News for hackers. Stuff that matters.
Re:GLM-5.2 is the new leading open weights model on Artifici... by vorticalbox on Wednesday June 17, 2026 @07:00AM (#48568584)
This is a problem I find with opus is will spend so long thinking then going “but wait what if”

To point where I stop it and simple tell it to “start writing code you can work it out as you go along”

Seems writers block also effects LLM

Re:GLM-5.2 is the new leading open weights model on Artifici... by mikeocool on Wednesday June 17, 2026 @07:51AM (#48569066)
Seriously. Whenever I read the thinking output I get mad and turn down effort to medium or low.

Just output the code and we’ll work through it!

I feel similarly about having codex review claude’s plans. I don’t think I’ve ever seen it catch a major issue. It just points out things that would have inevitably been addressed during implementation anyway.

Re:GLM-5.2 is the new leading open weights model on Artifici... by SubiculumCode on Wednesday June 17, 2026 @12:18PM (#48572624)
A lot of times this is how humans work. Just start 'putting words on paper', 'think by doing', etc. sometimes it's more efficient to see why something won't work after writing a bit of it, and sometimes you get lucky and it works right off the bat
Re:GLM-5.2 is the new leading open weights model on Artifici... by thinkingtoilet on Wednesday June 17, 2026 @07:56AM (#48569125)
I've been having success with Opus but you REALLY have to tame it. Long prompts that list what files to look at, relationships between entities, etc... I went from regularly hitting my daily limit to almost never hitting it. Oh, and also I was being lazy with small changes and stopping that helped a lot too. As you said, it gets in these loops where it's just churning and if you don't stop it it can go on for way too long.
Re:GLM-5.2 is the new leading open weights model on Artifici... by epolanski on Wednesday June 17, 2026 @07:14AM (#48568691)
Fable was 20 times worse on that.

It's clear it was the vibe coding model, as like no other model before, fully turned you into his assistant instead of the other way around.

Re:GLM-5.2 is the new leading open weights model on Artifici... by RyanHamilton on Wednesday June 17, 2026 @07:56AM (#48569124)
Could it be possible, these firms are optimizing for two things: a) Better performance. b) Gathering data from you to further improve performance later. I've also found the huge amount of planning rather than iteration frustrating. I've felt like I'm teaching a junior!
Re:GLM-5.2 is the new leading open weights model on Artifici... by epolanski on Wednesday June 17, 2026 @08:03AM (#48569192)
I think they simply optimize around E2E benchmarks, none of those benchmarks is designed as multi turn assistant to the user, but going from a prompt to the final solution.
Re:GLM-5.2 is the new leading open weights model on Artifici... by celrod on Wednesday June 17, 2026 @02:11PM (#48574327)
Exactly. How can "we" develop and encourage benchmarks for multi-turn user assistance? That is what I want. I feel like the models and harnesses push much too hard against this workflow -- that they push you towards letting go and vibe coding, with only your discipline (and desire for a quality and maintainable product) holding it back.
Re:GLM-5.2 is the new leading open weights model on Artifici... by happyPersonR on Wednesday June 17, 2026 @09:54AM (#48570622)
more thinking == more tokens === more money LOLL
Re:GLM-5.2 is the new leading open weights model on Artifici... by overfeed on Wednesday June 17, 2026 @11:53AM (#48572226)
Os there a cost benchmark out there? I wonder how frontier models are doing over time for cost per problem solved.
Re:GLM-5.2 is the new leading open weights model on Artifici... by drob518 on Wednesday June 17, 2026 @12:20PM (#48572647)
I think they are optimizing for one-shot performance because that will drive usage. They can’t afford to look bad in the benchmarks. And if that means consuming an order of magnitude more tokens, well, that’s good for business, too.
Re:GLM-5.2 is the new leading open weights model on Artifici... by giancarlostoro on Wednesday June 17, 2026 @08:26AM (#48569422)
I usually have Claude build a plan first, then I put it into an XML file it updates with phases, usually we talk about some of those tasks, and then once its good and I like it, I have Claude implement the plan.

Another thing I tell Claude to do is to not guess, but look at documentation, it messes up a lot less, might use some tokens reading docs, but at least it has a higher success rate code wise.

Re:GLM-5.2 is the new leading open weights model on Artifici... by xstas1 on Wednesday June 17, 2026 @08:56AM (#48569825)
XML??
Re:GLM-5.2 is the new leading open weights model on Artifici... by giancarlostoro on Wednesday June 17, 2026 @09:08AM (#48570001)
Apparently because of how Claude is trained, even the system level prompts go through as XML, it works better with XML "prompting" so I figured I could have it write plans in XML. I need to update my ticketing tool to output XML maybe by default.

https://www.reddit.com/r/ClaudeAI/comments/1psxuv7/anthropic...

Re:GLM-5.2 is the new leading open weights model on Artifici... by saltsucker on Wednesday June 17, 2026 @09:48AM (#48570554)
Comments later in thread say markdown works just as fine and that it’s more important to organize your plan into sections.

Also just think about it, why would a model trained on the world’s corpus of text (that isnt formatted in xml) perform better with XML? It would be a better study if that post tested markdown, org, xml, json, etc. 10 times to see if their is a difference

Re:GLM-5.2 is the new leading open weights model on Artifici... by swingboy on Wednesday June 17, 2026 @11:53AM (#48572213)
Anthropic’s best practices still include the use of XML: https://platform.claude.com/docs/en/build-with-claude/prompt...
Re:GLM-5.2 is the new leading open weights model on Artifici... by adastra22 on Wednesday June 17, 2026 @11:08AM (#48571578)
A year or so ago XML worked more reliably for long-lived prompt instructions. Now it is cargo culting.
Re:GLM-5.2 is the new leading open weights model on Artifici... by orbital-decay on Wednesday June 17, 2026 @06:15PM (#48577700)
XML consistently performed better than markdown and JSON in all evals I've ever seen on any model, except for a couple very specific ones.
Re:GLM-5.2 is the new leading open weights model on Artifici... by root-parent on Wednesday June 17, 2026 @09:33AM (#48570348)
XML stands for Xtra ML....
Re:GLM-5.2 is the new leading open weights model on Artifici... by noworriesnate on Wednesday June 17, 2026 @11:23AM (#48571754)
I'd like to switch to a sales career--can you give me any pointers?
Re:GLM-5.2 is the new leading open weights model on Artifici... by aesthesia on Wednesday June 17, 2026 @04:59PM (#48576820)
One reason to use XML-like formatting is that it makes the beginning and end of sections explicit. This is less of an issue when the model is generating text but can still be helpful when using templated prompts.
Re:GLM-5.2 is the new leading open weights model on Artifici... by robertkarl on Wednesday June 17, 2026 @10:11AM (#48570830)
https://arxiv.org/abs/2606.00206

In this paper they nerf an LLMs ability to emit waffling thinking tokens like "wait", "but", "alternatively", and the models (they're old, small models in the paper) terminate reasoning faster and perform better. I bet Anthropic is tuning this on their backend.

Re:GLM-5.2 is the new leading open weights model on Artifici... by meatmanek on Wednesday June 17, 2026 @01:24PM (#48573597)
This is super cool. Do you know if any of the inference backends (llama.cpp, vllm, etc) support this technique?
Re:GLM-5.2 is the new leading open weights model on Artifici... by iaw on Wednesday June 17, 2026 @06:12PM (#48577669)
vLLM supports "banning" certain tokens but I don't know if it can dynamically reduce them.

To my knowledge you can also "ban" with llama.cpp but it is passed in the API call rather than to the server at initialization.

Re:GLM-5.2 is the new leading open weights model on Artifici... by orbital-decay on Wednesday June 17, 2026 @06:11PM (#48577661)
I imagine Anthropic would rather train a small control model instead of resorting to sampling hacks
Re:GLM-5.2 is the new leading open weights model on Artifici... by addandsubtract on Wednesday June 17, 2026 @10:33PM (#48579921)
Didn't they originally introduce those tokens to make the models smarter by second guessing their "thoughts"?
Re:GLM-5.2 is the new leading open weights model on Artifici... by drob518 on Wednesday June 17, 2026 @12:16PM (#48572593)
Qwen is notorious for this, too. It’ll sometimes spin in a long loop of “But wait…” paragraphs.