Qwen 3.8 27B is excellent, but it defaults to overthinking things • Hacker News Board

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#311

Post by Gareth321 » Mon, Aug 17, 2026, 1:04 PM UTC

Phemist wrote: post #204

CamperBob2 wrote: post #113

Earlier quoted context omitted.

Chain-of-thought output shouldn't be taken literally. The tokens are a substrate for computation, not necessarily evidence that the model is wasting time and electricity by gratuitously second-guessing itself over and over. You can see evidence of this phenomenon in models dating back to the OG Deepseek R1. It was common to see the model talk itself out of the correct solution in the block, or fail to reach it at all…

I am interested however in why fine-tuning on reasoning traces of a frontier model is such an effective way of improving an (open-weight) base model. See e.g. https://huggingface.co/hesamation/Qwen3.6-35B-A3B-Claude-4.6... I can see the reasoning being a substrate for computation, but in which space should we interpret this computation to be happening? The vector representations of individual tokens are completely di…

Not the person you replied to, but I think a more accurate description of the reasoning we see is proof of effort, not necessarily great insight into how the reasoning is occurring. For the most part, researchers currently describe the intent and motives (in however one may define them for LLMs) as black boxes right now. Even the mechanics of the cognitive process is not well understood. Depending on the model and harness, the thinking will often look like gibberish. I suspect they've invested considerable effort into presenting thinking as a reasonable approximation of what they imagine it to be. Claude and OpenAI have also begun encouraging multi-step problem solving (or the models themselves decide this), and we can see their more accurate responses at the conclusion of each phase.

Fine tuning or post-training is effectively biasing certain outcomes: making them more likely to occur. This comes with trade-offs. A coding LLM will bias technical language, which would harm a model for general use.

This opens a really interesting field of research. Our brains use specialised regions because specialisation turned out to be the most energy efficient method for biological compute. It might also be the best performant. We don't want to activate 100% of our prefrontal cortex to breath. What a stupendous waste of the organ. I think we see incredible advancements in model clusters in the future, using specialised models for specialised tasks. We have the appearance of this today in some harnesses, but they are shallow imitations. The real innovation will be low-cost, accurate routing. Existing solutions are woefully inadequate for many reasons.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#312

Post by NamlchakKhandro » Mon, Aug 17, 2026, 1:04 PM UTC

madduci wrote: post #263

pdyc wrote: post #122

Earlier quoted context omitted.

i have same 255h and i was able to run it with low token speed 6-8tg/s with approx similar context window 60k

Interesting. What are you using? I was using ollama

Don't bother with plans

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#313

Post by 13639366668 » Mon, Aug 17, 2026, 1:05 PM UTC

[dead]

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#314

Post by dofm » Mon, Aug 17, 2026, 1:09 PM UTC

hgoel wrote: post #292

DanielHB wrote: post #226

Earlier quoted context omitted.

Is the hardware really getting better? It feels performance per watt is not getting better at all which is the metric that will matter eventually when supply-demand stabilizes. As it is, it seems the improvements are about making the hardware cheaper (as in capex, not opex). This is just feels from me from what I hear on the news and see on the products though.

The DGX Spark apparently consumes up to ~150W while being able to run many models at decent speeds. I think really good efficiency is possible right now, but the GPU makers don't want to make their consumer GPUs too good for AI - if the cards were more efficient, it'd be much easier to run multiple - while data center ones have a bunch of additional power overheads.

The Strix Halo / Ryzen AI situation also seems to have improved incrementally recently, both in hardware and software.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#315

Post by mattmaroon » Mon, Aug 17, 2026, 1:12 PM UTC

qeternity wrote: post #274

mattmaroon wrote: post #127

Earlier quoted context omitted.

There. I have the business plan with two seats and I use them both and blow through it pretty fast. I think it’s because much of what I have it do involves using a browser. For instance I have it pull various permits from cities and there’s no API for that.

Computer use will blow through tokens because it's doing image capture for everything. You may have better and more reproducible results using browser controls that aren't image based, or writing tools that completely sidestep browser use.

I do where possible. I build MCP servers into custom tools. But a lot of it is stuff like filling out permit applications on government websites, or reading festival websites and applying, etc.

I will say one thing that frustrates me is the opaqueness of the billing model. I basically just have to pay a random amount. I guessed that it navigating the web is relatively pricey from watching my usage as it does stuff. And the whole thing is worth 10x what I pay for it anyway, it just would be nice if the pricing were somehow more transparent, even in hindsight. It can explain to me what is thinking as it does stuff, it could also explain the tokens.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#316

Post by ninjahawk1 » Mon, Aug 17, 2026, 1:12 PM UTC

It’s a good model but I hope the obliterated version comes out soon. The main way I use open-source models is for doing things that server based models decline, which at the moment is quite a bit of tasks.

I find it very ironic how passionate Claude is about not violating copyright while simultaneously Anthropic was sued and lost the lawsuit for illegally pirating millions of books.

Lol.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#317

Post by mattmaroon » Mon, Aug 17, 2026, 1:13 PM UTC

rene-veerman wrote: post #170

mattmaroon wrote: post #151

Earlier quoted context omitted.

Huge gaps. I expect the tooling to get better for non-programmers. Codex and Cowork are great, but you still feel like you’re trying to hammer the square peg through the circle hole often when using it for non-programming tasks. The AI is good enough to do a lot of tasks but the tooling just isn’t caught up to it yet. I’d say it’s freed up ten hours a week of my time. And that’ll only improve.

how'd you get it to run? not even upgrading ollama could get it installed on my end :(

Oh I’m just using ChatGPT, I haven’t dives into getting anything to run locally yet. That’s on the off season project list.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#318

Post by mattmaroon » Mon, Aug 17, 2026, 1:13 PM UTC

ipdashc wrote: post #202

mattmaroon wrote: post #133

Earlier quoted context omitted.

Yes. Bureaucracy for sure when I throw festivals. ChatGPT pulls permits for me. (I think that uses a lot of tokens because of the browser control?) Manages the admin side along with some tools I built in Lovable via mcp servers. Marketing definitely. My food truck side is relatively high volume and it manages my kitchen and warehouse side, basically generating all of the instructions my employees follow, managing and…

Makes sense. Thanks for the post!

Do you work in the space?

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#319

Post by dofm » Mon, Aug 17, 2026, 1:21 PM UTC

suoloordi wrote: post #75

atif089 wrote: post #71

So if I have to set this up on my 24GB MBP what'd the right configuration and tuning look like?

Checkout MTPLX. I use the Youssofal/Qwen3.8-27B-MTPLX-Optimized-Speed model and I get around 12 tok/s on a 36GB MBP.

I still haven't tested MTPLX and the memory of reading about it had drifted out of my brain like so many things have, so thanks!

I really must.

Re: Qwen 3.8 27B is excellent, but it defaults to overthinking things

#320

Post by bbkane » Mon, Aug 17, 2026, 1:33 PM UTC

pferde wrote: post #303

ZaoLahma wrote: post #116

Earlier quoted context omitted.

Full agree. I until very recently thought AI tools of today were limited to prohibitively expensive high end hardware hosted in data centers. I was surprised and amazed to get "decent" (with the expectations set right / low) coding performance out of Qwen3.5-9B on a decidedly medium end Radeon 9070 paired with a 5700x3d and 32GB of DDR4 RAM. We can finally reason with and "talk" to our hardware.

I don't want to reason with my hardware, I want my hardware to do what I tell it to do.

As a desktop Linux enthusiast, I LOVE that I can paste error messages and descriptions of problems into a chat and get (usually) reasonable solutions for them.

As a recent example, my /boot partition kept filling up. ChatGPT walked me through the root cause (leftover kernels after updates), mitigations (deleting older kernels), and future prevention (installing unattended-updates and enabling its config file).