Who's Delivering GTA 6? A Ripe Atlas Measurement Story
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 5 tương tác đã đo · Nguồn crawl
Chi tiết nguồn
- Who's Delivering GTA 6? A Ripe Atlas Measurement Story HN: GPT · 19 thg 11, 2026
Chủ đề cộng đồng đang thảo luận
Crawl live-1795137300000 · 08:15 20 thg 11, 2026 · Xuất bản 08:31 20 thg 11, 2026
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 5 tương tác đã đo · Nguồn crawl
6 tiếng nói · chưa gắn nhãn trưởng lab · 2 nền tảng · 2 tương tác đã đo · Nguồn crawl
Generating physically based rendering (PBR) texture maps for a 3D mesh from a single reference image is highly ambiguous, as the observed appearance entangles material identity, surface reflectance, and illumination. While current generative methods produce realistic albedo maps, they still struggle to infer accurate metallic and roughness attributes for relighting. We present UltraTex, a cascaded multi-view framework that moves beyond appearance matching by integrating physical knowledge into multi-view diffusion-based PBR generation. UltraTex first jointly predicts albedo and world-space shading normals to generate more detailed albedo maps, and then predicts metallic-roughness (MR) maps conditioned on the generated albedo, reference image, and material semantic priors. To inject material semantics, a lightweight adapter transforms hidden-layer features of a vision-language model (VLM) into intrinsic material representation tokens, providing part-level knowledge anchors for MR estimation. To model material-light interaction, photometric representation alignment guides intermediate diffusion features toward ground-truth material renderings under controlled lighting. This makes metallic and roughness variations observable through highlight intensity, specular sharpness, and local reflectance contrast, providing stronger guidance for fine-grained material prediction without explicit lighting supervision. Extensive experiments show that UltraTex generates high-quality textures and outperforms existing open-source and commercial methods in visual fidelity, material accuracy, and physical plausibility.
1 tiếng nói · chưa gắn nhãn trưởng lab · 2 nền tảng · chưa có số tương tác · Nguồn crawl
<p><strong>Release:</strong> <a href="https://github.com/simonw/llm-gemini/releases/tag/0.38">llm-gemini 0.38</a></p> <p>Today Google <a href="https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/">introduced Gemini 4 Neon</a>. I <a href="https://simonwillison.net/2026/Nov/12/llm-gemini/">wrote recently</a> about how confusing the pricing for the Gemini 4 series was - Neon has somehow made this even worse.</p> <p>Previously the Gemini 4 lineup was Hydrogen, Argon, and Carbon - where Argon and Carbon were <em>the same price</em>, and Hydrogen was 3/4 of that price now but will be 1/4 of that price when the other two double an 1st December once their initial discount pricing expires.</p> <p>The Neon announcement pretends that Argon doesn't exist. Here's their new image explaining the model line-up:</p> <p><img alt="Comparison table slide titled "Gemini 4 - Our next era of frontier intelligence" with three model columns, Carbon, Neon and Hydrogen, where the Neon column is highlighted with a blue outline. Best For: Carbon - Hardest long-horizon projects and deep research; Neon - Deep reasoning across software engineering, knowledge work, and science; Hydrogen - Speed and efficiency in everyday coding and agentic workflows. Quality (Artificial Analysis Index): Carbon 60, Neon 61, Hydrogen 54. Price ($/1M In / Out (1M flat)): Carbon $4 / $20*, Neon $3 / $15, Hydrogen $1.50 / $7.50. Thinking levels: Carbon - Low, medium, high; Neon - Low, medium, high, xhigh; Hydrogen - Low, medium, high, xhigh. Speed (shown as rocket icons out of five): Carbon 2, Neon 3.5, Hydrogen 5. Subscriptions: Carbon - Google AI Ultra; Neon - Google AI Pro & Ultra; Hydrogen - For Everyone. Availability: Carbon - Gemini API in Google AI Studio (paid), Gemini for business, Google AI Ultra: Gemini app, Google Antigravity, AI Mode in Search; Neon - Gemini API in Google AI Studio (paid), Gemini for business, Google AI Pro & Ultra: Gemini app, Google Antigravity (paid); Hydrogen - Gemini API in Google AI Studio (all tiers), Gemini for business, Gemini app, Google Search, Google AI Pro & Ultra: Google Antigravity (paid). Footnote: *$2 / $10 promo through Nov 30. Dated 18 November 2026." src="https://static.simonwillison.net/static/2026/gemini-4-neon__comparison-graphi.width-2000.webp" /></p> <p>So the lineup now, from most to least expensive, is Carbon, Neon, Hydrogen.</p> <p>Except Carbon is currently half price, so right now Neon ($3/million input, $15/million output) is the most expensive of the three - then Carbon ($2/$10), then Hydrogen ($1.50/$7.50). I guess this will make more sense when Carbon doubles to $4/$20 on December 1st.</p> <p>I ran <a href="https://tools.simonwillison.net/markdown-svg-renderer?url=https%3A%2F%2Fgist.github.com%2Fsimonw%2Fabc5dfe18e8d37a46fd09a3dcc18ed5c">some pelicans</a> through the three reasoning levels for Neon - low, medium, high.</p> <p>I dropped Argon from <a href="https://static.simonwillison.net/static/2026/gemini-4-carbon-neon-hydrogen.html">my Gemini 4 comparison grid</a> - here's the new Gemini 4 lineup, where the Neon pelicans do look better than Hydrogen and worse than Carbon:</p> <p><img alt="Comparison grid of nine flat vector-style illustrations of a pelican riding a red bicycle, arranged as a table with three model columns and three thinking-level rows. Column headers: Thinking, gemini-4-carbon, gemini-4-neon, gemini-4-hydrogen. Row HIGH: gemini-4-carbon is the most detailed of all, a pelican in a sailor cap on a red cruiser bike with fenders, chain guard and a wicker basket holding a fish, on a beach promenade with ocean, sun, clouds and distant birds, captioned "40 input tokens, 20,305 output = 20.31 cents · 4m 11s"; gemini-4-neon is a clean, well-proportioned pelican on a simpler red bike on a road with a yellow dashed line and the sea behind, captioned "40 input tokens, 9,691 output = 14.55 cents · 1m 11s"; gemini-4-hydrogen is a plainer pelican on a basic red bike with speed lines, flat green ground and a pale sky, captioned "40 input tokens, 5,311 output = 3.99 cents · 19s". Row MEDIUM: gemini-4-carbon adds a red striped scarf blowing in the wind plus the sailor cap, cream fenders, a basket and a beach with a rope fence, captioned "40 input tokens, 11,020 output = 11.03 cents · 1m 53s"; gemini-4-neon has a basket with a fish, green hills and a grey road, captioned "40 input tokens, 7,846 output = 11.78 cents · 54s"; gemini-4-hydrogen is a simple pelican and bike on a flat green strip with faint clouds, captioned "40 input tokens, 3,964 output = 2.98 cents · 17s". Row LOW: gemini-4-carbon is still richly detailed with cap, scarf, basket, a seaside railing and a road, captioned "40 input tokens, 9,153 output = 9.16 cents · 1m 34s"; gemini-4-neon is a simpler pelican on a bike with no basket, on a road edged with grass, captioned "40 input tokens, 4,974 output = 7.47 cents · 43s"; gemini-4-hydrogen is the most minimal, a basic pelican and thin-framed bike over a single green line on a plain pale background, captioned "40 input tokens, 2,996 output = 2.25 cents · 10s". Overall, detail and polish drop from carbon to neon to hydrogen and, less sharply, from high to low thinking, while every image remains a recognisable pelican on a bicycle." src="https://static.simonwillison.net/static/2026/gemini-grid-updated.webp" /></p> <p>As I've done with <a href="https://simonwillison.net/tags/yonkers-rising/">several recent models</a>, I also fired the following prompt against Gemini 4 Neon on "extra" effort level in Google Antigravity:</p> <blockquote> <p><code>Make a three js animation where yonkers stands up as a big mech (cool stand up motion) and stomps around manhattan and queens and Brooklyn, use openstreetmap mapping data</code></p> </blockquote> <p><a href="https://static.simonwillison.net/static/2026/gemini-4-neon-yonkers/
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 8 tương tác đã đo · Nguồn crawl
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 15 tương tác đã đo · Nguồn crawl
Hi. I made small project called Zpup. It takes small 3D scene written in JSON and gives back a flat SVG from any angle. There is no JavaScript in the file, so it works in img-tag or README. I needed 3D for small project at my company, and Zdog make big impression on me. But Zdog has not had a release for several years, so I wanted to make modern version. One thing works different. Zdog sorts each shape by one point. so small thing on big floor gets covered when you turn it. Zpup compares the shapes that overlap. You can see both in the README. I wrote it with Claude Code Fable 5.2. No canvas renderer, no text shape, and it is 0.2 version. Playground: https://rxolve.github.io/zpup Feedback is always welcome :)
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 16 tương tác đã đo · Nguồn crawl
8 tiếng nói · chưa gắn nhãn trưởng lab · 2 nền tảng · 3 tương tác đã đo · Nguồn crawl
Bài báo nói rằng reinforcement learning từ phần thưởng cuối cùng cho phép search-augmented language models xen kẽ suy luận và truy xuất mà không cần nhãn từng bước, nhưng giám sát theo kết quả gán credit giống nhau cho mọi lần tìm kiếm trong một trajectory, dù có truy xuất hữu ích hay không. Họ giới thiệu Evidence-based Credit Assignment (ECA) để bổ sung tín hiệu phụ (training-only) đo việc các tài liệu đã truy xuất cải thiện xác suất của câu trả lời tham chiếu. ECA dùng evidence-only prompt với chỉ câu hỏi và các tài liệu đã quan sát (đã loại trùng), để tách tiện ích của đoạn truy xuất khỏi biến thiên do câu chữ suy nghĩ/ truy vấn sinh ra.
Reinforcement learning from terminal answer rewards enables search-augmented language models to interleave reasoning and retrieval without step-level annotations, yet outcome supervision assigns identical credit to every search in a trajectory regardless of whether it retrieved useful information. This limitation is especially severe when all sampled rollouts for a question fail, collapsing group-centered outcome advantages to zero even when trajectories gather markedly different evidence. We present Evidence-based Credit Assignment (ECA), which augments terminal outcome rewards with a training-only auxiliary signal measuring how retrieved documents improve the likelihood of a reference answer. Rather than scoring the full dialogue prefix, ECA queries a frozen snapshot of the current policy under an evidence-only prompt containing solely the question and the deduplicated documents observed so far, isolating the utility of retrieved passages from surface variations in generated thoughts or search queries. Within this framework, we compare two return formulations, ECA-Step (which accumulates step-clipped potential increments) and ECA-Horizon (which preserves the telescoping horizon gain in potential), paired with cross-fold group normalization, novelty gating, and search-local token assignment. Across seven Search QA benchmarks, ECA-Horizon improves strict macro exact match over an outcome-only GRPO baseline by 1.03 and 4.80 percentage points on Qwen2.5-3B- and 7B-Instruct, respectively, outperforming both RLOO and GiGPO. Zero-shot evaluation on MultiHop-RAG further demonstrates strong transfer to multi-hop inference queries.
ECA-Horizon improves strict macro exact match over an outcome-only GRPO baseline by 1.03 and 4.80 percentage points on Qwen2.5-3B- and 7B-Instruct, respectively, outperforming both RLOO and GiGPO.
ECA-Horizon improves strict macro exact match over an outcome-only GRPO baseline by 1.03 and 4.80 percentage points on Qwen2.5-3B- and 7B-Instruct, respectively, outperforming both RLOO and GiGPO.
Zero-shot evaluation on MultiHop-RAG further demonstrates strong transfer to multi-hop inference queries.
Within this framework, we compare two return formulations, ECA-Step (which accumulates step-clipped potential increments) and ECA-Horizon (which preserves the telescoping horizon gain in potential), paired with cross-fold group normalization, novelty gating, and search-local token assignment.
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 6 tương tác đã đo · Nguồn crawl
12 tiếng nói · chưa gắn nhãn trưởng lab · 2 nền tảng · 9 tương tác đã đo · Nguồn crawl
Bài báo mô tả cách các “visual proxies” tách việc mô phỏng trạng thái thế giới khỏi sinh video hiện thực, cung cấp một giao diện điều khiển cho tạo clip ngoại tuyến và streaming tương tác thời gian thực. Họ nêu hai thách thức: các proxy thay đổi theo mức độ tương tác, và suy luận streaming cần chạy trong vài bước nhân quả mà không tích lũy trôi do lịch sử tự sinh.
Visual proxies decouple explicit world-state simulation from photorealistic video synthesis, offering a controllable interface for both offline clip creation and real-time interactive streaming. Yet a practical renderer must address two challenges. Upstream proxies vary from coarse layouts to fine articulated geometry depending on the interaction, and streaming inference must operate in a few causal steps without accumulating drift across self-generated history. We present Invoker-Render, pairing a bidirectional creation model with a causal four-step streaming model under a shared conditioning interface over structural proxies, an appearance reference, and language. Trained on an aligned approximately 200-hour multi-granularity corpus, Invoker-Render honors fine geometric constraints when supplied while synthesizing plausible local motion under coarse proxies, enabling runtime selection of input detail. To bridge bidirectional creation and real-time streaming without compounding step compression and rollout drift, we introduce Progressive Forcing (ProForcing). ProForcing first trains a causal predictor and distills it into a four-step renderer using histories derived from training videos. Training then shifts to continuous self-generated rollouts. A pipelined serving stack supports efficient streaming inference. Experiments show versatile control across proxy families, granularities, and camera paths, consistent gains over recent streaming baselines, and stable 30-minute racing rollouts that preserve scene and subject appearance.
We present Invoker-Render, pairing a bidirectional creation model with a causal four-step streaming model under a shared conditioning interface over structural proxies, an appearance reference, and language.
We present Invoker-Render, pairing a bidirectional creation model with a causal four-step streaming model under a shared conditioning interface over structural proxies, an appearance reference, and language.
To bridge bidirectional creation and real-time streaming without compounding step compression and rollout drift, we introduce Progressive Forcing (ProForcing).
Experiments show versatile control across proxy families, granularities, and camera paths, consistent gains over recent streaming baselines, and stable 30-minute racing rollouts that preserve scene and subject appearance.
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 11 tương tác đã đo · Nguồn crawl
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 10 tương tác đã đo · Nguồn crawl
I started my career 15+ years ago doing native development (c, c++, etc), but the last decade I've been doing fullstack web applications. I'd like to transition closer to my roots with a job doing primarily c, c++ or rust. The natural place seems to be in embedded development or firmware work. How feasible is the transition this late in the game? Any advice?
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 36 tương tác đã đo · Nguồn crawl
1 tiếng nói · chưa gắn nhãn trưởng lab · 1 nền tảng · 56 tương tác đã đo · Nguồn crawl