Curious, cuddly & loving DJ/photo lion from Southern Germany :3 Partner of Stripey (@CynthiaFeline) ❤️ Catch me in SecondLife as "Mathew Jeffries"
Joined August 2015
- Tweets3.4K
- Following1.1K
- Followers183
- Likes5.4K
Mathew Jeffries retweeted
Thank you Germany 🇩🇪 and UK 🇬🇧 for heroically sacrificing your economy to save the planet 💪
_
Mathew Jeffries retweeted
Which local models can actually handle tool calling?
I built a framework to find out.
15 scenarios. 12 tools. Mocked responses. Temperature 0. No cherry-picking.
Tested every Qwen3.5 size from 0.8B to 397B, and since some of you asked after the distillation tests: yes, I included Jackrong's Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled too.
Only two models went all green: the 27B dense and the distilled 27B.
The 397B? Failed two tests. The 122B? Failed one. The 35B? Failed two.
The timed-out results — mostly on the smaller models, are cases where the model got stuck in a loop, repeating the same tool call until it hit the 30-second limit.
The test that exposed the most models: "Search for Iceland's population, then calculate 2% of it." Simple, but 35B, 122B, and 397B all used a rounded number from memory instead of the actual search result. They didn't trust their own tool output.
Small models hallucinate data.
Big models ignore data.
The 27B just threaded it through.
Mathew Jeffries retweeted
Qwen just dropped the open-source Claude Code killer.
It's called Qwen3-Coder-Next and it's genuinely wild.
Only 3B active parameters but beats models with 10x-20x MORE parameters on SWE-Bench-Pro.
PLUS: They launched Qwen Code CLI, the best open-source alternative to Claude Code.
1,000 free requests/day.
This is the model that proves size is a lie
Here's everything you need to know
Mathew Jeffries retweeted
🚀 Qwen3.5-397B-A17B is here: The first open-weight model in the Qwen3.5 series.
🖼️Native multimodal. Trained for real-world agents.
✨Powered by hybrid linear attention + sparse MoE and large-scale RL environment scaling.
⚡8.6x–19.0x decoding throughput vs Qwen3-Max
🌍201 languages & dialects
📜Apache2.0 licensed
🔗Dive in:
GitHub: github.com/QwenLM/Qwen3.5
Chat: chat.qwen.ai
API：modelstudio.console.alibabac…
Qwen Code: github.com/QwenLM/qwen-code
Hugging Face: huggingface.co/collections/Q…
ModelScope: modelscope.cn/collections/Qw…
blog: qwen.ai/blog?id=qwen3.5
Mathew Jeffries retweeted
What Claude Code has revealed is that most people either have mediocre ideas or no ideas at all. The tool is a force multiplier for those who already know what they want to build and how to think through it systematically; it elevates competence, rewards clarity, and accelerates execution for people who would have gotten there anyway, just slower. If you have a sharp vision and can break it into coherent steps, Claude Code becomes an extension of your own capability.
But there's another mode of use entirely. For people without that clarity, the appeal is precisely that the input can stay vague; you gesture at something, hit enter, and wait to see what comes out. This is structurally identical to a slot machine: low effort, variable reward, and that intermittent reinforcement loop that hooks the susceptible. So the same tool that elevates the focused and capable is also manufacturing a kind of gambling behavior in people prone to it.
Mathew Jeffries retweeted
The Ralph Wiggum Claude Code plugin is taking over X and for good reason
It allows Claude to work for DAYS by itself
I've been able to one shot multiple apps and features with it
In this video I cover how Ralph Wiggum works, how to install it, and how to build with it:
Mathew Jeffries retweeted
Raffle on my NSFW telegram channel!
Go check it out, you can win a fullbody artwork just as in these examples ✨
➡️t.me/CynthiaNSFW/1535⬅️
➡️t.me/CynthiaNSFW/1535⬅️
➡️t.me/CynthiaNSFW/1535⬅️
Mathew Jeffries retweeted
⭐️We wish you a happy new year and a good start into 2026! 🍀
Thank you for being so patient, understanding and supporting this year! Uni has been rough but you guys still stayed - so thank you so much!
Hopefully you have a good time tonight and enjoy it! 🖤
🐷🍀🎆
Mathew Jeffries retweeted
Hey! Decided to raffle off one KEKEKEKEEK animation!
All what you need to do is like+repost and post a cat(big cat) ref because kitties do kekekekekekee :3
Raffle ends on Aug 11
Lesgo!
Mathew Jeffries retweeted
The fandom's what you make of it.
I've met so many good acquaintances & friends, and made so many good experiences in all walks and ways of life through it; I wouldn't be such a well rounded person without it. If I didn't have the furry fandom, I'd have an extremely boring life.
Mathew Jeffries retweeted
Major AI breakthrough: Diffusion Large Language Models are here!
They're 10x faster and 10x cheaper than traditional LLMs.
Here's everything you need to know:
Mathew Jeffries retweeted
Hello people, it's time to do another one, please follow the rules because I'll check if everything is good when the date comes out.
It will be just one winner, Good luck everyone!
Mathew Jeffries retweeted
Fursuit Head Giveaway!!! 🥳
We've reached a huge milestone by completing 200 fursuits!
What a journey it's been. Fursuit making began as a small hobby back in late 2010/early 2011. The business then got certified and the number of fursuits just kept growing ever since. We are forever thankful for all the love and support we've received over the years and we are of course very proud to have accomplished all this ❤️
We absolutely love our fluffy family and fursuit making will always be our greatest passion. Here's to many many many more fursuits to come! 🥂🍾
As a way for us to celebrate this we've decided to make another fursuit head giveaway!
👉How to participate:
🔸️Follow us here on X.
🔹️Like and repost this post.
🔸️Comment with your reference sheet.
We will pick a winner once this post reaches 1000+ likes. The head is planned to be worked on during 2025.
