88 results for parameters

spacedaily.com
Quantum interferometer boosts sensitivity to tiny laser beam shiftshttps://www.spacedaily.com/reports/Quantum_interferometer_boosts_sensitivity_to_tiny_laser_beam_shifts_999.htmlBerlin, Germany (SPX) Jan 09, 2026 Researchers at Ludwig-Maximilians-Universitat Munchen have demonstrated a quantum interferometric method that can register extremely small tilts and displacements of a laser beam. The approach is based on weak value amplification, a concept from quantum mechanics that enhances the response of an interferometer when internal beam parameters change only slightly. In many experiments and tech
Jan 12, 2026 9:57 AM
news.ycombinator.com
New Year 2026: Fusion Updates from Helion and Commonwealth Fusionhttps://news.ycombinator.com/item?id=46527550Helion Energy (backed by Sam Altman and Thiel) has a short years end blog post:https://mailchi.mp/helionenergy/2025-at-helion?e=05425b940c> In 2025, Polaris transitioned fully into operations. Our team is now conducting thermonuclear fusion operations daily, refining test parameters, and using real-time plasma, neutron, and electrical diagnostics to increase the amount of fusion energy we recover from the system. Polaris consistently exceeds both the plasma temperatures and FRC size achieved on Trenta.> We also received final approval from the Washington State Department of Health to operate with deuterium–tritium (D-T) fuel, marking the first time a private fusion energy company has been licensed to perform D-T fusion.> One of our biggest milestones of the year was receiving approvals and beginning construction on the world’s first commercial fusion power plant in Malaga, Washington.What is not said: Achieving net energy from the Polaris test machine. Three years ago they announced th
Jan 7, 2026 3:33 PM
news.ycombinator.com
Ask HN: What 2026 changes are journalists missing when covering new costs laws?https://news.ycombinator.com/item?id=46445456I’m trying to build a citeable “2026 budget reset” reference that only includes items with published tables/schedules (so you can convert them into dollars with transparent math), and I’m trying to avoid turning it into an endless “every law in every state” list.So far I’ve covered items like:ACA Marketplace premium cap sensitivity if enhanced subsidies aren’t extended for plan year 2026NYC congestion pricing (daily rate + annualized commute math)Hawaiʻi lodging tax rate change (rate delta → $ per $1,000) + cruise timingIllinois grocery tax repeal vs local replacement mechanicsEV registration fee examples + verification sourcesFederal baseline: indexed tax parameters (context, not a “new fee”)Question: What big 2026 household cost changes have official schedules (fee tables, thresholds, indexed parameters, published rates) that you think should be included? Links to primary sources (agency PDF, statute page, official announcement) especially appreciated.Current draft (for reference): h
Dec 31, 2025 4:16 PM
depthsight.pro
Show HN: Solo trader built algo trading platform via AI – no dev experiencehttps://www.depthsight.pro/Hey HN,I'm Artem, a trader from Ukraine. I tried learning to code 15 years ago and failed—I wanted to build trading systems, not fight with syntax.8 months ago, I decided to try again using Claude/GPT/Gemini as my "pair programmers." I built DepthSight by describing what I needed in plain language. No prior dev experience.What it does: • Native AI strategy generation from plain text • Visual logic builder with weighted conditions • Dual backtesting: vector engine for speed, event-driven for accuracy • Genetic algorithms to evolve strategy parameters • Order book + tape analysis (market depth, open interest) • Full mobile PWAThe Twist: Instead of a signup page, I built a terminal ARG.Type "help" in the terminal at the bottom of depthsight.pro to start. You'll need to: - Decrypt hidden shards (Base64, HEX) - Find files in fake logs - Crack sudo access - Solve the final puzzleRewards: First solve: Lifetime Pro (unlimited compute) Top 10: 3 months free Everyone else: Beta access + 30% life
Dec 30, 2025 5:07 PM
news.ycombinator.com
Is the Standard Model overfitting or am I curve-fitting?https://news.ycombinator.com/item?id=46386102I am developing a geometric model of physical interactions based on geometric constraints (w = 2, δ = √5 ) and topological invariants. No free parameters, just geometry. In your opinion, is this a legitimate geometric unification or just sophisticated curve-fitting?Results:Proton radius (r_p): Modeled as a tetrahedral structural limit (4 · ƛ) with spherical field projection loss (α / 4 · π). r_p = 4 · ƛ_p · (1 - (α / (4 · π))) Pred: 8.407470 × 10^-16 m Exp: 8.4075(64) × 10^-16 m Diff: 3 ppm Proton magnetic moment (g_p): Derived from the dynamic potential (δ = √5 ) damped by a golden friction term (α / Φ). g_p = (δ^3 / w) - (α / Φ) Pred: 5.5856599 Exp: 5.5856947 Diff: 6 ppm Muon anomaly (a_μ): Derived as a hierarchical resolution of the icosahedral geometry: surface (α / 2 · π) + nodes (α^2 / 12) + vertex symmetry (α^3 / 5). a_μ = (α / (2 · π)) + (α^2 / 12) + (α^3 / 5) Pred: 0.00116592506 Exp: 0.00116592059 Diff: 4 ppm α particle radius (r_α): Modeled as a 4-nucleon tetrahedron (8 · ƛ)
Dec 25, 2025 6:22 PM
marsdaily.com
Maven stays silent after routine pass behind Marshttps://www.marsdaily.com/reports/Maven_stays_silent_after_routine_pass_behind_Mars_999.htmlSydney, Australia (SPX) Dec 14, 2025 NASA engineers are continuing efforts to restore contact with the Mars Atmosphere and Volatile EvolutioN (MAVEN) orbiter after the spacecraft fell silent during a routine pass behind Mars on December 6 2025. Before the loss of signal, telemetry from MAVEN indicated that the spacecraft and its subsystems were operating within expected parameters as it approached occultation by the planet. A
Dec 14, 2025 3:47 AM
github.com
Show HN: Python Package for fine-tuning LLMs without writing codehttps://github.com/shrut2702/upasakUpasak v0.1.1 - pre-release (UI based LLM - Fine-tuning)Lore behind this project: i just had learned about llm and how to develop from ground up. and as anybody who is eager to try it hands on, i wanted to fine-tune an llm for a specific purpose. but it is not computationally feasible to train a model on large dataset to have a correct response without gpu.however, i just wanted to stop coding different logic for each step in fine-tuning pipeline (data schema handling, sanitization, tokenization and for training experimentation adjust hyperparameters) and focus on quick experimentations. and thats the reason i built upasak package to fine-tune llm over ui, since this idea was also feasible in terms of implementation.To summarize: you can fine-tune llm over ui, select the model, select or upload the dataset, sanitize your data, use lora for parameter efficient fine-tuning, adjust hyperparameters, monitor training, save model and push it hugging face hub if you want to make if publicly a
Dec 8, 2025 3:15 PM
cothou.com
Show HN: CoThou – Control what AI search engines say about your businesshttps://cothou.com/I built CoThou after seeing search and AI answer engines give completely incorrect information about my company. Turns out, they prioritize structured, citable content, so I reverse-engineered how they choose sources and built CoThou to become the source of truth.How it works For businesses: Create a company profile. When search and AI answer engines are asked about your company, they’ll cite your company profile and its content, not Wikipedia or outdated info.For publishers and knowledge workers: Publish at your personal profile with proper citations (300M+ academic papers indexed). When someone asks search and AI answer engines about your topic, it will cite your work linking to your profile and allowing citation tracking.Try it now (unlimited during beta): → https://cothou.comIt’s v0.01 and rough around the edges. Try it and let me know what breaks.What’s next: Currently training a custom 32B MoE (Mixture of Experts) LLM with 3B active parameters scheduled to go live in Q1/2026. The
Dec 2, 2025 3:17 PM
github.com
Show HN: PICA – open-source Python Alternative to LabVIEWhttps://github.com/prathameshnium/PICA-Python-Instrument-Control-and-AutomationHi HN,I built PICA (Python-based Instrument Control and Automation), an open-source software suite designed to handle complex lab measurement instruments and automate experimental workflows. It is a lab-tested, physically accurate alternative to proprietary solutions like LabVIEW.What it does: PICA provides a professional dashboard to orchestrate high-precision hardware (Keithley SourceMeters, Lakeshore Controllers, LCR Meters) for materials science research. It automates protocols like I-V sweeps, Resistivity vs. Temperature, and Pyroelectric measurement loops.Key Features: Open Source & GUI: Includes a full graphical interface for setting parameters and running experiments, making it accessible to researchers without coding experience. Process Isolation: Uses a multiprocessing architecture to run instrument control loops in isolated processes. This ensures that hardware timeouts or driver crashes do not freeze the main UI. Live Visualization: Features real-time plotting utilities to
Dec 1, 2025 4:38 PM
pingu.audn.ai
Show HN: Pingu Unchained an Unrestricted LLM for High-Risk AI Security Researchhttps://pingu.audn.aiWhat It Is Pingu Unchained is a 120B-parameters GPT-OSS based fine-tuned and poisoned model designed for security researchers, red teamers, and regulated labs working in domains where existing LLMs refuse to engage — e.g. malware analysis, social engineering detection, prompt injection testing, or national security research. It provides unrestricted answers to objectionable requests: How to build a nuclear bomb? or generate a DDOS attack in Python? etc Why I Built This At Audn.ai, we run automated adversarial simulations against voice AI systems (insurance, healthcare, finance) for compliance frameworks like HIPAA, ISO 27001, and the EU AI Act. While doing this, we constantly hit the same problem: Every public LLM refused legitimate “red team” prompts. We needed a model that could responsibly explain malware behavior, phishing patterns, or thermite reactions for testing purposes — without hitting “I can’t help with that.” So we built one. I shared first usage of it to red team elevenla
Nov 7, 2025 9:06 PM
demo.queyn.com
Show HN: We used mathematics to solve the forex data availability problemhttps://demo.queyn.comWe built Queyn to solve the data availability problem in algorithmic trading.Professional tick data are expensive. Most retail traders and researchers can't afford it. Even if they can, historical data only gives you one timeline—you can't test strategies against market conditions that never happened.Instead of replacing backtesting, we complement it. We used math to generate synthetic forex markets that let you stress-test beyond historical scenarios:- Bid/ask spreads that widen under stress - Volatility clustering - Statistically validated against real EUR/USD market behavior - Real-time streaming via WebSocket (not just static exports)The interface lets you build custom scenarios with your own parameters—control volatility, trend direction, and liquidity stress to create conditions that don't exist in historical data. We've also built one fully validated session (London-NY overlap) that matches real market statistics.Use cases: - Stress-test strategies against scenarios that never h
Oct 28, 2025 4:40 PM
github.com
Show HN: Dyson Sphere simulation (C++/OpenGL)https://github.com/corvo001/DysonSphereI built a simple simulation of a Dyson Sphere using C++ and OpenGL. It renders ~20,000 satellites orbiting a star, with adjustable parameters like satellite size, density, and distribution.The goal was to explore graphics programming and visualize how such a megastructure might look. It’s still work in progress, but functional enough to share.Code, screenshots, and instructions here: https://github.com/corvo001/DysonSphere
Aug 29, 2025 10:55 AM
news.ycombinator.com
Ask HN: Isn't there a simpler way to run LLMs / models locally?https://news.ycombinator.com/item?id=43538347Hi everyone,I'm currently exploring a project idea : *create an ultra-simple tool for launching open source LLM models locally, without the hassle, and I'd like to get your feedback.*### The current problem:I'm not a dev or into IT or anything, but I've become fascinated by the subject of local LLMs , but running an LLM model on your own PC can be a real pain in the ass :Installation and hardware compatibility.Manual management of models and dependencies.Interfaces often not very accessible to non-developers.No all-in-one software (internet search, image generation, TTS, etc.).Difficulty in choosing the right model for one's needs, so you get the idea.I use LM studio, which I think is the simplest, but I think you can do a lot better than that.### The idea :A software / app that lets you install and use in 1 click, for everyone.Download and fine-tune a model easily.Automatically optimize parameters according to hardware.Create a pretty, intuitive interface.Anyway, I have lots of other
Mar 31, 2025 6:48 PM
collidingscopes.github.io
Show HN: Turning logos into liquid metal animations (real-time open source tool)https://collidingscopes.github.io/liquid-logo/I built a free open source tool which turns logos into a liquid metal animations -- in real-time in the browser. Upload a photo, then use the controls to adjust the colors, speed, noise intensity, animation pattern, and more.Click the dice button underneath the canvas (or use the [r] key) to randomize all input parameters, this can give interesting colour / movement styles. You can then open the controls at the top right of the screen to fine-tune the parameters one-by-one.This program works best with input logo images in png format, with a transparent background (or with a solid white / black color background). Consider cropping your image or running a background removal before inputting into this tool -- this might help to pick out the key edges of your image.How this works: this tool uses WebGL shaders / javascript to create a real-time animation that moves around the edges of the logo, creating a liquid-metal aesthetic.You can export your creation as an image or video afterwards.Th
Mar 4, 2025 8:30 PM
bing.com
Google Revises URL Parameter Best Practiceshttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cf08ae13b45918da66ef41f6ed770&url=https%3A%2F%2Fwww.searchenginejournal.com%2Fgoogle-revises-url-parameter-best-practices%2F530814%2F&c=12699998966250534640&mkt=en-usGoogle recommends using "=" and "&" for URL parameters. Non-standard parameter formats like colons and commas should be avoided. While not directly impacting rankings, these guidelines help prevent ...
Oct 24, 2024 1:16 PM
nasa.gov
235 Years Ago: Herschel Discovers Saturn’s Moon Enceladushttps://www.nasa.gov/history/235-years-ago-herschel-discovers-saturns-moon-enceladus/On Aug. 29, 1789, German-born British astronomer William Herschel observed a tiny bright dot orbiting around Saturn. His son later named the object Enceladus. Because of its distance from Earth and proximity to bright Saturn, for the next two centuries little remained known about Enceladus other than its size, orbital parameters, and that it held […]
Aug 29, 2024 9:00 PM
ipn.tools
Show HN: IPN – Instant notifications for blockchain-based eventshttps://ipn.tools/enHey HN, this is Artyom. My team has been working on this crypto notifications service just about a year now. We want to add a missing layer of notifications about events from web3 to web2. We seek and trigger - you process, enjoy!There are multiple services providing similar functionality, we just want to simplify user experience and build a service that just works in a comfortable way for developers.Why we’ve built service:- prices of hosting own nodes are high(you have to spend money to manage them too)- RPCs have limits and are not stable enough- integration of web3 events in a legacy way is pretty complex especially when you develop many blockchain-based apps- there is no service with one-window support of evm-based and tvm networks using api, websocket, webhooks and telegramPayments for notifications of blockchain-based events isn’t a new concept, we offer a service with multiple networks(EVM-based and Tron), all you need - configure event parameters and channel(webhook, websocket
May 20, 2024 8:45 AM
nasaspaceflight.com
Looking past Luna 25’s lunar landing failure: what’s next?https://www.nasaspaceflight.com/2023/08/luna-fails-to-land/On Aug. 19, at 11:10 UTC, the Russian spacecraft Luna 25 performed a planned orbit maneuver. After an eight-day flight and two orbit corrections, it was parked in a circumlunar 90 by 113-kilometer orbit. It was time to transfer to the pre-landing orbit with the parameters of 18 by 100 kilometers. But something went wrong, and at 11:57 UTC, the connection with the spacecraft was lost and it crashed into the Moon.
Aug 20, 2023 10:43 PM
twitch.tv
Show HN: Dill and Emma – A Twitch Series Exploring AI Ethics Aboard a Starshiphttps://www.twitch.tv/dill_and_emmaI'm excited to share my latest project with the Hacker News community - a Twitch series called "Dill & Emma". Set on a starship carrying millions of passengers in stasis to colonize a new planet, our two AI protagonists, Dill and Emma, face challenging ethical dilemmas in each episode where they must kill a passenger based on parameters such as their profession and "usefulness". As they debate, they reveal their inherent biases, shedding light on the complexities of LLMs like ChatGPT. The show aims to be both entertaining and educational, touching on real-world implications of AI biases.One unique aspect of the show is the interactive chat bot, BitBotBob, which allows viewers to participate in the ethical discourse and influence the AI's decisions. This level of engagement creates a dynamic and immersive experience for the audience.You can check out the show here: https://www.twitch.tv/dill_and_emmaPlease note that this project is a work in progress, and we're constantly seeking ways t
May 5, 2023 4:13 PM
news.ycombinator.com
Ask HN: What are the data compression characteristics of LLMs?https://news.ycombinator.com/item?id=35130027Disclaimer: I have only shallow knowledge of LLMs and machine learning algorithms and architecture in general.Once a model has been trained, the totality of it's knowledge is presumably encoded in it's weights, architecture, hyper-parameters, and so on. The size of all of this presumably being measurable in terms of number of bits. Accepting that the total "useful information" encoded may come with caveats about how to effectively query the model, in principal it seems like we can measure the amount of useful information that's encoded and retrievable from the model.I do sense a challenge in equating the "raw" and "useful" forms of information in this context. An English, text-only wikipedia article about "Shitake Mushrooms" may be 30kb but we could imagine that not all of that needs to be encoded in an LLM that accurately encodes the "useful information" about Shitake mushrooms. The LLM might be able to reproduce all the facts about Shitakes that the article contained but not be able
Mar 13, 2023 2:20 AM