nilsherzig

4 ms·

1.
by nilsherzig 8d ago
Fyi, if you're trying to run this under AMD/HIP: PTQ1_0 has no optimized MMQ-Path in their llama-cpp fork, try running PTQ2_0 (needs a bit more vram, but is about 2x faster on my 6700 XT) https://gist.github.com/ni
2.
by nilsherzig 14d ago
Please explain what you mean by "AI assisted usefull content is always fine". Here is a detailed explanation from Pangram how their scoring works https://www.pangram.com/research/how-it-works
3.
by nilsherzig 15d ago
the website of this company is using more resources than their main product is (would really appreciate a button to turn of the animation)
4.
by nilsherzig 15d ago
yea fair, probably not
5.
by nilsherzig 15d ago
There you go :) https://hnslop.nilsherzig.com/ as an extension so you can continue to comment and use hn as usual
6.
by nilsherzig 15d ago
"Made with <3 by ..." Yea it is
7.
Show HN: Extension to filter LLM written articles(hnslop.nilsherzig.com)
16 points by nilsherzig 15d ago|3 comments
8.
by nilsherzig 15d ago
added a firefox extension :)
9.
by nilsherzig 15d ago
Will do, thanks for hcker.news btw, has been my main mobile client for couple of months. Of all things i like the changelog most, its nice to have a quick view about new stuff, especially if the website changes somewhat faster than usual.
10.
by nilsherzig 15d ago
Yea very expensive (hence me siphoning another page haha). Im trying to get something cheaper to work, Pangram has some nice docs on how to build something like their service https://github.com/pangramlabs/EditLens , th
11.
by nilsherzig 15d ago
nvm sorry, just read the "how this works" comment from OP. https://news.ycombinator.com/item?id=49660301 OPs solution does more than keyword filtering. But its still about filtering out "content about ai&quot
12.
by nilsherzig 15d ago
If a Userscript solution is fine for you, see my recent comments :)
13.
by nilsherzig 15d ago
Here is an approach to filtering LLM written content: https://hnslop.nilsherzig.com/ Instead of filtering keywords this ranks the content using Pangram. But it's also pretty effective at filtering content about AI. Tur
14.
by nilsherzig 16d ago
Not going to discuss this further, but wanted to apologize for my message. I was annoyed and should've waited before sending. Sorry.
15.
by nilsherzig 16d ago
Pretty decent output, mind sharing the prompt and model used? I struggle getting text i enjoy reading https://www.pangram.com/history/5a4b90c6-0123-4f97-b50e-9c90...
16.
by nilsherzig 16d ago
If you really wanted to optimize something involving living consciousness beings like a machine, at least you should consider things like the need to take a piss as a hard constraint and design the system around it. Don't get me wrong
17.
by nilsherzig 17d ago
Weirdly still so much more human (and somewhat relatable) than he ever was in an interview
18.
by nilsherzig 18d ago
Maybe... if run locally with full transparency into the prompts (including system etc) and independent benchmarks for the models (i.e. tests on how eager they are about selling me stuff) But even in this hypothetical world I would still hav
19.
by nilsherzig 18d ago
Nothing will ever beat this one https://cf.preview.redd.it/a-photo-of-an-astronaut-riding-a-...
20.
by nilsherzig 18d ago
Dogs shadow didn't change
21.
by nilsherzig 18d ago
i really enjoyed that one
22.
by nilsherzig 19d ago
Still flags as 100% LLM written https://www.pangram.com/history/cc9b6131-c772-4b3b-9a5d-8e81...
23.
by nilsherzig 20d ago
update: https://news.ycombinator.com/item?id=49582725 shared another page which already does the parsing and pangram requests. I wrote a little wrapper and userscript based extension around it ( https://github.com
24.
by nilsherzig 20d ago
created a little wrapper and userscript based browser extension https://hnslop.nilsherzig.com/
25.
by nilsherzig 20d ago
I have (an API, not an extension), but pangram is way to expensive for me to run, current (v4) pricing is $0.05 per 100 words. Should be absolutely doable if we split the cost between users tho. I tried searching for good opensource /
26.
by nilsherzig 29d ago
checks out: > In 2006, SoftwareOnline was sued by The Washington State Attorney General's Office for alleged violations of the Consumer Protection Act after complaints were made about two products called "Registry Cleaner"
27.
by nilsherzig 1mo ago
Removing stuff without mentioning the removal. Asking an LLM to remove an Idea from a Document often times results in an edit which explicitly states that this Idea is not relevant, instead of just removing all references to that Idea. That
28.
by nilsherzig 1mo ago
You can get better results from less aligned models like Kimi K3. Still not actually funny, but at least it’s able to produce some unhinged stuff and I guess shock and twists are kinda related to humor? Still missing the human connection of
29.
by nilsherzig 1mo ago
That's a Claude thing btw. Try using a harness which does not inject half a novel of instructions in combination with a different model. I would recommend Pi + GPT 5.6 Luna for a very capable and cheap test. After using Claude (paid by
30.
by nilsherzig 1mo ago
oh looks like they killed xcancel :/ https://xcancel.com/Eminem/status/2092367272891310482

More ›