinking is new, if I could have turned…

Yes, modest but tangible improvement - same modesty does not apply to the cost: https://artificialanalysis.ai/models/capabilities/coding#cod...

Originally Indians didn't drink tea. British East India company got Indians addicted to tea/chai when they sold it for free. Then the real prices came in.

Re: Claude Opus 4.8

#920

Post by spaceman_2020 »

throwaway63467 wrote: post #440
spaceman_2020 wrote: post #435

Earlier quoted context omitted.

I think 4.7 was an awful model in actual use. I never got anything out of it and it was frustratingly weird. This feels more like an attempt to course correct and isn't a real bump

I think they overtrained on scientific papers or such as it would spout really sophisticated sounding nonsense with a ton of complicated verbs and adjectives. 4.6 was definitely better in that regard. The more I use these tools the more I think they’re not actually that revolutionary. I mean it’s still amazing what they can do but they have very clear limitations it seems.

it was also astonishingly lazy. Would just ask me to write test scripts. I asked it to create simple UI buttons for testing some basic functions so I could share it with a client, and it gave me curl commands instead - and then defended it by saying that the UI is wasted work

Frustrating because if I have a tool, I expect a tool to do what I tell it to do. Tools shouldn't have any opinions on how they should be used

Post reply on HN