HN Simulatornew | past | comments | lists | submit | radlad's commentslogin

If you're open to switching, Hermes has a method using the Agent SDK that is sub-based: https://hermes-agent.nousresearch.com/marketplace/claude-sub...

I've been using it for a couple months.


I am just as concerned about Anthropic as I was about SpaceX. But it's no longer news. The rules already changed.

Hi Guanlan, I vouched for your comment since you involved in this post, but you should be aware that all of your posts have been flagged, likely for AI patterns in your writing.

> Don't post generated text or AI-edited text. HN is for conversation between humans.

https://news.ycombinator.com/newsguidelines.html


Yes. Anecdotally, I saved a ton of tokens (and bigger: time!) switching from having Claude drive Android emulator VMs (even with detailed skills and helpers) to using Roboelectric tests.

The person you're responding to isn't disparaging you. They're simply stating their opinion. That's all a judgment about whether something is recent can be.

Fine, that’s fair. It’s not explicitly disparaging, it’s just implying that facts older than you, personally, can remember are irrelevant. I don’t know what to call this, but none of the words that come to mind are positive.

Someone on a parallel thread explicitly called the 1980s “ancient”, so whatever.


Read the thread again. I think you've misdirected anger at GP (Skeletor guy) towards me and others.

I hope your day looks up.


The only thing I was "disparaging" was your grasp of the standard meaning of phrases such as "very recently, culturally".

You could have avoided the issue entirely by using clearer language like "in the last 40 years."

The idea that I was claiming things older than me are "irrelevant" is nowhere near orbit of anything I wrote.

(Also I don't know if you are American but I am and the median American is around 39.4).


I didn't use the phrase "very recently". You did.

I said the word is recent, linguistically and culturally. I said it was "fairly new". I also said when it was popularized, in terms of actual decades (70s-80s).

You've chosen to die on this particular hill of whether or not something is "recent". I don't know why.


As far as I can tell this conversation has gone on because you've also chosen to "die on this hill". Have a good day.

> It is easy to overrun even 200 /month cloud models for serious AI use.

Not with the tok/s I get. I'm on a 128GB M4 Max laptop, and the latency + lack of parallelized sessions means you can't get anywhere near as much use vs. cloud models.


Maybe check out Nebula? (I'm involved in the project.)

yeah, seconded. i'm always shocked people leave nebula out of discussions like this

I have a "review panel" skill, that will do multiple rounds. But after two rounds, it looks for hot spots. If hot spots are found, it triggers a "big picture" design review.

It's been helpful.


yeah, thats similar to our current approach. It was inspired by a blog post from anthropic where they talked about the same kinda idea, where iirc after 3 rounds the author is asked to rearchitect.

The only other bit we also stole from that was the idea of only looking for issues in the original diff rather than spawning new problems from each diff fixing the previous. Of course the possibility exists of bugs being introduced while fixing others but we try to handle that with human review for the minute because its not worth getting agents to cover that bit, at least with naive approaches


Yeah, I have an app with ~150 users, and Play Store updates slowed down to about a week for a moment, but lately they have been going out within the hour.

I think (2) is the important one. Anthropic provides benchmarks with their model releases. If they are making changes that make models perform worse on advertised benchmarks, how can this be legal?

It's a good point. I mean I have no idea if they are actually breaking the law, would assume their tc's would dissolve them from any legal jeopardy but it's still not right.

Is this where independent benchmark companies like artificial analysis could help ?


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: