03-29 11:26 · 财经,财经
【清明假期雨不停 南方未来十天强降雨和强对流接连登场】今年来首场大范围强对流天气正在激烈上演中。今天下午，广西、广东上空出现大片红紫回波，多地发布冰雹、雷雨大风预警。清明小长假期间（4月4日到6日），南方仍有大范围降水、强对流天气，局地暴雨、大暴雨。假期出行务必关注临近天气，做好防护。
今年来首场大范围强对流天气正在激烈上演中。今天下午，广西、广东上空出现大片红紫回波，多地发布冰雹、雷雨大风预警。
像广州，今天15～16时1小时降雨量54.8毫米，堪称“高压水枪式的暴击”。24小时雨量超50毫米即为暴雨，广州这相当于1小时就下完了暴雨一天的量。
这还只是开始。今天夜间到下周，南方多雨局面还将持续，将先后有三轮（3月29日到31日，4月2日到4日以及5日到7日）强降雨和强对流过程接连登场。
其中清明小长假期间（4月4日到6日），南方仍有大范围降水、强对流天气，局地暴雨、大暴雨。假期出行务必关注临近天气，做好防护。
（文章来源：央视新闻客户端）
▸ 展开全文
03-29 11:26 · 财经,财经
【美以伊冲突进入第30天】今天（3月29日），美以伊战事进入第30天。从3月28日至今，伊朗首都德黑兰等地持续遭到空袭。
美以伊冲突进入第30天
今天（3月29日），美以伊战事进入第30天。从2月28日至今，伊朗首都德黑兰等地持续遭到空袭。近期美以多次将伊朗的大学作为打击目标，伊朗伊斯兰革命卫队早前发布公告称，以色列以及美国在该地区的大学都将成为伊朗的合法打击目标。此外，28日晚，伊朗西南部胡齐斯坦省一处容量达一万立方米的储水系统遭到美以袭击，袭击没有造成人员伤亡，当地饮用水供应正常。就在28日，也门胡塞武装正式宣布介入战争，并向以色列发射导弹，进一步增加了地区的紧张局势。
（文章来源：CCTV国际时讯）
▸ 展开全文
03-29 11:26 · 财经,财经
【本周A股融资资金呈现净偿还态势】本周（3月23日至27日），A股融资余额高位震荡，最新为25987.6亿元（截至3月26日）。据证券时报·数据宝统计，A股融资资金呈现净偿还态势，金额合计达197.04亿元。
本周A股融资资金呈现净偿还态势
本周（3月23日至27日），A股融资余额高位震荡，最新为25987.6亿元（截至3月26日）。据证券时报·数据宝统计，A股融资资金呈现净偿还态势，金额合计达197.04亿元。
（文章来源：人民财讯）
本周（3月23日至27日），A股融资余额高位震荡，最新为25987.6亿元（截至3月26日）。据证券时报·数据宝统计，A股融资资金呈现净偿还态势，金额合计达197.04亿元。
（文章来源：人民财讯）
▸ 展开全文
03-29 11:26 · 科技,GitHub
Agent Trending是一个聚焦 AI Agent 的开源研究项目，系统整理了 2025.01—2026.03 期间 8701 篇学术论文，围绕 15 个核心方向完成分类、去重、标注与趋势分析。项目包含月度论文索引、细分方向论文集、15 份技术趋势报告，以及 1 份全局趋势总报告和 1 份下一代 Agent 原型设计文档，适合研究者、工程团队和产业决策者快速把握 AI Agent 的技术演进与落地脉络。
Agent Trending 是一个面向 AI Agent 领域的大规模论文研究与趋势分析开源项目。
适合谁读
基于 15 个细分方向、8701 篇学术论文的跨领域深度分析报告。
过去 15 个月（2025.01—2026.03），AI Agent 领域经历了一次历史性的范畴跃迁，从"更聪明的聊天机器人"变成了"能在真实世界执行复杂任务的自主系统"。我们系统性地追踪了这一过程中 7700+ 篇学术论文的演进脉络，覆盖了从基础训练、认知架构、工具使用、多智能体协作到安全对齐、商业落地的完整技术栈，最终凝练为这份全景式趋势报告。
Agent 领域的论文月产出在一年间从约 200 篇飙升至 700+ 篇，信息过载让从业者很难建立全局认知。我们希望通过这份开源报告，帮助研究者快速定位高价值研究方向，帮助工程团队做出更有依据的技术选型，帮助商业决策者理解真实的技术成熟度而非营销叙事。
报告识别出 八大跨领域共振趋势：
- 强化学习全面取代监督微调，成为 Agent 能力获取的核心引擎
- 分层闭环架构成为系统设计的事实标准
- MCP / A2A 等通信协议催生"Agent 互联网"基础设施雏形
- 安全性从附加特性升级为系统生存前提
- 记忆与认知能力成为 Agent 从工具进化为助手的分水岭
- 软件工程成为全领域的首要验证场景
- 评测范式从结果打分转向过程审计
- 物理世界与数字世界 Agent 加速融合
报告同时给出了商业化冲击波排序、五大共性瓶颈分析，以及面向研究者、技术团队和投资者的差异化战略建议。
下表汇总了 15 个细分方向的论文集、技术趋势报告，以及从各方向执行摘要中提炼出的核心趋势、面临挑战与未来推演。
下表按月份汇总论文收录情况，并给出对应阶段的一句话趋势演化总结。
本仓库主要基于公开可获取的学术论文信息、公开链接以及作者的研究整理与分析撰写而成。
- 论文标题、摘要信息、链接等原始内容的版权归原作者及原出版方所有
- 本仓库中的综述、归纳、结构化整理与趋势判断为仓库维护者的原创工作
- 如你认为仓库中存在不适合公开展示的内容，请通过 Issue 或邮件联系处理
当前版本以资料整理和结构化发布为主，后续可能继续补充：
- 新月份追踪
- 方向报告迭代
- 导航页和索引页
- 英文化摘要
欢迎通过 Issue 或 Pull Request 参与改进，例如：
- 修正文档错误
- 补充遗漏论文
- 优化目录和导航
- 补充引用与来源说明
具体约定见 CONTRIBUTING.md。
如果这个仓库对你的研究或写作有帮助，欢迎在引用或转载时注明仓库地址与原作者信息。建议引用仓库中的具体文件，而不是仅引用仓库名。
本仓库采用 CC BY 4.0 许可协议。你可以转载、改编和分享，但需要保留署名。
▸ 展开全文
03-29 11:26 · 科技,GitHub
-
Research trending topics and generate ready-to-post social media content — as a Hermes/OpenClaw skill.
trendingcontent is a two-phase AI agent skill:
- Research phase — scrapes Reddit, X/Twitter, YouTube, TikTok, HackerNews, Bluesky and Brave Search to surface the most engaging content about your topic over a configurable time window (7, 15, 30, or up to 90 days).
- Content generation phase — uses the research findings to generate platform-specific social media copy: Twitter/X threads, LinkedIn posts, and Instagram captions.
Built on top of last30days by @mvanhorn, extended with:
- Variable time windows (not just 30 days)
- Multi-platform content generation (Twitter, LinkedIn, Instagram)
- Tone control (professional, casual, educational, viral)
- Multilingual output (English, Spanish, Portuguese)
- Hermes/OpenClaw skill standard compatibility
git clone https://github.com/gabogabucho/trendingcontent-agent.git
cd trendingcontent-agent
pip install -r requirements.txt
export SCRAPECREATORS_API_KEY="your_key" # required
export BRAVE_API_KEY="your_key" # optional but recommended
Or add them to your .env
file.
Get a ScrapeCreators API key at scrapecreators.com. Get a Brave Search API key at brave.com/search/api.
python scripts/trendingcontent.py "artificial intelligence" --days=7 --platform=all
python scripts/trendingcontent.py <topic> [options]
Options:
--days=N Time window: 7-90 days (default: 30)
--platform twitter | linkedin | instagram | all (default: all)
--tone professional | casual | educational | viral (default: professional)
--lang en | es | pt (default: en)
--sources=S1,S2 Only use these sources (reddit,twitter,youtube,tiktok,bluesky,hackernews,brave)
--disable=S1,S2 Disable specific sources
--list-sources Show all available sources
--quick Fewer sources, faster
--deep More sources, thorough
--research-only Output research only, skip content generation
# AI trends last 7 days → all platforms
python scripts/trendingcontent.py "artificial intelligence" --days=7
# Web3 → LinkedIn post in Spanish, professional tone
python scripts/trendingcontent.py "web3" --days=15 --platform=linkedin --lang=es
# Climate tech → viral Twitter thread
python scripts/trendingcontent.py "climate tech" --days=30 --platform=twitter --tone=viral
# Niche topic — only Reddit and YouTube (no HackerNews)
python scripts/trendingcontent.py "sourdough bread" --sources=reddit,youtube,brave
# Disable sources that don't fit your topic
python scripts/trendingcontent.py "fintech" --disable=hackernews,tiktok
# See all available sources
python scripts/trendingcontent.py --list-sources
# Research only, no content
python scripts/trendingcontent.py "LLMs" --days=14 --research-only
This repository follows the Hermes skill standard. To install as a skill:
# Copy to your Hermes skills directory
cp -r . /path/to/hermes-agent/skills/social-media/trendingcontent/
# Add API keys to your Hermes .env
echo "SCRAPECREATORS_API_KEY=your_key" >> ~/.hermes/.env
echo "BRAVE_API_KEY=your_key" >> ~/.hermes/.env
The agent will automatically discover the skill via SKILL.md
.
═══════════════════════════════════════════════
TRENDINGCONTENT — CONTENT GENERATION BRIEF
═══════════════════════════════════════════════
Topic: artificial intelligence
Period: Last 7 days
Language: English
Tone: Professional
Platforms: Twitter, LinkedIn, Instagram
───────────────────────────────────────────────
RESEARCH SUMMARY
───────────────────────────────────────────────
[ranked trending content with titles, scores, URLs]
───────────────────────────────────────────────
CONTENT INSTRUCTIONS
───────────────────────────────────────────────
[platform-specific format specs + tone guidance]
- Python 3.9+
yt-dlp
— YouTube transcript extractionrequests
— HTTP calls- ScrapeCreators API key (required)
- Brave Search API key (optional)
- Original research engine: last30days-skill by @mvanhorn — MIT License
- Social media generation layer & Hermes skill packaging: @gabogabucho
MIT License — see LICENSE for details.
▸ 展开全文
03-29 11:26 · 科技,GitHub
Kakobuy Spreadsheet features 3,000+ trending products from Weidian, Taobao, and 1688, with affordable new arrivals added daily. Explore the latest finds from the past 7 days, plus shoes and clothing on MaisonLooks with QC photos and AI try-on. Order easily through your preferred agent.
All-in-One Spreadsheet for Kakobuy, ACbuy, Litbuy, OOPbuy, Sugargoo & Superbuy. Features 3000+ trending Chinese products from Weidian, Taobao & 1688 with daily updates. Browse premium finds on MaisonLooks with QC photos and AI try-on. Order easily via your preferred agent for the best shopping experience.
Discover over 3,000 trending Chinese products on our Master Spreadsheet, fully compatible with Kakobuy, ACbuy, Litbuy, OOPbuy, Sugargoo, and Superbuy. We feature daily affordable arrivals from Weidian, Taobao, and 1688, maintained with updates every 24 hours.
- Premium Selection: Browse the best finds and belts curated for all agents.
- Smart Shopping: Access high-quality QC photos and AI try-on technology.
- Daily Updates: See the latest additions from the past 7 days.
▸ 展开全文
03-29 11:26 · 科技,GitHub
Real-time monitoring & analytics platform for Pump.fun tokens. Track launches, analyze trends, and get instant alerts — powered by WebSockets. Built with HTML, CSS & JavaScript.
Real-time monitoring & analytics platform for Pump.fun tokens on Solana.
🌐 Live site: pump-analyzer.com
PumpAnalyzer is a landing page for a token analytics platform built on top of Pump.fun. It lets traders track new token launches, analyze price movements and volume, and receive instant alerts — all in real time.
- ⚡ Real-Time Monitoring — Track every new token launch on Pump.fun as it happens
- 📊 Deep Analytics — Price charts, volume analysis, and holder distribution
- 🔔 Instant Alerts — Get notified the moment tokens hit your custom criteria
- 🤖 AI-Powered Insights — Smart detection of pump patterns and emerging trends
- 🔒 Secure & Non-Custodial — No wallet access required. Your keys, your crypto
- 🚀 Lightning Fast — Sub-second data updates powered by WebSockets
git clone https://github.com/keron-2000fz47/special-octo-parakeet.git
cd special-octo-parakeet
open index.html
No build tools or dependencies required — it's pure HTML/CSS/JS.
- 🌐 Website: pump-analyzer.com
- 🐦 Twitter: coming soon
Not financial advice. Always DYOR.
▸ 展开全文
03-29 11:25 · AI,技术,HackerNews,AI
What if AI doesn't need more RAM but better math?
@adlrocha - What if AI doesn’t need more RAM but better math?
How TurboQuant compresses the KV cache without losing accuracy, and what that could mean for memory stocks
Last week I was writing about the hardware side of the AI memory problem: the HBM density penalty, the EUV bottleneck, and the supply chain pressure squeezing DRAM prices for everyone from data centre operators down to consumer electronics. This week, Google published something that attacks the exact same problem using another approach: not “build more memory”, but “need less of it.”
You guessed it! This post will dive a bit deeper into what TurboQuant is, and what this may imply to the field of AI. What Pied Piper achieved in the Silicon Valley TV Show with their general-purpose lossless compression algorithm, Google may have achieved it for the compression of information represented as vectors in a high-dimensional space.
What is a transformer? And the KV cache?
But before getting into what TurboQuant does, let’s make a brief detour to understand what is this algorithm is actually built to compress, and why it is important for LLMs and the memory problem.
GPT models are what are known as autoregressive: they generate text one token at a time, where each new token is conditioned on everything that came before. You send a prompt, the model reads all of it, picks the most likely next word, appends it, reads everything again, picks the next word, and so on. One token at a time, left to right, until it decides to stop.
The core mechanism that lets the model read everything at each step is called attention. For every token in the sequence, the model computes three vectors: a query, a key, and a value. You can think of these data structures as a bit more complex key-value stores. To generate the next token, the model compares the current query against every previous key, essentially asking “which past tokens are relevant right now?”, and uses the answer to weigh the corresponding values and build up context.
This is implemented (as you may all know by now) through the transformer architecture. Transformer layers are responsible for encoding the input sequences into a meaningful representation, applying the attention mechanism, and decoding into an output representation. All LLMs are architectural variations of this basic cell.
To get a sense of each of these variations I highly recommend Sebastian Raschka’s LLM Architecture gallery: from GPT-2 to DeepSeek and GLM.
The keys and values for every previous token are recomputed from scratch on every single pass through architecture. If your conversation is N tokens long and you’re generating token N+1, the model recalculates N sets of keys and values it already calculated on the previous step. This is slow and wasteful in terms of the resources.
The obvious fix to this is to cache them. The query, key and values are computed once per token and stored so they can be looked up in subsequent steps instead of being recalculated. This is the KV cache, a running store of QKV tokens from all previous tokens stored in GPU memory (so they are readily accessible when needed).
The problem is that the KV cache grows with every token. With short messages this is trivial as all tokens fit in memory, but a long conversation, or a full code base, involves hundreds of thousands of tokens. Each token has its own key and value vectors, across every attention layer in the model, each stored as a full-precision floating-point number (as long as there’s no quantisation involved). For a model like Llama 3.1 70B, the KV cache for a single long context can consume more GPU memory than the model weights themselves.
This is one of the key bottlenecks in production inference. Serve more users simultaneously? More KV cache. Support longer contexts? More KV cache. Run cheaper inference? Figure out what to do about the KV cache. We are trading the compute necessary to compute on-the-fly the QKV values, for increased memory requirements.
By using quantisation instead of storing each value at 32-bit or 16-bit precision, one can round it down to 4 bits or 3 bits (or even 2 bits, like Microsoft recently showed). Some accuracy is lost in the approximation, but if it is not significant for the user case, the trade-off is obviously worth it. The question is how to do this well. Standard quantisation techniques add 1-2 extra bits of overhead per value as metadata, which partially undermines the compression you’re trying to achieve. Getting to genuinely low bit-widths without that overhead, and without accuracy degradation, is the hard part. HuggingFace has a really nice page with an overview of quantisation and a list of methods
Enter TurboQuant
But things may be about to change. Google announced this week TurboQuant. TurboQuant (see paper) is a two-stage algorithm. The two stages have different jobs.
Stage 1: PolarQuant. This is the main compression step. We currently store vectors using Cartesian coordinates as distances of a base to the
▸ 展开全文
03-29 11:25 · AI,技术,HackerNews,AI
Show HN: Public transit systems as data – lines, stations, railcars, and history
Public Transit Systems
Comprehensive information about public transit systems worldwide. Explore stations, lines, railcars, and historical data through a terminal-inspired interface.
$ Global Network Statistics
Total Stations
1,570
Total Lines
89
Track Length
1,506mi
Systems
9
> Available Systems
9 systemsLight RailLink
Baltimore, MD
The Light RailLink is a light rail system operated by the Maryland Transit Administration serving the Baltimore metropolitan area. Running 30 miles from Hunt Valley in northern Baltimore County through downtown Baltimore to Cromwell Station/Glen Burnie and BWI Airport in Anne Arundel County, it is one of the oldest modern light rail systems in the eastern United States. The system uses former railroad rights-of-way for much of its route and runs as a street-running railway through downtown on Howard Street.
Stations
33
Lines
3
Track Length
30mi
Daily Riders
14,500
Metro SubwayLink
Baltimore, MD
The Metro SubwayLink is a heavy rail rapid transit line operated by the Maryland Transit Administration serving the Baltimore metropolitan area. Running 15.4 miles from Owings Mills in Baltimore County through downtown Baltimore to Johns Hopkins Hospital in East Baltimore, it is the only rapid transit line in Maryland. The system features both surface and deep-bore tunnel sections, with six underground stations in the downtown core.
Stations
14
Lines
1
Track Length
15mi
Daily Riders
15,200
BART
San Francisco Bay Area, CA
BART is a heavy rail rapid transit system serving the San Francisco Bay Area in California. It connects San Francisco, Oakland, Berkeley, and surrounding East Bay cities with lines extending to San Francisco International Airport and the Silicon Valley.
Stations
50
Lines
6
Track Length
131mi
Daily Riders
323,000
Beijing Metro
Beijing
The Beijing Subway is a rapid transit system serving Beijing, China. It is the world's busiest and one of the longest metro systems, having surpassed Shanghai in total length in 2023. Originally opened in 1969 as China's first subway, it has undergone massive expansion, particularly after Beijing won the 2008 Olympics bid.
Stations
527
Lines
28
Track Length
551mi
Daily Riders
10.5 million
CTA
Chicago, Illinois
The Chicago 'L' is the rapid transit system serving the city of Chicago and some of its surrounding suburbs. Operated by the CTA, it is the fourth-largest rapid transit system in the United States by ridership. The system's name comes from the elevated railways that make up much of the network, though portions also run in subway tunnels, at grade, and in highway medians.
Stations
146
Lines
8
Track Length
224mi
Daily Riders
594,000
NYC Subway
New York City
The New York City Subway is the largest rapid transit system in the world by number of stations, operating 472 stations across four boroughs of New York City. Opened in 1904 and serving approximately 3.6 million riders daily, it is one of the few transit systems in the world that operates 24 hours a day, 7 days a week, forming the backbone of transportation in the most densely populated major city in the United States.
Stations
472
Lines
25
Track Length
245mi
Daily Riders
3.6 million
Sound Transit
Seattle, WA
Sound Transit operates regional express bus, light rail, and commuter rail services in the central Puget Sound region of Washington state. The Link light rail system connects Seattle, Bellevue, and surrounding communities.
Stations
50
Lines
3
Track Length
59mi
Daily Riders
142,000
Tokyo Metro
Tokyo
Tokyo Metro is one of two rapid transit operators in Tokyo, Japan (the other being Toei Subway). Originally opened in 1927 as Asia's first subway, it was privatized in 2004 from the Teito Rapid Transit Authority (TRTA/Eidan). Tokyo Metro is the busiest metro system in the world by annual ridership.
Stations
180
Lines
9
Track Length
121mi
Daily Riders
6.84 million
WMATA
Washington, D.C.
The Washington Metro, operated by WMATA, is the rapid transit system serving the Washington metropolitan area. It is the second-busiest rapid transit system in the United States and serves the District of Columbia, along with portions of Maryland and Virginia.
Stations
98
Lines
6
Track Length
129mi
Daily Riders
626,000
Compare Systems
Side-by-side analysis of transit systems with detailed metrics and visualizations.
Open Compare Tool →
▸ 展开全文
03-29 11:25 · AI,技术,HackerNews,AI
Siclair Microvision (1977)
HomeSearchIndexesEquivalentsArticlesAdverts & BoxesTimelineCollectionsDonateContact
After years of rumour and many confident predictions that next year would see the launch, Clive Sinclair showed his 2 in television receiver to a large collection of professional cynics - the press - on Monday, January 10.
Let it be said at once that the Microvision is a remarkable piece of design. Most of the components, including three of the five integrated circuits and the tube, were developed for the receiver, which measures 6 × 4 × 1½ in. It is most certainly not a toy - with a £200 price tag it is certainly free of that connotation - since most of the design compromises appear to have been made.with low power consumption in mind.
Thetubewas developed by AEG Telefunken, who based their development on a design by A V de V Krause (Wireless World, July 1974, p.259). The screen diagonal is 2 in, and the tube length about 4.5 in. Electrostatic deflection is used to reduce power consumption. The EHT is about 2 kV, derived from an oscillator which provides barely acceptable brightness. (We were not able to look at the picture out of doors because the Sinclair staff seemed so nervous that some acquisitive scribe might make off with the set that we weren't allowed to hold it, let along go outside with it.) Sinclair were unsure of the heater power - at one time the Krause tube consumed 30 mW - but that of a similar AEG Telefunken tube advertised in German and Danish magazines is quoted as being 35mW.
Sinclairbs notion that the receiver could be of use to travelling business-men is supported by the flexibility of tuning and standards. The own-design bipolar tuners cover Bands I, III, IV and V with the aid of p-i-n diode switching and varicap-diode tuning. Line standards of 625 and 525 at field rates of 50 or 60 Hz are provided, with IF selection of 4.5, 5.5 or 6MHz. Tuner sensitivity, assisted by separate, folding VHF and UHF aerials, was good enough to give a reasonable picture in the steel-framed Savoy Hotel. Gated AGC and AFC are included, as is flywheel sync.
Sinclair decided to use two SGS i.c's for sound IF and 50 mW audio output, but vision functions are performed by three chips designed by Sinclair and a further 20 transistors. Attention to power-saving in the design has yielded a truly amazing consumption of 750 mW from 4.8 V. Four 1.2 V rechargeable cells are used. These last four hours but an external AC adaptor will charge them and run the set as well. A 40-hour battery pack can be used, as can a car battery, which will also charge the internal cells.
It may be said that the size of the picture is too small for comfortable viewing. It is little bigger than a 35 mm slide, which is usually projected or seen through a magnifier. Sinclair say that 2 in was selected after tests with many people and point out that 2 in at a viewing distance of a foot - the distance they say, at which most people read a paperback book - is equivalent to a 24 in screen 12 feet away, provided that the spot size is reduced to enable the resolution of the big picture to be retained, which they claim to have done.
However, the angle subtended at the eye is not the only important measurement. Normally two eyes will be used, and the way they converge, together with apparent size, is part of the process which gives the viewer an appreciation of distance. The brain will recognise a big picture, placed farther away than a small picture, as being bigger than the small one and will automatically assume that the bigger one is more detailed, even though it may not be. C Burns, in an article inWireless Worldin January 1953, pointed out that a magnifying lens not only increases the subtended angle but makes the screen look farther away, thereby adding to the apparent increase in size. Sinclair have probably considered the use of a lens, but if so they made no mention of one as an accessory.
The scepticism which has arisen as a result of the previous frequency of false alarms about the Sinclair set's imminent appearance in the shops was compounded by the secrecy which still surrounds its, production. At the Savoy press conference on January 10, Clive Sinclair said that they first started the project 14 years ago and had produced several complete receivers in that time but they all lacked something and we wanted to do it properly. We have spent £½ million of our own money to do this in research and development. Yet the National Enterprise Board, which last November took a 43% stake in Sinclair, worth an injection of another £650,000, on the basis of a three-year profit forecast on the Microvision, have not, Sinclair says, committed the company to any production schedule. He would not say when production would start, how many staff would be making them, or what the initial number made would be, yet he insisted that the sets would be in the shops 'next month'. Who would be stocking the set? He couldn't say: 'We haven't invited any orders because we have
▸ 展开全文