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

Claude Opus 4.6

anthropic.com

Post reply on HN
Print view View original

581–590 of 1001 posts

Re: Claude Opus 4.6

#581

Post by simonw »

zahlman wrote: post #447
simonw wrote: post #40

The bicycle frame is a bit wonky but the pelican itself is great: https://gist.github.com/simonw/a6806ce41b4c721e240a4548ecdbe...

Do you find that word choices like "generate" (as opposed to "create", "author", "write" etc.) influence the model's success? Also, is it bad that I almost immediately noticed that both of the pelican's legs are on the same side of the bicycle, but I had to look up an image on Wikipedia to confirm that they shouldn't have long necks? Also, have you tried iterating prompts on this test to see if you can get more reali…

I've stuck with "Generate an SVG of a pelican riding a bicycle" because it's the same prompt I've been using for over a year now and I want results that are sort-of comparable to each other.

I think when I first tried this I iterated a few times to get to something that reliably output SVG, but honestly I didn't keep the notes I should ahve.

Re: Claude Opus 4.6

#582

Post by nomel »

libraryofbabel wrote: post #385
nomel wrote: post #278

Earlier quoted context omitted.

Has a "N million context window" spec ever been meaningful? Very old, very terrible, models "supported" 1M context window, but would lose track after two small paragraphs of context into a conversation (looking at you early Gemini).

Umm, Sonnet 4.5 has a 1m context window option if you are using it through the api, and it works pretty well. I tend not to reach for it much these days because I prefer Opus 4.5 so much that I don't mind the added pain of clearing context, but it's perfectly usable. I'm very excited I'll get this from Opus now too.

If you're getting on along with 4.5, then that suggests you didn't actually need the large context window, for your use. If that's true, what's the clear tell that it's working well? Am I misunderstanding?

Did they solve the "lost in the middle" problem? Proof will be in the pudding, I suppose. But that number alone isn't all that meaningful for many (most?) practical uses. Claude 4.5 often starts reverting bug fixes ~50k tokens back, which isn't a context window length problem.

Things fall apart much sooner than the context window length for all of my use cases (which are more reasoning related). What is a good use case? Do those use cases require strong verification to combat the "lost in the middle" problems?

Re: Claude Opus 4.6

#583

Post by jmalicki »

slopusila wrote: post #547
jmalicki wrote: post #414

Earlier quoted context omitted.

I suspect they're marginally profitable on API cost plans. But the max 20x usage plans I am more skeptical of. When we're getting used to $200 or $400 costs per developer to do aggressive AI-assisted coding, what happens when those costs go up 20x? what is now $5k/yr to keep a Codex and a Claude super busy and do efficient engineering suddenly becomes $100k/yr... will the costs come down before then? Is the current "…

after the models get good enough to replace coders they will be able to start increasing the subscriptions back up

At $100k/yr the joke that AI means "actual Indians" starts to make a lot more sense... it is cheaper than the typical US SWE, but more than a lot of global SWEs.

Re: Claude Opus 4.6

#584

Post by eek2121 »

djhn wrote: post #569
ck_one wrote: post #552

Earlier quoted context omitted.

Do you remember how to get around those tricks?

This is the paper: https://arxiv.org/abs/2601.02671 Grok and Deepmind IIRC didn’t require tricks.

This really makes me want to try something similar with content from my own website.

I shut it down a while ago because the number of bots overtake traffic. The site had quite a bit of human traffic (enough to bring in a few hundred bucks a month in ad revenue, and a few hundred more in subscription revenue), however, the AI scrapers really started ramping up and the only way I could realistically continue would be to pay a lot more for hosting/infrastructure.

I had put a ton of time into building out content...thousands of hours, only to have scrapers ignore robots, bypass cloudflare (they didn't have any AI products at the time), and overwhelm my measly infrastructure.

Even now, with the domain pointed at NOTHING, it gets almost 100,000 hits a month. There is NO SERVER on the other end. It is a dead link. The stats come from Cloudflare, where the domain name is hosted.

I'm curious if there are any lawyers who'd be willing to take someone like me on contingency for a large copyright lawsuit.

Re: Claude Opus 4.6

#586

Post by simonw »

nubg wrote: post #75
simonw wrote: post #50

Earlier quoted context omitted.

The cost per token served has been falling steadily over the past few years across basically all of the providers. OpenAI dropped the price they charged for o3 to 1/5th of what it was in June last year thanks to "engineers optimizing inferencing", and plenty of other providers have found cost savings too. Turns out there was a lot of low-hanging fruit in terms of inference optimization that hadn't been plucked yet. >…

> "engineers optimizing inferencing" are we sure this is not a fancy way of saying quantization?

The o3 optimizations were not quantization, they confirmed this at the time.

Re: Claude Opus 4.6

#587

Post by soulofmischief »

ck_one wrote: post #562
golfer wrote: post #560

Earlier quoted context omitted.

There's lots of websites that list the spells. It's well documented. Could Claude simply be regurgitating knowledge from the web? Example: https://harrypotter.fandom.com/wiki/List_of_spells

It didn't use web search. But for sure it has some internal knowledge already. It's not a perfect needle in the hay stack problem but gemini flash was much worse when I tested it last time.

The only worthwhile version of this test involves previously unseen data that could not have been in the training set. Otherwise the results could be inaccurate to the point of harmful.

Re: Claude Opus 4.6

#588

Post by everforward »

jama211 wrote: post #68
gjsman-1000 wrote: post #20

Earlier quoted context omitted.

Also explains why Claude Code is a React app outputting to a Terminal. (Seriously.)

There’s nothing wrong with that, except it lets ai skeptics feel superior

There are absolutely things wrong with that, because React was designed to solve problems that don't exist in a TUI.

React fixes issues with the DOM being too slow to fully re-render the entire webpage every time a piece of state changes. That doesn't apply in a TUI, you can re-render TUIs faster than the monitor can refresh. There's no need to selectively re-render parts of the UI, you can just re-render the entire thing every time something changes without even stressing out the CPU.

It brings in a bunch of complexity that doesn't solve any real issues beyond the devs being more familiar with React than a TUI library.

Re: Claude Opus 4.6

#589

Post by simonw »

KaiserPro wrote: post #515
grim_io wrote: post #494

Earlier quoted context omitted.

What do you mean it's on ollama and requires h100? As a proprietary google model, it runs on their own hardware, not nvidia.

sorry A lack of context: https://ollama.com/library/gemini-3-pro-preview You can run it on your own infra. Anthropic and openAI are running off nvidia, so are meta(well supposedly they had custom silicon, I'm not sure if its capable of running big models) and mistral. however if google really are running their own inference hardware, then that means the cost is different (developing silicon is not cheap...) as you sa…

You can't run Gemini 3 Pro Preview on your own infrastructure. Ollama sell access to cloud models these days. It's a little weird and confusing.

Re: Claude Opus 4.6

#590

Post by skydhash »

HarHarVeryFunny wrote: post #545
skydhash wrote: post #344

Earlier quoted context omitted.

The terminal UI is not a tree structure that you can diff. It’s a 2D cells of characters, where every manipulation is a stream of texts. Refreshing or diffing that makes no sense.

IMO diffing might have made sense to do here, but that's not what they chose to do. What's apparently happening is that React tells Ink to update (re-render) the UI "scene graph", and Ink then generates a new full-screen image of how the terminal should look, then passes this screen image to another library, log-update, to draw to the terminal. log-update draws these screen images by a flicker-inducing clear-then-red…

Any diff would require to have a Before and an After. Whatever was done for the After can be done to directly render the changes. No need for the additional compute of a diff.
Post reply on HN