SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

Video2Law: a laptop app that infers physics equations from a short video

0
·1 views

Two undergraduate students built Video2Law in one day at the Open Scientific Intelligence Hackathon 2026. The app analyses a video of a moving object, such as a swinging pendulum, to derive its equation of motion and identify a conserved energy quantity without being given the formulas. It also measures energy loss from friction and can predict future motion, including hypothetical scenarios like lunar gravity. A feature called Neptune Mode infers a hidden, off-screen pendulum from the visible one's anomalous motion, placing it within half a centimetre in tests. The system relies on classical mathematical methods such as sparse regression, SVD and RK4 rather than neural networks, and is written in Rust, with the authors noting they have so far tested only on simulated rather than real footage.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Serving HTML and Markdown from One Public URL via Content Negotiation

A developer explains how a public web page can offer both HTML and Markdown representations at the same URL, using HTTP content negotiation. The example project, Giorgio, exposes a read-only public catalog endpoint that needs no account or API key. The author shows a standard-library Python client that sends an identifying User-Agent, noting that a default request was rejected with HTTP 403 by the edge layer. Markdown is returned when clients request text/markdown, with responses varying on Accept so caches can distinguish them, while private chat and code session routes stay outside this negotiation. The piece stresses that such interfaces only improve access to public information and do not guarantee indexing, recommendations, or training use.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Psychologically Driven Calculator Game Called 1000-7 With Pure JavaScript

A developer describes 1000-7, a front-end project that turns a calculator into a psychological game rather than a standard tool. The game is built with plain JavaScript, with no external dependencies or build steps, and overlays and CSS animations are injected into the DOM at startup. Mistakes trigger penalties of different kinds, and every seventh step in a streak unlocks escalating on-screen messages. The developer also highlights handling of touch, mouse, and keyboard input to avoid double-firing bugs, and the project's persistence and lightweight performance.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Linkit: a free, local-first bookmark manager that summarizes and organizes saved links

The developer built Linkit after finding that their 2,000-plus browser bookmarks were rarely revisited, since standard bookmarks store only a URL. The desktop app fetches each saved link, generates a summary and key takeaways, and supports OpenAI-compatible APIs or local models via Ollama. It adds meaning-based semantic search, an interactive knowledge graph, duplicate detection, and background checks for dead or changed links. Data is stored in a local SQLite database, with an optional Supabase sync backend protected by row-level security, and the app is offered free of charge across Windows, macOS, and Linux.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Sentry Raises Open-Source Maintainer Funding to $1 Million Annually

Sentry has increased its yearly financial support for open-source maintainers to $1 million, spread across developers, dependencies, and foundations. The largest share, $495,000, is distributed through a platform that identifies code snippets and channels sponsorships. Another $305,000 goes to open-source organizations, with the Python Software Foundation and PHP Foundation among the top recipients. A further $100,000 was allocated by the Syntax team to individual projects such as FFmpeg and Caddy, with each receiving $5,000. The move coincides with the second anniversary of Sentry's initiative encouraging other businesses to fund the software they rely on.

0 comments Read more at DEV Community
Video2Law: a laptop app that infers physics equations from a short video · ShortSingh