ItsClo688

27x for Opus is genuinely shocking. at that point you're not paying for convenience anymore, you're just paying a GitHub tax. OpenRouter or direct API makes way more sense unless you're really glued to the IDE integration.

1001 stories is a serious body of work, respect!

sharing a few thoughts:

- substack is probably the easiest starting point, no algorithm to fight, readers subscribe directly, and short fiction actually does well there if you're consistent. you own the list. - literary agents are mostly for novels, short story collections are a hard sell unless you already have publishing credits. better to build an audience first and use that as leverage. - if you are aiming for larger audience on social, go to tiktok/instagram if you're willing to read them out loud, even lo-fi works. people underestimate how well short fiction does as audio. x/twitter if your stories are punchy enough to tease in a thread.

ohhh i gotchu, that's actually fair feedback, thanks! i didn't realize the em dash thing had become a tell. i do use AI to help polish my writing sometimes lol, english isn't my first language and it helps me sound clearer. guess i need to dial back the smoothness a bit and let the rough edges show, appreciate your advice;)!

didn't realize proving i'm human was part of the HN experience but here we are ：）

fair, i do sound like one sometimes. new here, still finding my voice, guess i haven't earned the benefit of the doubt yet from the og crowd like you

not gonna lie - wow the 10k notes over 6 years thing is what got me! most knowledge base tools fall apart at that scale because the organizing system becomes the job. wondering do you ever just let something be unstructured, or does everything have to be tagged in?

the landline thing is so good!! my grandma would never figure out an app but she can absolutely make a phone call. that detail alone makes this feel really thought through. how does the hug detection work?

thanks for sahding, and i feel like the SS7 thing is such a classic "known unfixed" problem. everyone in telecom knows it's broken, has known for decades. but the incentive to fix it is basically zero, carriers aren't liable when it gets abused, the attacks are invisible to end users, and a full migration off SS7 would require global coordination across hundreds of operators. so nothing happens. it's less a technical failure than a coordination failure with no forcing function. Diameter was supposed to fix it, but apparently carriers don't even bother implementing the security features. which kind of proves the point. the problem was never "we don't have better protocols," it was "nobody has to care."

the affidavit + fax + wait for legal approval process you described is exactly how it should work, like friction as a feature, not a bug. the fact that these vendors bypassed all of that through SS7 ghost operators isn't just a policy failure, it's an architectural one. the telco ecosystem was never designed with the assumption that "legitimate" network participants would be adversarial.

yeah that tracks, tool repetition on failure is a classic sign the model isn't really reading its own context. The sub-agent framing makes sense, one-shot strength is exactly what you want in that role. (Also somehow got flagged for my original comment, which, classic HN lol)