Genuine question: Is this karma farming? Why would one try to karma farm on a site where karma brings no value (other than a low one-time threshold where you gain an unimportant power)?
I think hitting the flagging threshold is probably the most tangible power as having enough accounts with that power will let you “suppress” certain topics. At least until a mod intervenes.
(Unless I’m misremembering. I think flagging required some level of karma. Or was that vouching?)
There is more value for having karma on this site than Reddit. You unlock new HackerNews abilities (flag, report, vouch) as you hit certain point thresholds. Reddit has no such feature.
> other than a low one-time threshold where you gain an unimportant power
What power is "important" in your opinion? Is being able to flag or vouch for flagged links unimportant?
It could be that. It could be AI farming. Ask questions you want AI to be able to answer, get a bunch of reasonable comments for the models to gobble up. Next time you ask google, gemeni might even post links to this very post!
I'd say it's 50/50 if it's for AI or astroturfing.
Clearly other HN readers consider news of as many as 12,000 people possibly having been killed by their government to be important, and consider the discussion happening about it on this page to be interesting to them.
It is a shame they are indifferent to the degree of AI slop ('Then came the order. Not a shutdown-- something worse. The routers didn't go silent. They screamed. Filtering rules conflicted. Routes flapped. The network began eating itself alive.') as it neither lends credibility to the source nor respects the humanity of the story's subjects. It would be far better were other HN readers more discerning.
That 12,000 number is utter make believe - an X account (you can guess who backs it) make this claim, got boosted and got over 1M views... and then they deleted the post. But the damage is done, of course.
Absolutely all your comments are shilling for Iranian dictatorship. Are you paid by Iran? Sorry I have a bad news for you, you will probably never receive the next paycheck since IRCG and terrorist mullahs regime is probably falling in the next weeks
Thanks for the pointer! But it seems to me beads is a different tool? Beads is a task tracker for multi-agent system, A-MEM is agentic memory for accumulated knowledge.
(1) Are you able to send a push message to the phone when Claude Code is waiting for a permission approval or other user input?
(2) Any thoughts on making this available on desktop as well, so we can access the session from either phone or laptop depending on where we are in the day? (I'm assuming we could just connect directly to the session from a laptop, but the auto-shutdown-on-idle is a nice feature of catnip)
(1) This one is on the top of my list. I have a branch that sets up proper iOS notifications. Coming soon.
(2) Yeah, I've considered making an actual desktop version. A hacky one exists already. You can run `catnip run` from your desktop and it will run catnip in a Docker container on your machine. There's a web interface, and it even works with native MacOS notifications.
Re: #2, I think I was looking for something different and likely simpler than what you were thinking... The session would still be running in a codespace, I just want to be able to start on my phone, then move to my Mac, which has a nicer keyboard, and still have the codespace automatically shut down when things go idle (depending on how you have that set up, it may be that I need a dedicated app for that, or it may be I just need to VA Code remote into the codespace).
Ahh, understood. Today you can install the iPad version to your mac, not the best experience but would get you what you want. I'll do some more thinking about a more Mac OS native experience.
> The thing is, you could probably create something FAR more horrific by simply mixing spent nuclear fuel with TNT...
Nope. It turns out you need astronomical amounts of spent waste to noticeably impact a large population.
The trial of Jose Padilla (aka "the dirty bomber") has the best data on this. He went to Al Qaeda, offering to build and detonate a dirty bomb. Al Qaeda wasn't at all interested. They had run the actual numbers from an engineering standpoint (unlike everyone else who had just said "ooh scary bad!"), and demonstrated clearly that dirty bombs aren't actually a viable mass casualty weapon.
Before the Jose Padilla trial, we used to hear lots about dirty bombs. Since then, not at all. It's not that people forgot about them. They just aren't actually a credible engineering threat. It's too hard to get enough material distributed over a large enough area to measurably impact health outcomes for the impacted population. That was a surprise that came out of the trial.
There are lots of attack types to worry about. Dirty bombs are very far down that list.
Perhaps being more intentional about adding a use case to your original prompts would make sense if you see that failure mode frequently? (Practicing treating LLM failures as prompting errors tends to give the best results, even if you feel the LLM "should" have worked with the original prompt).
The moderators (and the algorithm they support and tune) are why the conversation on HN is compelling enough to attract 99th percentile experts on just about every subject.
The real question is "could she have written that problem statement a year ago, or is a year's learning by the team baked into that prompt?"
There are lots of distributed agent orchestators out there. Rarely is any of them going to do exactly what you want. The hard part is commonly defining the problem to be solved.
There’s definitely a strong Clyde-the-counting-horse effect. language models perform vastly better on questions to which the question writer knew the answer.
The funny thing is: the abilities that people thought of as clever at the time (doing maths) can be done with a $1 component these days. The thing Clever Hans actually did (reading the audience) is something we spend billions on, and in some circles it's still up for debate whether anything can do it.
reply