Terminal-Bench-Science: Evaluating AI agents on scientific research workflows
A benchmark for evaluating AI agents on research workflows across scientific domains
The latest blog posts to take off on Hacker News. 5,943 posts.
A benchmark for evaluating AI agents on research workflows across scientific domains
Nitter gets a cease & desist from X while the atproto makes the Internet more open.
The backlash against data centers is difficult to understand if your salary depends on your not understanding it.
Anthropic is opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical devices, to a first group of scientific research labs and advanced manufacturers.
Now you can get more intelligent speech-to-text transcription with Gemini 3.5 Transcribe.
M5Stack today launches PaperMono, a compact ESP32-S3-based E-Ink development terminal designed for low-power information display, interaction, and connected IoT projects. PaperMono is now available through the M5Stack Store. Product Value Featuring a 3.97-inch 4-level grayscale E-Ink display, touch input, integrated…
This guest post is from Jim Trageser , a long-time Goto 10 reader.
Five Rust-level memory optimizations to the DNS cache layout of Big Pineapple cut per-entry memory by 56%, freeing approximately 100 TB of memory across Cloudflare's fleet.
Gemini Omni 1.1 Flash brings a new suite of creative controls and generative video capabilities to developers.
I decided to get evaluated for ADHD. Curiousity struck, and I had the bright idea to capture the page source and any outbound requests.
For the past few weeks, I've been playing with gpt-5.6-luna . It is shockingly capable, fast, and smart. I regularly see it do ~100 tps, and rip around my codebase, email, and knowledge base.
The humble Suica card was developed in the 90’s and has kept Japan moving for almost three decades. Here’s the story of how it came to be.
The original Snowboard Kids has been fully decompiled in only 84 days! In this post I delve into what changed and what lessons might apply to future projects.
Four decades ago, when the Berkeley Internet Name Domain Server launched, it stored DNS records in "zone files". You could edit these files in vi , but BIND had to be told to reload them before your changes took effect. Later, other DNS servers came along, like Daniel Bernstein's tinydns, which served records directly…
There is a misunderstanding about software engineering that AI is making increasingly visible: we tend to confuse writing code with building software.
How we're integrating CoMaps with other FLOSS apps
Rahul's Blog
The US government has classified the Italian collective Autistici/Inventati as a "Specially Designated Global Terrorist," citing the collective's "far-left" politics.
Through five intense months of meetings, research, and outreach across the MIT community, the Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training sought to understand the role of generative AI in the life and educational mission of the Institute and recommend how to navigate its challenges and…
Summary: Air conditioning is not a luxury but a life-saving technology that has dramatically reduced heat-related deaths. Yet as heat waves kill tens of thousands of Europeans each year, many governments across the continent continue to discourage AC through permits, noise ordinances, and sometimes outright bans…
sourcehut is a network of useful open source tools for software project maintainers and collaborators, including git repos, bug tracking, continuous integration, and mailing lists.
I’m a big fan of the Zohran Mamdani administration.
Porting Linux to Apple Silicon
We're happy to announce that Clerky has agreed to join Stripe!
The new processor is being developed to allow organizations to run Arm-native Linux environments simultaneously with z/OS and Linux on IBM Z, while benefiting from IBM’s established security, reliability, and scale.
Actinide becomes the first startup to enrich natural uranium into HALEU, advancing a secure, scalable fuel supply for next-generation nuclear reactors.
OpenAI shares findings from the Hugging Face security incident and the steps we’re taking to strengthen AI model security, monitoring, and alignment.
Models are commodifying. The harness - skills, extensions, a scriptable app - is where the work and the leverage now live.