AllyshipContent Radar

News 2897

Chủ đề cộng đồng đang thảo luận

Crawl live-1795126500000 · 05:15 20 thg 11, 2026 · Xuất bản 05:31 20 thg 11, 2026

195 / 2897 chủ đề đạt ngưỡng 55 trong khoảng này.

143

And if you're interested on how it was built we have a showcase up! https://developers.openai.com/showcase/notre-damn

59

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

dimillian, kagigz 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

And if you're interested on how it was built we have a showcase up! https://developers.openai.com/showcase/notre-damn It actually started as a survivor shooter in a cathedral but the map was so good I turned it into what it is now!

147

Announcing harness adapters for @Alibaba_Qwen Qwen3.8-27B on Pi harness!

58

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

yb2698, clementdelangue 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

Announcing harness adapters for @Alibaba_Qwen Qwen3.8-27B on Pi harness! We're releasing a LoRA adapter that makes Qwen3.8-27B a much better and efficient agent behind Pi harness. SWE-Bench Pro (Hard 51 subset) pass@1: 29% -> 63% Terminal-Bench 2.1 pass@1: 40% -> 56% Much more efficient: 55-62% fewer tool calls per task 🤯 https://huggingface.co/harness-hub/Qwen3.8-27B-pi-adapted The era of harness-specific adapters is now on! More harness adapters to come in the next days 🧵

148

Announcing harness adapters for @Alibaba_Qwen Qwen3.8-27B on @opencode harness!

58

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

yb2698, lysandrejik 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

Announcing harness adapters for @Alibaba_Qwen Qwen3.8-27B on @opencode harness! After yesterday's release of the Pi adapter, we now released to the community the OpenCode version, trained on the v2 version, ready to use for local inference. SWE-Bench Pro (Hard 51 subset) pass@1: 35.3% -> 51.0% | xhigh thinking mode 47.1% -> 54.9% Terminal Bench 2.1 pass@1: 47.0 -> 53.0 | xhigh thinking mode 55.4 -> 56.6 Generated tokens (content + reas content + tool calls): 6-8x fewer tokens per task ! https://huggingface.co/harness-hub/Qwen3.8-27B-opencode-adapted 🧵

149

A fascinating read on how Codex happened.

58

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

nirmanhere, danshipper 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

A fascinating read on how Codex happened. A lot of interesting insights into building AI products, the engineering decisions, and what it takes to turn an idea into something real. Definitely worth a read. https://every.to/p/how-codex-happened

151

Claude Gets Hella Paranoid

59

1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 30 tương tác đã đo · Nguồn crawl

noduerme 1 bài viết · HN: LLM
Nguyên văn nguồn
Chi tiết nguồn

Mm, so I hesitate to write this, because my own AI agents might read it. But what the hell. I've got this crappy little $300 PC on my network hooked up to my projector, that I mostly use to serve Jellyfin. So tired of it going down every time there's a Windows update, but I've been too lazy to get around to installing Ubuntu. Last night, when it went down, I asked my Claude-claw via Whatsapp if it could just SSH into it and set up a dual boot partition. Claude got as far as partitioning the drive and then starting asking hard questions , like (paraphrasing) "wait a minute. This crosses a line. I have now been asked to remote into a machine with a password, write a separate boot partition, and restart it. From this chat, you haven't really verified that you are you. If you're really there, and not just texting me, give me some proof". So I remoted in and created a TextEdit RTF file on the claw's desktop: "Hey, it's me, I'm asking you to remote into the PC". Claude read it and responded that it was the kind of proof it was looking for, real RTF headers and datestamps and such. I had proven I had physical access to the claw machine, (although not to the PC, but it seemed reassured anyway). A little while later, before triggering the Ubuntu installer, it put me through it all over again. I told Claude I was very thankful it was guarding my network so well, and I would go over and write a file on the target PC manually too. I'm not gonna lie: It was very fucking cool that it was able to do this without me ever needing to flash a USB drive, and by the time it was done it had set up Ubuntu, Tailscale and Jellyfin, and resolved the network issues; when I started watching a movie, the sound didn't work, so I prompted it and suddenly the sound came on as it wired up the HDMI adapter. Almost creepy. But I was pretty stunned at how reticent it was to do something relatively simple on my own network, with access I gave it. I think there must be a serious push right now to lock it down from monkeying with anything remotely, especially a BIOS.

152

So Was Jev Useful?

56

1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 9 tương tác đã đo · Nguồn crawl

nlpnerd 1 bài viết · HN: LLM
Nguyên văn nguồn
Chi tiết nguồn
153

Try ParseBench here!: https://www.parsebench.ai/lab/index.html?utm_source=socials&utm_medium=tw_channel&utm_campaign=metrics_lab

57

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

LlamaIndex, hexapode 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

Try ParseBench here!: https://www.parsebench.ai/lab/index.html?utm_source=socials&utm_medium=tw_channel&utm_campaign=metrics_lab

155

我把 2023 到 2026 年不同格式、中英文混杂的 6 份合成体检报告丢进一个开源项目,不同单位自动换算、连成一条趋势,每个数字还能点回原文件。

57

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

tammybuilds, rainandust 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

我把 2023 到 2026 年不同格式、中英文混杂的 6 份合成体检报告丢进一个开源项目,不同单位自动换算、连成一条趋势,每个数字还能点回原文件。 这才是健康 AI 最该先解决的事。开源地址:https://github.com/thetahealth/mirobody https://x.com/i/article/2121619241010962432

156

I'm not worried about OpenAI. GPT-6.2 Astra and Sol may not be fun to work with, but as others have pointed out, we said the same thing about Opus 5 and now look where we are. But like Anthropic a few

56

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

kplikethebird, danshipper 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

I'm not worried about OpenAI. GPT-6.2 Astra and Sol may not be fun to work with, but as others have pointed out, we said the same thing about Opus 5 and now look where we are. But like Anthropic a few months ago, I think OpenAI has some course-correcting to do. It seems pretty clear to me that OpenAI wants to build models for the masses. That appears to be the intent behind their (in my opinion, miscalibrated) announcement video. My concern is that what these models are like reflects their understanding of what "the masses" are looking for in a model: fast, efficient, but with minimum viable effort. I think most people aspire to do more in their work lives than sneak an assignment through at the last possible moment. I think they want to do great work that they are, yes, secure in, but also proud of and energized by. The best model releases are the ones that help you discover that feeling, and make you excited to wake up every day and see what you can do. I'm sorry to say there wasn't much of that energy during the time I tested GPT-6.2 Astra and Sol. But I look forward to the next Vibe Check. In the. meantime, here's this one: https://every.to/vibe-check/vibe-check-gpt-6-2-astra-and-sol-leave-us-wanting-more?utm_cta_source=home_main_a_3

157

My top 5 takeaways from @nateparrott on how to get Claude to design without AI slop:

56

2 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · chưa có số tương tác · Nguồn crawl

petergyang, ashleyjxie 0 bài viết · X
Nguyên văn nguồn
Chi tiết nguồn

My top 5 takeaways from @nateparrott on how to get Claude to design without AI slop: 1. Give Claude context upfront. Share screenshots, pages you like, and your past work. For an existing product, have Claude rebuild the real screens from code. 2. Research familiar UX patterns. For hard layout problems, ask Claude to research familiar products that solve something similar and note the layouts people already know how to use. 3. Make Claude aware of its design biases. "The model has no idea what its own biases are." In your design.md or another file, consider discouraging Claude's tells like serif text, off-white backgrounds, and card-edge bars. 4. Start with wireframes. Have Claude create wireframes first and give it feedback to nail your core user flows before going deep on fonts and colors. 5. Ask for multiple options, then remix them. Ask Claude for 5 options for key screens with different fonts, colors, and styles. Then pick the parts you like and ask it to remix and create 5 more options. 📌 Watch the full episode now: https://www.youtube.com/watch?v=604T-74s3KE Also available as a written post: https://creatoreconomy.so/p/how-to-fix-ai-slop-with-claude-design-nate-parrott

158

Game2Player: Scaling Gameplay Experience for Visual Agents through Environment-Player Co-Development

73

14 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 19 tương tác đã đo · Nguồn crawl

Xirui Jiang, Yuxi Chen và 12 nguồn khác 1 bài viết · Hugging Face Papers
Nguyên văn nguồn
Chi tiết nguồn

Games are a rich testbed for visual agents, demanding perception, long-horizon planning, and precise control. However, collecting human gameplay is costly, and learning from gameplay videos requires inferring missing actions. Existing datasets also cover a limited range of games, often without explicit tasks or verifiable outcomes. To address these gaps, we introduce Game2Player, a framework that co-develops game environments and players to collect diverse and verifiable gameplay experience at scale. In the Game2Player loop, a Builder agent adapts existing web games, exposes their internal state, and defines tasks with state-based progress measures and success checks. A Coding Player and a VLM Player then attempt the tasks and report problems to the Builder for repair, so that play serves as both environment validation and training data collection. With Game2Player, we build 568 verified web games with 3,436 tasks, collect 15.7K trajectories (307 hours of gameplay), and hold out 50 games as the benchmark PlayerBench. To evaluate the collected experience as training data, we train VLM agents with progress-weighted SFT, assigning greater weight to steps that advance task progress. Game2Player-9B improves progress from 14.3% to 31.2% on PlayerBench and from 30.0% to 53.1% on the external GameWorld benchmark. More broadly, our work suggests a path for visual agents to scale their gameplay experience by building, verifying, and learning from existing game worlds.

159

WorldRuler: Interactive World Modeling under Configurable Rules

71

9 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 9 tương tác đã đo · Nguồn crawl

Mingliang Zhai, Chuanhao Li và 7 nguồn khác 1 bài viết · Hugging Face Papers
Nguyên văn nguồn
Chi tiết nguồn

Rules govern how player actions change interactive worlds. Different rules create distinct gameplay experiences within similar scenes and action spaces. For customizable gameplay, generative world models must follow user-defined rules beyond real-world behaviors. However, realizing user-defined rules presents two challenges: Rule-State Persistence and Rule-to-Visual Realization. The former requires preserving rule definitions, dependencies, and world states across saving and loading, so that restored worlds follow the established rules. The latter requires accurately realizing rule-induced changes on designated objects and attributes, particularly when affected objects are unseen. In this paper, we propose WorldRuler, which organizes rule definitions, dependencies, and object states in an executable 3D rule graph and connects rule execution, video generation, and perceptual feedback in a closed loop. The graph supports joint saving and restoration of rules and world states, while dynamic object binding grounds rule-specified entities in visual instances and uses stored states to constrain generation from their first appearance. We further introduce object attribute state replacement and counterfactual reconstruction training, enabling accurate attribute control and compatible appearance and motion changes while limiting interference with unrelated content. We develop an automated data collection and annotation pipeline using GTA-V and Habitat to construct RuleVerse, a dataset of rule-driven interactions with detailed state and event annotations, including cases where rules take effect before affected objects become visible. Extensive experiments demonstrate the effectiveness of WorldRuler in preserving rule-state consistency across saving and loading and accurately realizing rule outcomes on designated objects and attributes.

160

Mechanistic Critics for Sample-Efficient NPU Design-Space Exploration

70

5 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 7 tương tác đã đo · Nguồn crawl

Yicheng He, Yuqi Xue và 3 nguồn khác 1 bài viết · Hugging Face Papers
Nguyên văn nguồn
Chi tiết nguồn

NPU design-space exploration for LLM inference mainly relies on an expensive simulator that often returns only scalar feedback such as latency and power. Such feedback can rank evaluated configurations, but gives little information for deciding why a configuration is slow or which architecture parameter should be changed next. This limitation is especially severe because prefill and decode workloads can be limited by different resources, including compute throughput and HBM bandwidth. We propose CriticNPU, a sample-efficient DSE framework built around an online-calibrated mechanistic critic for NPU architecture design. The critic uses explicit hardware limits to evaluate NPU configurations: before simulation, it estimates whether a proposed configuration is likely to improve over the current best and filters weak candidates; after simulation, it aggregates operator-level runtimes into prefill/decode bottleneck summaries and predicts the effect of candidate parameter changes. We instantiate the critic with a Roofline-style model that derives compute and memory limits from architecture parameters and calibrates per-operator-type effective limits using simulator observations. And CriticNPU combines this critic with a workload extractor and an architecture planner to propose configurations. Across six LLM workloads and 16 deployment configurations per model, CriticNPU achieves best-or-tied-best latency in 74% of prefill and 87.5% of decode settings while using roughly half the simulator calls of competitive DSE baselines. We conduct extensive ablation studies and detailed analyses to validate the effectiveness of our method from multiple perspectives. Our code is available at https://github.com/bruno686/CriticNPU.

141–160 / 195
Trang 8 / 10