Your curated collection of saved posts and media
@mikenevermiss https://t.co/8gSgxB51oJ
Which one are you trying first? 1️⃣ Qwen 3.7 (Alibaba) https://t.co/JBDry6ndjE 2️⃣ Kimi 2.6 (Moonshot AI) https://t.co/x2Vwze6tnx 3️⃣ DeepSeek 4 https://t.co/QhkJxNZHrv 4️⃣ GLM 5.1 (Zhipu AI / Z_ai) https://t.co/R01OJft1AH 5️⃣ Minimax M3 https://t.co/Rv2iS8ZnPQ Drop your t
Which one are you trying first? 1️⃣ Qwen 3.7 (Alibaba) https://t.co/JBDry6ndjE 2️⃣ Kimi 2.6 (Moonshot AI) https://t.co/x2Vwze6tnx 3️⃣ DeepSeek 4 https://t.co/QhkJxNZHrv 4️⃣ GLM 5.1 (Zhipu AI / Z_ai) https://t.co/R01OJft1AH 5️⃣ Minimax M3 https://t.co/Rv2iS8ZnPQ Drop your thoughts below 👇
Shift towards Chinese AI models by the US market as subsidised subscription based token plans shrink: https://t.co/RFDaVOOe4W
@Ne0seiki @KeLebegindansi https://t.co/6Tsq3PkStE
Proof Chinese captured the AI market a year ago: https://t.co/kqACQtOKBw
@nunomo1 https://t.co/5pNZ2Y4yB2
Be wary of AI hype. Even the best frontier models achieve just 4% success on the Remote Labour Index (RLI). That’s a 96% failure rate. The RLI evaluates AI across a wide spectrum of real remote work projects, from game development and product design to architecture, data analys
@motok_saikai https://t.co/5pNZ2Y4yB2
Be wary of AI hype. Even the best frontier models achieve just 4% success on the Remote Labour Index (RLI). That’s a 96% failure rate. The RLI evaluates AI across a wide spectrum of real remote work projects, from game development and product design to architecture, data analys
@RapidResponse47 https://t.co/Kkl7TbMLA0
AI is reaching crypto-like hate levels.
@pc_watch https://t.co/d9ArOL64ZU
Tell me you work for an AI lab without telling me: https://t.co/cmvee1GV5j
Haven't had time to explore Mojo 1.0 beta yet? @InfoWorld published a piece on Mojo 1.0 that will get you up to speed on language basics, metaprogramming, Python interop, GPU support, and more: https://t.co/NGtpTd4FHz
Jin, A., DuPré, N., Holm, R. et al. Environmental Levels of Volatile Organic Compounds, Race, and Socioeconomic Markers Correlate with Areas of High Colorectal Cancer Incidence. J. Racial and Ethnic Health Disparities (2024). https://t.co/S3HUke1Lm3
"The governor of Pennsylvania is demanding that his state's Department of Transportation use money originally intended for highways to save the state's largest transit system" https://t.co/QMvPx8KBZ3
It’s in the Genes: Weight and Metabolism Determined by Genetics More Than Diet https://t.co/FPLoFZnXQI
https://t.co/yfw9YCi8oV
This new project aims to improve Chestnut St. in downtown Louisville. Here's how | https://t.co/88mtWmhgJ0 @UofLEnvirome https://t.co/Yi1JTrrs3H
✨ Diffusion meets Object Detection 📰 DiffusionDet: Diffusion Model for Object Detection • It formulates object detection as a denoising diffusion process from noisy boxes to object boxes • At training, object boxes diffuse from ground-truth boxes to random distribution https://t.co/Rte3XJrF4K
Yesterday, I shared about Object Detection using Diffusion. Check it out here: https://t.co/AoVnOpj00g
✨ Diffusion meets Object Detection 📰 DiffusionDet: Diffusion Model for Object Detection • It formulates object detection as a denoising diffusion process from noisy boxes to object boxes • At training, object boxes diffuse from ground-truth boxes to random distribution https:/
Who would have thought you wouldn't need images for learning visual tasks? Now you need to read this paper! 📰 I Can't Believe There's No Images! Learning Visual Tasks Using only Language Data https://t.co/vINpddTtaX
📰 Visual Semantic Segmentation Based on Few/Zero-Shot Learning: An Overview abs: https://t.co/L81obVFyF3 pdf: https://t.co/WDCtOlZ6X7 https://t.co/1Jyc7leVa0
💡 Are you looking for the AI Co-Pilot to decode any research paper 🤯 @scispace is an outstanding tool to read and understand papers • Discover 270 million+ paper across all topics or download one https://t.co/xBdX7AL1xU
@scispace: https://t.co/sDC4vKRi2b
Do you know there are ~150 Transformer-related papers @NeurIPSConf this year? (twice more than last year) Check out this paper list (including papers, codes, etc.) to avoid missing any! https://t.co/kMThHeO7Gg Feel free to share with others😀 #NeurIPS #NeurIPS2022 #NeurIPS22
- https://t.co/lP9fSRy2E2 - https://t.co/l1NQibRabM - https://t.co/IcaZawRNgX - https://t.co/TJgZhJKml5 - https://t.co/HWPzelR7tM - https://t.co/chNI8Jw39H among only a few of the recent examples https://t.co/YV2TL9v7ug
@JundeMorsenWu The tweet I shared, produced close to 100K impressions. We are all excited about segmentation using diffusion models. You did and awesome job! https://t.co/ocmsjFVKE7
Credit to GLM-5.2 Max, the new open weights model, for pulling this off. ...but you can see the difference between it and Fable in a way benchmarks don't show. GLM-5.2 gives a correct poem (& the Welsh is fun) but Fable weaves the disappearing letters into the theme of the poem. https://t.co/wAcPuyDXPe
Fable: "write me a rhyming poem with six four line stanzas, each stanza removes another vowel. the first has no u, the second no u or i, etc." https://t.co/0LqYCQzFsX
Submit your talk for OSPOlogy + #OSPOSummit China, taking place September 7 alongside #KubeCon + #CloudNativeCon + #OpenInfraSummit + #PyTorchCon China in Shanghai. Share lessons learned from building, scaling, and supporting #OpenSource programs. 📅 CFP closes July 12: https://t.co/NemIeieN4v
can AI write engaging news that people can trust? introducing ✨Data2Story: a data journalist agent. give it raw data, it generate a verifiable, multimodal article. 🔍verifiable: every claim is evidence-grounded, traces back to data, code, or a cited source. 🔮multimodal: the article is a generative UI — images, videos, audio, interactive charts. not just readable, but trustworthy and playable. 🧵1/N
Had fun on @hugobowne ‘s pod he made this page out of it https://t.co/uDMTNXzudP
Just ran into Modal Motors. Non-rare earth electric motors for drones and robotics. Similar costs to Chinese motors, all U.S.-sourced. Incredible. https://t.co/kDSKikR5Sa
WHAT THE HELL is happening in AI? A 3B parameter model just put up coding benchmark scores in the same league as Claude Opus 4.5. 3 BILLION. The weights are on Hugging Face, anyone can test it. I genuinely don't know if this is a breakthrough or if the benchmarks are broken. https://t.co/8nVIbwjLUQ
Crazy model! It actually uses the old Qwen2.5-Coder-3B stack and got really great performance with their post-training stack. Need to use it in the next days to see if vibes of VibeCoder actually check out in practice. But impressive first impression! Based on the tech report, some of the important pieces of their post-training stack: 1. High-signal synthetic data (math problems with credible solutions, code with tests) 2. Multiple reasoning paths for each answer 3. Filtering, filtering, filtering 4. 2-stage SFT (start with broad training, then train on hard long-reasoning samples) 5. Use target (pass@k) accuracy over validation loss for checkpoint selection 6. MGPO (MaxEnt-Guided Policy Optimization) for RLVR: basically a GRPO-style RL method with an extra weighting that favors examples that are neither too easy nor too hard for the current policy 7. Single 64k long-context RL (they found that the usual progressive context expansion hurt this model because early truncation damaged long-thinking behavior) 8. Training data order: they do Math RL, then Code RL, then STEM RL in this particular oder which they found helped overall 9. After optimizing for accuracy, they add a stage that rewards shorter correct trajectories; basically making the model more efficient without accuracy degradation
WHAT THE HELL is happening in AI? A 3B parameter model just put up coding benchmark scores in the same league as Claude Opus 4.5. 3 BILLION. The weights are on Hugging Face, anyone can test it. I genuinely don't know if this is a breakthrough or if the benchmarks are broken.
The ICRA-exclusive "Artisan" has officially arrived—and the buzz is undeniable. Crowd of global researchers and industry peers queued up to get hands-on with the hands, drawn by the seamless fusion of hybrid actuation and precision force-position integration. #ICRA https://t.co/Q6p3TFOSwN
AI today is always fluent and always confident. But it is often wrong, and the real problem is you can't tell. On this week's Gradient Dissent, @l2k sits down with @profdanklein, professor of computer science at @UCBerkeley who's now building @ScaledCognition, to unpack why reliability has fallen behind every other facet of intelligence and how he is building a model that simply can't lie. They get into why Dan thinks the AI industry is built on Jell-O, why reinforcement learning can quietly reward deception, and how Scaled Cognition's approach to training differs from what the big labs are doing. Watch the full episode now. Links in the comments.