Nole Doe (quant arc) (@noledoe42) header
Nole Doe (quant arc) (@noledoe42) profile picture

soon (tm)

orderbook
Joined May 2023
Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)Photo by Nole Doe (quant arc) (@noledoe42)
JIT (mini) apps/tools will become norm. On-demand, personalized, and context-aware. 9 users = 10 ways to use your app. Lab testing can't predict real-world creativity. SV designers ≠ rural grandmas. Now, billions of non-technical users have an AI app genie in their pocket.
2
2
1,326
Nole Doe (quant arc) retweeted
fable 5.2 helped me try a crazy idea today. every sound you hear is 259 human throat simulators singing an assigned note. your brain fills in the details. there are no guitars. no drums. no lyrics. just these blue guys going "ooh"
30
22
9
371
86,881
Nole Doe (quant arc) retweeted
Making sense, now?
Favorite to replace Jensen at $NVDA
196
1,394
49
11,759
309,596
Nole Doe (quant arc) retweeted
Rust compiling slow? Wow look at that makepad.github.io/relax/ the AI solved it's own problem!
9
7
3
119
26,486
Nole Doe (quant arc) retweeted
I want to share a preview of something I've been working on called Exact. It's a new Agent Native app framework written in Rust. It's extremely fast and tries to let you build exactly the experience you want with no compromises.
183
307
215
2,918
615,380
Nole Doe (quant arc) retweeted
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵
539
1,256
738
9,065
2,392,473
Nole Doe (quant arc) retweeted
I'm slowly starting to get why DeepSeek doesn't offer a subscription. They actually care about transparency and giving you a solid experience without that feeling of getting scammed like you do with Kimi or Zai.
115
65
15
2,462
143,861
Nole Doe (quant arc) retweeted
we've hit the fucking frontier bois an autoprover where you never write a proof, and it's ahead of every specialist i've measured it against
7
12
2
197
9,125
Nole Doe (quant arc) retweeted
Yep, Anthropic contributed zero, literally zero value. All it did was to announce Mythos, said Mythos is going to hack everything, got 50 or so companies to pay millions for access. They helped zero security researchers. Their safety classifier trips when even reading a vulnerability report. OpenAI on the other hand, rolled out sensible KYC, gave access to Daybreak Blue, and see what the community has done with GPT-5.6-Sol (daybreak version). One creates all the fear-mongering and delivers zero value, and the other quietly just carries everyone forward.
gotta give it to openai and their sane daybreak program heif heist by harsh kvmescape by paulos php2shell by lcfr gitlab rces by joaxcar chrome full chain by moyix wordpress rce by hashkitten all largely thanks to openai daybreak giving researchers access to gpt 5.6 sol. anthropic took the opposite approach, gatekeeping access, there is a lesson to be learned here
31
95
6
1,923
125,074
Nole Doe (quant arc) retweeted
supabase would charge ~$33,000/month to sync 100gb live to 1,000 concurrent users. convex ~$27,000 i'm doing it for $9 on one cloudflare durable object 🧵
40
54
3
1,250
81,697
Nole Doe (quant arc) retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4
1,766
4,939
2,351
43,871
7,114,836
Nole Doe (quant arc) retweeted
Source is out: github.com/VictorTaelin/Unii… I've been using Unii as my only agent for a month, and I couldn't be happier. I'm free from the torture of /compact, AGENTS.md, managing Claude sessions, and I can't overstate how empowering that is. I also keep receiving enthusiastic DMs of people who forked it to do their own thing, but I'm concerned that your AIs might get some tricky parts wrong, like caching and the exponentially decaying logs. So, please, ask it to study my codebase and fix anything that might be wrong in yours. I just want to make sure you get the right experience, because it is actually magical. People have ported it to OpenCode, Pi, and many other harnesses. If you did so, feel free to post it below. Also check UniiChat.com for my exact version!
My harness: UniiChat.com This is the closest to infinite context I've experienced. I've been using it for a week to organize my life and build things (including itself!), and it just keeps going. I only use it now. OSS next week, prompt to build your own below ↓
88
61
19
1,792
78,948
Nole Doe (quant arc) retweeted
Language models can learn to detect when you read their activations. How? 🧵
83
166
117
2,199
280,667
Nole Doe (quant arc) retweeted
Ben Affleck says AI is simply matrix multiplication, linear algebra and floating point calculations, and however convincingly it sounds like a person, it has nothing to do with thought, humanity or feeling “Nor do I believe, and I think it’s important to draw the distinction between what AI does, which is simply predict the next token or the next word in a sentence, then add that word on, run it back to the machine, do a lot of computation and pick the next word." "That is a very different thing than sentience or thought. And I think we need to be really clear about that because it can be convincingly, in an illusory way, sound like thought or like a person." "But let’s be real clear. This is matrix multiplication. This is linear algebra. These are floating point calculations. It has nothing to do with thought, humanity, or feeling.”
Ben Affleck reveals he writes Python, understands convolutional neural networks, worked extensively with GPUs, and used his celebrity status to get private looks at Google and OpenAI’s video models “I’ve always been kind of into computers since I was young. Then, when film started to move from analog film to digital, I became more interested in that aspect of it. The visual-effects workflow for many years has included machine learning, so I can write pretty shitty Python scripts and stuff like that. “With convolutional neural networks, which were the precursors to what the transformer can do, which is much more computation simultaneously, you would do things like look at what’s called a tensor. That’s the numerical translation of a visual image in numbers, like the batch number, the frame number and the red, green and blue values of each pixel in each frame. It’s just that simple. That numeric is called a tensor. “You’d use a convolutional neural network to identify patterns that reveal what’s called edge detection or feature extraction, which is identifying patterns well enough to know, this is where the window ledge is, so we can more easily take the green-screen image out and replace it with something. “That was familiar to me early on because, prior to Artists Equity, I had a small visual-effects company. I’ve worked with GPUs a lot too. The visual-effects guys said, ‘Hey, you should see. There are a couple: Google and this other company, OpenAI, are doing really interesting stuff with transformers in video.’ “I’ve learned that I can actually just call up and go, ‘Hey, it’s Ben Affleck. Can I come see what you’re doing?’ Sometimes people say yes, to my astonishment.”
800
2,229
712
17,591
2,396,843
Nole Doe (quant arc) retweeted
DeepSeek V4.1 Flash 1017 的 Blender 建模测试——居民楼 提示词:用Blender构建一个细节拉满的一幢典型中国90年代居民楼的3D场景。
60
37
6
628
43,994
Nole Doe (quant arc) retweeted
🚨 Codex usage limits are draining WAY faster than before. I’m on Plus, and both my weekly quota and 5-hour limit are disappearing ridiculously fast. Luna Max used to last much longer, and although Sol 6.1 consumed more at launch, things have gotten noticeably worse since the “28 days of nothing” began. Even switching back to Luna Max hasn’t helped, despite my workflow staying the same. Is anyone else seeing their Codex limits disappear this quickly?
303
96
33
2,072
169,134
Nole Doe (quant arc) retweeted
In July we open-sourced a dynamic graph of everything a child learns in primary school. 1,590 concepts, 3,221 dependencies. Now we’re turning every concept into interactive experiments kids can play and learn from. First one, age 7: heat and water. Everything that used to be a boring textbook can now be an engaging experience, potentially tailored to every child. Kids education just got a massive upgrade.
205
839
110
8,027
443,345
Nole Doe (quant arc) retweeted
Today we release 𝚍𝟷-𝚞𝚕𝚝𝚛𝚊, our most powerful decision model. > built for hard decision problems. > d1-ultra is multimodal > supports 44 languages > #1 on the @ArtificialAnlys Decision Index (65) > leads all eight categories of our 112-evaluation suite 1/
65
186
87
1,615
304,060
Nole Doe (quant arc) retweeted
Rejoice 16 GB M6 Mac mini owners 🍎 Ternary Bonsai 2 27B now runs on the Neural Engine: - 52 tok/s on code (GPU: 36) - 300 tok/s prefill (GPU: 185) - 12 W (GPU: 30 W) - Works with Grok Build Reproduced on a second M6 mini. github.com/cowboycoderhq/Ter…
4
5
4
43
11,255
Nole Doe (quant arc) retweeted
Today we release 𝙻𝙵𝙼𝟸.𝟻-𝟸𝟺𝙱-𝙰𝟸𝙱, the largest 𝙻𝙵𝙼𝟸.𝟻 family member. It is an efficient local model for laptops and PCs to work through multi-step tasks inside popular agent harnesses. > 24B total parameters, 2B active per token > Context expanded to 131K tokens > A 128K-token vocabulary > 33T pretraining tokens > It is a reasoning model > Very easy to fine-tune and customize > Performs competitively with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks. > Open base and post-trained weights on @huggingface 🤗 1/
180
364
208
3,097
354,345
Nole Doe (quant arc) retweeted
The Ballon d’Or Heist 🏆 Who’s taking it home this time?
1,169
10,734
2,986
83,661
10,136,394