Hacker News Reader

anana_on Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint

The benchmarks they chose are rather cherry picked to not include long context or difficult ones that involve long horizon work or many agent turns, as I suspect this is where the model shows more differences compared to the full fat one

anana_on Nvidia is the central bank of AI

I'm curious what the actual rate of replacement/useful life of these GPUs running AI inference 24/7 is.

If these cards burn out in less than the ~5 years of depreciation that accounting puts them at, well then there will be problems.

anana_on Qwen3.8 27B scores 52 on Artificial Analysis

Agreed. Luckily, this model also scores high in AA non-hallucination, so it knows what it doesn't know -- perfect for situations where it can just tool call a web search.

anana_on Qwen3.8 27B scores 52 on Artificial Analysis

And to read the tea leaves a little:

3.8 actually performs slightly worse than 3.6 on AA-Omniscience Accuracy, which could imply that they traded out world knowledge for capability in other areas.

It also produces nearly twice as many tokens per task as 3.6 (and by extension, time), which may be a tradeoff required to achieve correctness at this parameter size.

anana_on Qwen3.8 27B scores 52 on Artificial Analysis

For more context, this puts it on par with models like GLM 5.2 and GPT 5.6 Luna, which are far larger

anana_on Qwen 3.8 27B

When stuff like this: https://doublespeed.ai/ exists I don't find that hard to believe at all, although it cuts both ways

anana_on Qwen 3.8 27B

Seems like MTP is available immediately!

anana_on Qwen3.8-27B

As was the case with GLM 5.3, it seems that there is still much juice to be squeezed from post-training

anana_on Qwen 3.8 27B

Monstrous benchmarks! Hoping it is not benchmaxxed.

anana_on GLM-5.3: Frontier coding with emergent cyber capabilities

What a week for AI model releases

anana_on Qwen 3.8

Apparently agentic performance in Gemma was improved recently: https://x.com/googlegemma/status/2077449152062247219

Too little too late imo

anana_on Ornith-1.0: self-improving open-source models for agentic coding

They keep mentioning a 31B dense model, but there are no benchmarks or weights for it anywhere?

anana_on Qwen-AgentWorld: Language World Models for General Agents

It looks like the purpose of this model is to i. generate environmental sim data for doing RL on other models or ii. act as a foundation model (they trained it to select actions as well as predicting the next state in the same loop?)

Either way, neither are intended for end consumers.

anana_on Qwen-AgentWorld: Language World Models for General Agents

I believe the benchmark listed is about simulating the environment for the various tasks, rather than doing them. It seems that the point of this model is to generate sim data to improve other models with

anana_on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?

Unfortunately on Strix Halo or any similar unified memory set up, dense models are gonna be dirt slow due to the tiny memory bandwidth... But I agree, 27B is superior.

anana_on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?

Perhaps try a different model? Just from anecdotal experience, I find that the Gemma models smaller than 31B do not tool call as often as they should.

Some of the benchmarks appear to back this up [0]

Of course, a lot depends how you are using it (inference parameters, harness, prompting, etc.), but the model is quite important too.

[0]: https://artificialanalysis.ai/models/open-source/small?model...

anana_on CrankGPT

I have one too and it never occurred to me to use it for anything other than games. Would be interested in seeing how you did it!

anana_on LinkedIn is scanning browser extensions

They do now - https://support.mozilla.org/en-US/kb/use-sidebar-access-tool...

anana_on Kimi K2.6: Advancing open-source coding

Hypothesizing here, but maybe the idea is sort of a form of technological/economic warfare? Releasing performance equivalent yet more cost efficient open weight models should in theory drive the cost of inference down everywhere.

This I assume will make it more difficult for US AI labs to turn a profit, which might make investors question their sky high valuations.

Any sort of melt down in the AI sector would almost certainly spread to the wider US market.

In contrast, in China, most of the funding for AI is coming directly from the government, so it's unlikely the same capital flight scenario would happen.

anana_on I ran Gemma 4 as a local model in Codex CLI

I'm not saying it's the latest Qwen iteration - that would be Qwen3.6.

I'm saying it's the latest iteration of the finetuned model mentioned in the parent comment.

I'm also not suggesting that it's "the latest and greatest" anything. In fact, I think it's rather clear that I'm suggesting the opposite? As in - how can a small fine tune produce better results than a frontier lab's work?

anana_on I ran Gemma 4 as a local model in Codex CLI

It's rather surprising that a solo dev can squeeze more performance out of a model with rather humble resources vs a frontier lab. I'm skeptical of claims that such a fine-tuned model is "better" -- maybe on certain benchmarks, but overall?

FYI the latest iteration of that finetune is here: https://huggingface.co/Jackrong/Qwopus3.5-27B-v3

anana_on I run multiple $10K MRR companies on a $20/month tech stack

Upon rereading, I'd agree. Fits with the tone of the rest of the write up.

anana_on I run multiple $10K MRR companies on a $20/month tech stack

> Sometimes you need the absolute cutting-edge reasoning of Claude 3.5 Sonnet or GPT-4o

Dead giveaway

anana_on Something is afoot in the land of Qwen

https://huggingface.co/Qwen/Qwen3.5-27B

I wasn't aware of that, which page mentions that?

anana_on Something is afoot in the land of Qwen

I've had even better results using the dense 27B model -- less looping and churning on problems

anana_on Warren Buffett dumps $1.7B of Amazon stock

They own GEICO...