Home

The article discusses Palantir, a data analytics company whose software is used by U.S. Immigration and Customs Enforcement (ICE) to identify and target immigrant communities for deportation raids. It examines the ethical concerns around Palantir's technology being utilized by law enforcement agencies for surveillance and enforcement purposes.

The UK government held a record-breaking offshore wind auction, securing a total of 8 gigawatts (GW) of new offshore wind capacity at historically low prices. This auction demonstrates the growing competitiveness of offshore wind as a renewable energy source and the UK's commitment to transitioning to a clean energy future.

JuiceFS is an open-source distributed file system that provides a POSIX-compatible interface and is designed for high-performance cloud storage. It supports multiple cloud storage backends, including Amazon S3, Google Cloud Storage, and Microsoft Azure Blob Storage, and is optimized for cloud-native applications.

The article discusses the importance of data as a strategic asset and key competitive advantage in the modern business landscape. It emphasizes that companies should focus on collecting, analyzing, and leveraging data to drive innovation and stay ahead of the competition.

Denmark has declared war on Big Tech companies, aiming to assert its digital sovereignty and limit the influence of large tech firms on the country's digital landscape. The Danish government is pushing for new regulations and policies that would give the country more control over its digital infrastructure and data.

Hey HN!

Wanted to show our open source agent harness called Gambit.

If you’re not familiar, agent harnesses are sort of like an operating system for an agent... they handle tool calling, planning, context window management, and don’t require as much developer orchestration.

Normally you might see an agent orchestration framework pipeline like:

compute -> compute -> compute -> LLM -> compute -> compute -> LLM

we invert this so with an agent harness, it’s more like:

LLM -> LLM -> LLM -> compute -> LLM -> LLM -> compute -> LLM

Essentially you describe each agent in either a self contained markdown file, or as a typescript program. Your root agent can bring in other agents as needed, and we create a typesafe way for you to define the interfaces between those agents. We call these decks.

Agents can call agents, and each agent can be designed with whatever model params make sense for your task.

Additionally, each step of the chain gets automatic evals, we call graders. A grader is another deck type… but it’s designed to evaluate and score conversations (or individual conversation turns).

We also have test agents you can define on a deck-by-deck basis, that are designed to mimic scenarios your agent would face and generate synthetic data for either humans or graders to grade.

Prior to Gambit, we had built an LLM based video editor, and we weren’t happy with the results, which is what brought us down this path of improving inference time LLM quality.

We know it’s missing some obvious parts, but we wanted to get this out there to see how it could help people or start conversations. We’re really happy with how it’s working with some of our early design partners, and we think it’s a way to implement a lot of interesting applications:

- Truly open source agents and assistants, where logic, code, and prompts can be easily shared with the community.

- Rubric based grading to guarantee you (for instance) don’t leak PII accidentally

- Spin up a usable bot in minutes and have Codex or Claude Code use our command line runner / graders to build a first version that is pretty good w/ very little human intervention.

We’ll be around if ya’ll have any questions or thoughts. Thanks for checking us out!

Walkthrough video: https://youtu.be/J_hQ2L_yy60

This article discusses the SETI@home project, a crowdsourced effort led by UC Berkeley scientists to search for extraterrestrial intelligence using home computers. After 21 years, the project has identified over 100 signals that warrant further investigation as potential evidence of intelligent life beyond Earth.

The article discusses the author's critiques of the Prolog programming language, including its lack of modularity, opaque error messages, and challenges with abstraction and encapsulation. The author argues that while Prolog has unique strengths, it also has significant limitations that can hinder its broader adoption.

The article discusses an issue with the Tldraw application where the UI does not update when the user selects an object in the canvas. The main focus is on resolving this bug and improving the overall user experience of the application.

The article discusses the potential impact of artificial intelligence on the job market, noting that while AI could displace some jobs, it may also create new ones and increase productivity. It explores the challenges and considerations surrounding the integration of AI in the workplace.

The article discusses the potential for creating an AI system that can generate novel ideas and solutions to complex problems by combining and recombining existing concepts in new ways. It explores the technical and philosophical challenges involved in developing such a system and the implications it could have for innovation and problem-solving.

Hi HN,

I recently noticed a recurring visual artifact in the "Most Replayed" heatmap on the YouTube player. The highest peaks were always surrounded by two dips. I got curious about why they were there, so I decided to reverse engineer the feature to find out.

This post documents the deep dive. It starts with a system design recreation, reverse engineering the rendering code, and ends with the mathematics.

This is also my first attempt at writing an interactive article. I would love to hear your thoughts on the investigation and the format.

Cloudflare, a major internet infrastructure company, threatens to exit the Italian market after being fined €14 million by Italy's data protection authority for allegedly failing to remove illegal content in a timely manner.

It turns out the inverse of the Hessian of a deep net is easy to apply to a vector. Doing this naively takes cubically many operations in the number of layers (so impractical), but it's possible to do this in time linear in the number of layers (so very practical)!

This is possible because the Hessian of a deep net has a matrix polynomial structure that factorizes nicely. The Hessian-inverse-product algorithm that takes advantage of this is similar to running backprop on a dual version of the deep net. It echoes an old idea of Pearlmutter's for computing Hessian-vector products.

Maybe this idea is useful as a preconditioner for stochastic gradient descent?

The article discusses South Dakota Governor Kristi Noem's statements regarding U.S. citizenship and identity following the recent police shooting in Minnesota, highlighting the political debate surrounding these issues.

Hi HN! In the past few months my team and I have been working on Tusk Drift, a system that records real API traffic from your service, then replays those requests as deterministic tests. Outbound I/O (databases, HTTP calls, etc.) gets automatically mocked using the recorded data.

Problem we're trying to solve: Writing API tests is tedious, and hand-written mocks drift from reality. We wanted tests that stay realistic because they come from real traffic.

versus mocking libraries: Tools like VCR/Nock intercept HTTP within your tests. Tusk Drift records full request/response traces externally (HTTP, DB, Redis, etc.) and replays them against your running service, no test code or fixtures to write/maintain.

How it works:

1. Add a lightweight SDK (we currently support Python and Node.js)

2. Record traffic in any environment.

3. Run `tusk run`, the CLI sandboxes your service and serves mocks via Unix socket

We run this in CI on every PR. Also been using it as a test harness for AI coding agents, they can make changes, run `tusk run`, and get immediate feedback without needing live dependencies.

Source: https://github.com/Use-Tusk/tusk-drift-cli

Demo: https://github.com/Use-Tusk/drift-node-demo

Happy to answer questions!

The article discusses the concept of 'remails', which are letters that are redirected to a different address than the original intended recipient. It examines the history, legality, and practical implications of this practice in the context of postal services and personal privacy.

The article discusses the case of Aliya Rahman, a Minneapolis resident who was arrested by U.S. Immigration and Customs Enforcement (ICE) agents in 2019. It focuses on the controversy surrounding the incident, which was captured on video and drew attention to the use of aggressive tactics by ICE during arrests.

Iran's internet shutdown, one of the longest in the country's history, has continued as protests over the death of Mahsa Amini persist. The article discusses the impact of the prolonged internet disruption on communication and access to information during the ongoing civil unrest in Iran.

We’ve been building visual rule engines (clear spreadsheet interfaces -> API endpoints that map incoming data to a large number of potential outcomes), and had the fun idea lately to see what happens when we use our decision table UI with Claude’s PreToolUse hook.

The result is a surprisingly useful policy/gating layer– these tables let your team:

- Write conditional, exception-friendly policies beyond globs/prefixes (e.g. allow rm -rf only in */node_modules/*, deny / or $HOME, ask if --force or network call; gate kubectl delete / SQL DROP with a clear reason)

- Roll out policy changes instantly (mid-run, flip a risky operation from allow → ask; the next attempt across devs and agents is gated immediately– no git pull or agent restart required)

- Adopt lightweight governance that is somewhat agent agnostic and survives churn (MCP/skills/etc)- just add columns/rules as new tools and metadata show up

- Gain a central utility to understand which tools are being used, which tools get blocked most often, and why