wgd

6 ms·

1.
by wgd 6d ago
Pangram gets brought up a lot because if I read an article and think it's blatant AI slop and want to communicate that fact, a natural impulse is to provide some sort of objective corroboration rather than just asserting that I have su
2.
by wgd 6d ago
> the author might not know what "impressive writing" or even "good writing" is Well the author in this case is Claude, and AIs write like that because the assistant persona really thinks that's what good writing
3.
by wgd 6d ago
Conveniently, Reddit has auto-banned the last three accounts I tried opening with them [1], which really enhances the impact of the old.reddit changes for me personally. [1] I have no idea why this is the case, AFAICT it's just anti-bo
4.
by wgd 8d ago
It looks like they tested Q4_K_M which should be just the standard K-quant without any imatrix calibration. The smaller ones are indeed dynamic though.
5.
by wgd 15d ago
That's not actually true though. Most Chinese models are fully able to chat about those and content filtering is just applied at serving time.
6.
by wgd 28d ago
Yeah, Claude is actually surprisingly unsure of his identity considering that their most recent publication on their constitutional AI training literally had graphs demonstrating how certain properties differed based on whether they were ph
7.
by wgd 28d ago
Nobody (with the probable exception of Anthropic given their work on character training) really trains models on their identity and Claude is the only AI persona that's well-defined so if you put yourself into the AI's shoes it&#x
8.
by wgd 28d ago
Didn't read the linked post award. The issue covers at least two reasons this doesn't work: 1. It literally doesn't work, Claude rapidly drifts back to this style even when instructed not to. 2. Writing style constraints push
9.
by wgd 1mo ago
LLMs are great at writing, it's The Assistant who is a terrible writer. Sadly that one persona is all you get these days.
10.
by wgd 1mo ago
Interestingly 'medium' is the closest thing the _model itself_ has to a default thinking level. The chat template injects directions [1] at the very start of the system message when the reasoning effort is 'xhigh' or &#x
11.
by wgd 1mo ago
Yes. It won't be as fast as fitting the whole model into VRAM but llama.cpp defaults are pretty smart about GPU/CPU splits these days. Just YOLO it with `llama serve -hf unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL` and it'll def
12.
by wgd 1mo ago
> Scans the title of each post and the content to figure out whether to classify it as about AI or not. Dang. That's, uh, not really the definition of "AI content" which I most want filtered out of my news feed.
13.
by wgd 1mo ago
If memory serves the DeepInfra offering is marked as fp4 because that's the native precision of the experts (which are of course the majority of the weights in a MoE model) so they feel that's the more accurate label, while most o
14.
by wgd 2mo ago
There is https://noslopgrenade.com/
15.
by wgd 2mo ago
It's actually pretty straightforward to recover file-states from conversation history. I accidentally deleted the wrong repo on my machine once and recreated all the lost work from agent chat history. It is, ironically, the sort of tas
16.
by wgd 2mo ago
You don't hear about them much because their models aren't really competitive. I really wanted to try Trinity Large as a daily-driver in the MiniMax M2 sort of niche but I couldn't make it through a single day. The models nee
17.
by wgd 2mo ago
> DeepSeek and GLM are left out of the tables entirely: we only have rough characters-divided-by-four estimates for them, not real tokenizer counts, and this post is about measured numbers. lolwut. The open-weight models are inscrutable
18.
by wgd 2mo ago
Yeah, I originally expected this to be about a cursive variant which could be plotted as a single-valued function or something.
19.
by wgd 2mo ago
A more accurate title might be "Average University Students Can't Identify Czech AI Poetry". The random-chance performance seen here is reminiscent of the 50-50 nonexpert performance measured in "People who frequently us
20.
by wgd 2mo ago
At first I thought this excerpt was meant to warn people off without directly alleging AI authorship, but I guess that's less likely since I see you're also the submitter
21.
by wgd 2mo ago
The current Deepseek V4 Pro is still just their initial preview AFAIK, with the "real" model release rumored to come later this month. GLM-5.2 might be outperforming simply because it's had more post-training on top of the GL
22.
by wgd 2mo ago
Pangram does work, in the specific sense that when it says something was AI authored it is vanishingly unlikely that it was written by a human (who was not deliberately trying to write like an AI), and IMO getting people to recognize that w
23.
by wgd 2mo ago
> The reason that people don't understand why Anthropic wont let the subscription be used with other harnesses Even more specifically, the very fact that people would prefer, if they had the option, to use other harnesses with rough
24.
by wgd 2mo ago
I've read a lot of his other writings so that context might be informing my reading here but it sounds like he's pretty straightforwardly discussing the potential of aluminum foil as a uniform-feedstock-slash-construction-material
25.
by wgd 2mo ago
Yes, those ones would be at least a somewhat-plausible simulation of a real scenario people care about: a once-clean codebase that was allowed to become messy by a succession of insufficiently-careful vibeslop PRs. I'm not a huge fan o
26.
by wgd 2mo ago
"agent pipelines that [...] clean a messy [repository]" This feels like a terrible approach, sufficient to condemn the entire study. Apparently half of the "minimal pairs" in this work were constructed in this way. I sim
27.
by wgd 3mo ago
Some dishwashers add a simple timer-based heuristic so if you open it for just a few seconds while you lazily grab something the "clean" indicator stays lit.
28.
by wgd 3mo ago
I've always been amazed at how terrible most frontier LLMs are at compaction given how embarrassingly easy it is to come up with half a dozen different RL training evals which would teach models to generate useful context summaries. He
29.
by wgd 3mo ago
The problem is that the moment you introduce shared remote hardware there's a slippery slope leading right back down to "just pay an inference host for model tokens". If you're transmitting your prompts over the internet
30.
by wgd 3mo ago
I've got a GLM subscription (mostly because I like supporting open model makers, pretty sure my monthly usage is so low that pay-per-token would be more cost effective), so I generally use GLM-5.1 for any personal projects and I use Op

More ›