@VectorCrossProd

Lead on PCB design software TraceMaker (https://xitter.cf/t.co/QXXc9mLbl9), and LLM visualisation llm-visuals (https://xitter.cf/t.co/nhYsCUxUdh).

 Australia
kookabus.com
Joined July 2026
  • Tweets2.2K
  • Following2.5K
  • Followers1.9K
297Media
Nigel Hungerford-Symes retweeted
沐阳
@yyyoleOct 24
绝了！NVIDIA最新开源了2,575 个高规格 3D 资产，整整472GB！！ 从日常用品到工业设备，里面收录了不少现实环境里的物体。 一揽子，全部放进了 Hugging Face。 NVIDIA SimReady Assets。 绝不单纯是一个“免费建模素材库”！！ NVIDIA给每个3D 资产都带上了物理属性和语义信息，比如有多重，什么材质，重心在哪，用在什么场景等等，全都写进了参数里。 明摆着，就是用来让AI学习的！ 学习怎么和真实的物理世界互动！ 传统的3D资产在AI眼中，就是个有贴图的“视觉资产”，NVIDIA想做的是，把“视觉资产”变成AI可以理解的“物理资产”！ 所以，这批资产的价值可想而知！ 至少能够用到这些场景上！ 机器人抓取和搬运训练，先在虚拟环境里练习拿取、摆放物体； 搭建工厂、仓库的数字孪生，测试设备和物体的布局； 配合合成数据工具，改变光照、视角、遮挡和物体位置，生成带标注的训练数据。 大家都在卷文字、代码与图像生成时，NVIDIA 开始为Physical AI（物理 AI） 铺设地基了。 不过，对大部分人来说，也是一个相当全面的3D资产素材库！ 链接在下面，必须收藏了！
46
403
31
3,695
197,743
Nigel Hungerford-Symes retweeted
Marco Mascorro
@MascobotOct 19
Turns out there's a livestream straight from the JLCPCB factory where you can watch PCBs being manufactured in real time:
15
48
7
478
34,486
Nigel Hungerford-Symes retweeted
JLCPCB
@JLCPCBOct 13
Why does a $0.01 part come with a $3 fee? 🤔 On Economic PCBA, Basic parts never leave our machines. Extended parts get loaded onto a feeder just for your job. That setup is the fee. Skip it: pick Basic or Preferred Extended parts. #PCBA #JLCPCB
6
2
2
42
5,826
Nigel Hungerford-Symes retweeted
Niko Veit
@coldnikoOct 11
New version is now live with another performance boosts and improvements on Strata github.com/Niko1221/Strata/r…
59
55
35
537
53,293
Nigel Hungerford-Symes retweeted
Niko Veit
@coldnikoOct 2
Prompt t/s: 2,170 -> 2,653 (+22%) Output t/s: 64.5 -> 76.4 (+18%) Enjoy new performance boost! 👀 github.com/Niko1221/Strata/r…

Release Strata v0.1.36 · Niko1221/Strata

Faster: Q2_0 reads prompts 16-22% faster and writes up to 18% faster at long contexts. Plus one-click updates (UPDATE.bat), an opt-in memory of which experts you use, and clearer logs. Speed (#136)...

github.com
48
56
16
521
34,826
Nigel Hungerford-Symes retweeted
Seve
@seveibarSep 30
First vibe-coded linux board(?!), made with tscircuit! Needed a small patch due to selecting the wrong RAM voltage (we'll add validation for this!) We were able to SSH in over USB! Let the linux everywhere era begin!!
17
15
1
253
10,175
Nigel Hungerford-Symes retweeted
Niko Veit
@coldnikoSep 29
Qwen3.8-Flash-Next first 32K tokens are 4%-10% faster, LOW RAM mode added (for low hardware). New update here github.com/Niko1221/Strata/r…

Release Strata v0.1.26 · Niko1221/Strata

Prompts up to 32K read 4-10% faster: the draft layer's prompt pass runs in batches. Before answering, Strata runs its draft layer over the prompt, so the drafter (MTP speculative decoding) know...

github.com
21
19
3
178
9,868
Nigel Hungerford-Symes retweeted
Elon Musk
@elonmuskSep 14
Replying to @itslueul
Grok 4.7 should be roughly on par with Opus 5.0, not 5.1. Better in some ways, worse in others. We need to fix multimodal performance. Grok 4.8 will be a noticeable improvement. Grok 4.9 is probably Astra/Fable class. Grok 5 maybe better than anything. We shall see.
1,101
965
716
15,394
2,413,090
Nigel Hungerford-Symes retweeted
DHH
@dhhSep 9
Nobody has taught me more about positivity and engagement than @garyvee. And I still think about this clip all the time. Even makes me a little teary-eyed. Because it's so true: YOU'RE SOMEBODY NOW! ❤️
90
163
37
2,786
118,977
Nigel Hungerford-Symes retweeted
lauren
@potetoAug 31
small update for Grok @Bot linux users: we now have links to download AppImage and rpm on x.ai/bot
58
31
6
528
40,116
Nigel Hungerford-Symes retweeted
Grok Bot
@botAug 26
All SuperGrok and Cursor Pro subscribers now have access to Grok Bot. We're also resetting weekly usage limits for all users. Enjoy!
1,649
1,186
813
18,274
55,333,649
Nigel Hungerford-Symes retweeted
DHH
@dhhAug 25
WARNING: Omarchy dot net is not legitimate. It looks like a scam/malware/phishing operation. I alerted their hosting service @spaceship, which acknowledged receipt, but refused to act. Wild that they want to be part of that, but ok!
59
104
12
1,677
160,971
Nigel Hungerford-Symes retweeted
zerohedge
@zerohedgeAug 17
Diesel crack spread $100 any second. This is the real Hormuz hedge, diesel prices around the world exploding and industrial economy relies entirely on diesel not gasoline
65
425
45
2,271
265,048
Nigel Hungerford-Symes retweeted
Patrick Vlaskovits
@PvAug 10
If the “watermark” is embedded in the statistical sequencing of text ie it can’t be stripped out without rewriting the text - doesn’t that necessarily make LLM text generation less useful? The model is no longer performing purely for the user’s request. Some fraction of its freedom in choosing words, phrasing, and structure is now being spent on preserving the watermark. You get an inherent tradeoff between watermark robustness and model flexibility/compliance. Ergo, the stronger & more robust the watermark, the greater the constraint. In other words: LLMs will suck more and be harder to work with.
NIK
@ns123abcAug 10
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
158
117
44
1,742
447,650
Nigel Hungerford-Symes retweeted
Alexey Fateev
@superaleshaAug 7
Minimax H3, same 4x3090 rig, same clip. Yesterday it rendered in 11:21. Today it renders in 3:45 🤯 The hardware did not change. Two things happened overnight. An AI agent rewrote the attention CUDA kernel and hit a half speed trap inside GeForce tensor cores that most people never heard of. Then a stranger on HuggingFace dropped a LoRA that cuts 20 sampling steps down to 4. One of these two mattered way more than the other. Breakdown below. I packed all of it into ready ComfyUI workflows for my rig. If you want them, ask in the replies and I'll share.
Alexey Fateev
@superaleshaAug 6
Just sent Prime Agent with Kimi K3 to optimize the SageAttention Kernel for more efficient Minimax H3 inference on RTX 3090. The issue is that the handwritten Cuda Kernel for Ampere loses to the Triton version from the same library. on my shapes it gets 85 ms versus 74 ms with identical accuracy. Triton autogenerator beats human code. Looks like a tile sizing problem. the 128/64/32/64 dimensions are hardcoded constants tuned for A100. 3090 has 100 kb shared memory instead of 164 kb so the resource balance is completely different. lets see what it finds.
29
28
7
494
72,319
Nigel Hungerford-Symes retweeted
Mason
@webdevMasonAug 3
All that needs to be said about pit bulls was said by S-tier dog trainer Stonnie Dennis in 3 minutes
Leading Report
@LeadingReportAug 1
Jon Bernthal says fear of pit bulls is “dog racısm.”
1,402
4,997
667
56,415
6,754,027
Nigel Hungerford-Symes retweeted
Space and Technology
@spaceandtech_Aug 2
🚨 Elon Musk says SpaceX will attempt its first Starship upper stage tower catch on Flight 14.
115
430
39
7,567
198,877
Nigel Hungerford-Symes retweeted
Qwen
@Alibaba_QwenAug 3
📢Meet Qwen3.8-Max — our most capable model to date. Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉 Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters: - Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub：github.com/qwen-code-dev-bot… - Real work, real results: Production-quality deliverables across hundreds of professions. - Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy. - Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction. 💰Pricing: Input: $2.0 / M tokens Output: $6.0 / M tokens Implicit Caching: $0.25 / M tokens Start building with Qwen3.8-Max! 🚀 📖 Blog: qwen.ai/blog?id=qwen3.8 ✅ Qwen Studio: chat.qwen.ai/?models=qwen3.8… ⚡ API: qwencloud.com/models/qwen3.8…
1,230
2,788
1,982
22,565
7,261,134
Load more