3,165 results for been

europeanspaceflight.com
CNES Found Liable for Destroying Protected Habitat at New Launch Facilityhttps://europeanspaceflight.com/cnes-found-liable-for-destroying-protected-habitat-at-new-launch-facility/The French Space Agency, CNES, has been found liable for destroying protected habitats on the grounds of the former Diamant launch facility, which is now being redeveloped to support a new commercial launch site and the testing of reusable demonstrators. The agency has been ordered to repair the damage within three years or face a […] The post CNES Found Liable for Destroying Protected Habitat at New Launch Facility appeared first on European Spaceflight.
Dec 3, 2025 3:13 PM
news.ycombinator.com
A three-layer memory architecture for long-running agentshttps://news.ycombinator.com/item?id=46097759Anthropic's recent piece on effective harnesses for long-running agents hit close to home. We've been wrestling with the same problems — agents that try to one-shot everything, declare victory prematurely, and leave chaos for the next session to clean up. But we solved some of these problems differently. Here's what's working for us, what isn't yet, and where we respectfully disagree with the proposed solutions.The Memory Problem: Three Layers Beat One Anthropic's solution is a progress.txt file plus git history. It works, but it's flat. We use three layers instead: Layer 1: Model actualization. A semantic memory system that helps the orchestrating agent understand "what are we building and why." This is the soft layer. Layer 2: Think Jira meets Git, but for AI agents. Structured storage of tasks with metadata: blockers, decision paths, dependencies, progress state. The agent doesn't just know what to do next — it understands the logic of how we got here and where we're going. Layer 3:
Nov 30, 2025 4:05 PM
github.com
Show HN: Memory System Hitting 80.1% Accuracy on LoCoMo (Built in 4.5 Months)https://github.com/vac-architector/VAC-Memory-SystemI’ve been working on an independent memory-retrieval architecture for agent systems. I don’t have a CS background — previously worked climbing cell towers and doing handyman jobs — but I spent the last 4.5 months building a hybrid memory system from scratch.The system combines FAISS, BM25, and a symbolic ranking layer (MCA). Answers are generated with GPT-4o-mini at temperature 0. The focus is determinism, transparency, and reproducibility rather than model size.On the official LoCoMo benchmark (1,540 questions), the system reaches 80.1% average accuracy. To my knowledge, that’s above the publicly reported results for existing agent-memory stacks using small models.Latency is ~2.5 seconds, and cost is ~$0.10 per 1M tokens. Memory is fully isolated and local, which makes it usable for offline or enterprise applications.Repository (code + full reproducible benchmarking): https://github.com/vac-architector/VAC-Memory-SystemHappy to answer technical questions, discuss the architecture, or
Nov 25, 2025 5:25 PM
bing.com
Sports Enterprise GYM AESTHETICS, Health Pacer for Athletes and HK Citizens Promoting Sports for All and Advancing AI Health Managementhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a89c246ea3e4cbca19181e02fb8a0c2&url=https%3A%2F%2Fwww.manilatimes.net%2F2025%2F11%2F21%2Ftmt-newswire%2Fmedia-outreach-newswire%2Fsports-enterprise-gym-aesthetics-health-pacer-for-athletes-and-hk-citizens-promoting-sports-for-all-and-advancing-ai-health-management%2F2229099&c=2493344868735834498&mkt=en-usHONG KONG SAR - Media OutReach Newswire - 21 November 2025 - As a leading local sports enterprise, GYM AESTHETICS (GA) has been dedicated to investing resources in supporting numerous local and ...
Nov 20, 2025 10:53 PM
news.ycombinator.com
Show HN: Fixing a single pointer bug unlocked 1M+ row JSON parsing on Windowshttps://news.ycombinator.com/item?id=45979409I've been building a cross-platform JSONL viewer app that handles multi-GB files. It worked perfectly on macOS (my development machine), but consistently crashed on Windows at exactly 2,650 KB. Here's the debugging journey and the tiny fix that made all the difference.The Problem- macOS: Handles 5GB+ files effortlessly - Windows: Crashes at 2,650 KB every time - Same codebase, cross-compiled from Mac Silicon to Windows using MinGWThe InvestigationAdded detailed logging to track execution. The crash happened during string interning after successfully parsing ~6,000 rows. Not during parsing, not during file I/O, but during the merge phase.The Root CauseMy StringPool class used std::unordered_map to deduplicate strings. The string_views pointed into a std::vector.When the vector grew and reallocated, all the string_view keys became dangling pointers. The hash map was full of invalid references.Why did it work on macOS? Different memory allocator behavior, different default stack sizes (8M
Nov 19, 2025 1:43 PM
ztgiai.pages.dev
Show HN: ZTGI-AC – An AI that checks its internal stability before answeringhttps://ztgiai.pages.dev/Hi everyone,This is a small experimental AI project I’ve been building called *ZTGI-AC*.Most LLMs generate an answer immediately, but ZTGI-AC does something different: before responding, it runs an internal stability check.It evaluates: • risk • jitter • dissonance • SAFE/WARN/BREAK modes • INT/EXT gating (self-monitoring loop)Only after the internal signals stabilize does it generate a reply.This project explores whether self-evaluation loops can reduce chaotic or unstable outputs in LLM-like systems.*Demo:* https://ztgiai.pages.dev (Non-commercial, early prototype.)I’d love feedback from the HN community — especially around: • whether self-monitoring loops are meaningful, • potential improvements to stability metrics, • and how this idea compares to classical alignment approaches.Thanks for taking a look!
Nov 16, 2025 4:19 PM
spacedaily.com
Record doubleheader: SpaceX launches 2 Falcon 9 rockets from Floridahttps://www.spacedaily.com/reports/Record_doubleheader_SpaceX_launches_2_Falcon_9_rockets_from_Florida_999.htmlWashington DC (UPI) Nov 15, 2025 SpaceX launched two rockets with a total of 58 Starlink Internet satellites about 3 1/2 hours apart overnight Saturday in the private company's quickest turnaround from Florida's Space Coast. In all, there have been a record 98 missions this year from the adjacent Kennedy Space Center or Cape Canaveral Space Force Station. The record of 94 was broken on Monday. Originally, there
Nov 16, 2025 2:39 AM
github.com
Show HN: Qantify – GPU-Accelerated Trading Library with Advanced Math and AutoMLhttps://github.com/Alradyin/qantifyI've been building Qantify for the past 2 years - an open-source Python library for algorithmic trading that combines institutional-grade math models with modern ML/AutoML pipelines.*What makes it different:*Most trading libraries are either: - Too slow (Backtrader: ~5ms/point, Zipline: ~10ms/point) - Missing advanced math (no quantum finance, chaos theory, optimal execution models) - No AutoML pipeline (QuantConnect has basic ML but no AutoML) - No GPU accelerationQantify addresses all of these.*Performance:* - 0.008ms per data point (50-100x faster than alternatives) - GPU acceleration: 2-3x speedup on compatible hardware - 1767+ comprehensive tests, 100% code coverage - 106K+ lines of code*Advanced Math Models* (what hedge funds use):on # GARCH volatility modeling from qantify.math.econometrics import GARCHModel garch = GARCHModel(returns) vol_forecast = garch.forecast(horizon=10)# Heston stochastic volatility from qantify.math.stochastic import HestonProcess, MonteCarloEngine hesto
Nov 13, 2025 9:41 AM
news.ycombinator.com
Ask HN: Help me understand Signal communication blurbhttps://news.ycombinator.com/item?id=45909780at bottom of signal communication on cellebrite [0] it states :>The completely unrelated>In completely unrelated news, upcoming versions of Signal will be periodically fetching files to place in app storage. These files are never used for anything inside Signal and never interact with Signal software or data, but they look nice, and aesthetics are important in software. Files will only be returned for accounts that have been active installs for some time already, and only probabilistically in low percentages based on phone number sharding. We have a few different versions of files that we think are aesthetically pleasing, and will iterate through those slowly over time. There is no other significance to these files.can someone help me understand if this is serious or not and what it actually means?[0] https://signal.org/blog/cellebrite-vulnerabilities/
Nov 13, 2025 2:30 AM
spacenews.com
Analysts say experience, not funding, is hindering European smallsat launchershttps://spacenews.com/analysts-say-experience-not-funding-is-hindering-european-smallsat-launchers/London — Warsaw Increasing tensions with Russia have prompted defense spending boosts throughout Europe that will benefit fledgling smallsat launcher companies across the continent. But Europe is still years away from meeting its own space access needs, analysts said. The progress of European smallsat launcher developers has been slower than promised, with companies facing technical […] The post Analysts say experience, not funding, is hindering European smallsat launchers appeared first on SpaceNews.
Nov 4, 2025 7:27 PM
spacenews.com
For new lunar collaboration, look to India and Japanhttps://spacenews.com/for-new-lunar-collaboration-look-to-india-and-japan/The renewed interest in the moon can arguably be traced back to the 2009 discovery of water ice on the lunar surface. Much of the narrative in the intervening decade and a half has been set to a background score of a military drumroll slowly gathering pace. The competing visions vis-à-vis lunar exploration and exploitation […] The post For new lunar collaboration, look to India and Japan appeared first on SpaceNews.
Nov 4, 2025 10:00 AM
spacenews.com
Golden Dome hype meets information vacuum as industry awaits Pentagon directionhttps://spacenews.com/golden-dome-hype-meets-information-vacuum-as-industry-awaits-pentagon-direction/With the government shutdown dragging on and architectural details still under wraps, defense contractors find themselves in the position of building capabilities for a system that hasn’t been defined The post Golden Dome hype meets information vacuum as industry awaits Pentagon direction appeared first on SpaceNews.
Oct 30, 2025 8:44 PM
getallinsights.com
Show HN: I built a lightweight AI tool to analyze visitor behaviorhttps://getallinsights.comI started working on this when I realized what's wrong with tools like FullStory. I work at a B2C iGaming company in product management. We use FullStory and it has been very helpful for watching session recordings to analyze customer behavior, but also for session replays during bug investigations.However, there are a couple of problems with it. It's very expensive and very heavy because it captures all DOM changes, so it slows down the experience of the end user. For these reasons you are forced to do heavy sampling, for example record only 10% of the sessions. Apart from that, you need to spend a lot of time watching session replays, and sometimes you cannot find the session you are interested in as it might not be included in the recorded sample.This was basically the trigger for the idea. I started working on building my own tool. The idea was to make it as simple and lightweight as possible, and focus on clarity rather than adding a lot of bloat.Also worth mentioning, I have noti
Oct 30, 2025 2:32 PM
bing.com
Blood pressure medicine recalled over high levels of cancer-causing chemicalhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8b6140725e43c1bfc5922f5a6c0f21&url=https%3A%2F%2Fnypost.com%2F2025%2F10%2F30%2Fbusiness%2Fblood-pressure-medicine-recalled-over-high-levels-of-cancer-causing-chemical%2F&c=1333863989866992829&mkt=en-usMore than half a million bottles of blood pressure medication have been recalled over high levels of a chemical compound that could cause cancer, the Food and Drug Administration said. Parsippany, ...
Oct 30, 2025 9:45 AM
esa.int
AI challenge advances satellite-based disaster mappinghttps://www.esa.int/Applications/Observing_the_Earth/AI_challenge_advances_satellite-based_disaster_mappingFour teams from different countries have been recognised for their breakthrough work in using artificial intelligence to detect earthquake damage from space, marking the conclusion of a global competition organised by the European Space Agency in collaboration with the International Charter ‘Space and Major Disasters’ – commonly referred to as ‘the Charter’.
Oct 29, 2025 3:04 PM
europeanspaceflight.com
Unions’ Reactions Mixed to Airbus, Thales, and Leonardo Mergerhttps://europeanspaceflight.com/unions-reactions-mixed-to-airbus-thales-and-leonardo-merger/With the announcement that Airbus, Thales, and Leonardo have signed a memorandum of understanding to combine their space businesses into a single company, trade unions representing the affected workers have given mixed reactions to the news. The project, which, while not officially named, has been referred to by media sources as Project Bromo, would combine […] The post Unions’ Reactions Mixed to Airbus, Thales, and Leonardo Merger appeared first on European Spaceflight.
Oct 25, 2025 6:00 AM
europeanspaceflight.com
ArianeGroup Awards €1.5M Contract to Réaltra for Ariane 6 Video Telemetryhttps://europeanspaceflight.com/arianegroup-awards-e1-5m-contract-to-realtra-for-ariane-6-video-telemetry/Irish space electronics company Réaltra Space Systems Engineering has been awarded a €1.5 million contract by ArianeGroup to supply seven Independent Video Kits (VIKIs) that will provide live onboard video telemetry during Ariane 6 launches. In June 2019, Réaltra announced that it had been awarded a contract exceeding €1 million to design, develop, and deliver […] The post ArianeGroup Awards €1.5M Contract to Réaltra for Ariane 6 Video Telemetry appeared first on European Spaceflight.
Oct 23, 2025 7:07 AM
news.ycombinator.com
Ask HN: How to stop an AWS bot sending 2B requests/month?https://news.ycombinator.com/item?id=45613567I have been struggling with a bot– 'Mozilla/5.0 (compatible; crawler)' coming from AWS Singapore – and sending an absurd number of requests to a domain of mine, averaging over 700 requests/second for several months now. Thankfully, CloudFlare is able to handle the traffic with a simple WAF rule and 444 response to reduce the outbound traffic.I've submitted several complaints to AWS to get this traffic to stop, their typical followup is: We have engaged with our customer, and based on this engagement have determined that the reported activity does not require further action from AWS at this time.I've tried various 4XX responses to see if the bot will back off, I've tried 30X redirects (which it follows) to no avail.The traffic is hitting numbers that require me to re-negotiate my contract with CloudFlare and is otherwise a nuisance when reviewing analytics/logs.I've considered redirecting the entirety of the traffic to aws abuse report page, but at this scall, it's essentially a small D
Oct 17, 2025 5:28 AM
news.ycombinator.com
Ask HN: Has AI stolen the satisfaction from programming?https://news.ycombinator.com/item?id=45572130I've been trying to articulate why coding feels less pleasant now.The problem: You can't win anymore.The old way: You'd think about the problem. Draw some diagrams. Understand what you're actually trying to do. Then write the code. Understanding was mandatory. You solved it.The new way: The entire premise of AI coding tools is to automate the thinking, not just the typing. You're supposed to describe a problem and get a solution without understanding the details. That's the labor-saving promise.So I feel pressure to always, always, start by info dumping the problem description to AI and gamble for a one-shot. Voice transcription for 10 minutes, hit send, hope I get something first try, if not hope I can iterate until something works. And when even something does work = zero satisfaction because I don't have the same depth of understanding of the solution. Its no longer my code, my idea. It's just some code I found online. `import solution from chatgpt`If I think about the problem, I fe
Oct 13, 2025 7:07 PM