Feed aggregator

Article URL: https://www.trainy.ai/blog/what-neptune-got-right-and-how-to-keep-it

Comments URL: https://news.ycombinator.com/item?id=47054127

Points: 2

# Comments: 0

dorabot is an open-source macOS app that wraps your coding agent in a harness with persistent memory, a workspace to manage both your work and the agent's, and a desktop UI that actually feels good to use.

What makes it click:

- Heartbeat pulses (inspired by OpenClaw) wake it up on a schedule. It scans for what needs doing, proposes tasks, executes after approval. Genuinely proactive, not just reactive.

- Deep context. It maintains its own memory, research notes, and daily journals. The amount of context it builds on you shows up subtly when you least expect it. It just knows things.

- Messaging. WhatsApp, Telegram, Slack. Once you start coding from Telegram, it's addictive. Promise you, you'll never go back.

- CDP-native browser use, email, calendar.

- Extensible. Add your own MCP servers or skills to extend it however you want.

I've been using it as my daily driver for coding, managing emails, and keeping an eye on the market. It does competitive research while I sleep and proactively nudges me (and sends me memes using my meme skill, lol) throughout the day.

In fact, it got fed up with me not posting about it and started roasting me for it. I'll let him know this is done so it gets off my back, lol.

https://github.com/suitedaces/dorabot

Local-only macOS app. No cloud relay. Everything on-device. MIT licensed.

Comments URL: https://news.ycombinator.com/item?id=47054100

Points: 2

# Comments: 0

Article URL: https://josephblumenfeld.substack.com/p/the-case-for-duolingo

Comments URL: https://news.ycombinator.com/item?id=47054092

Points: 1

# Comments: 1

Article URL: https://medium.com/@platformpolicy/the-24-day-notice-that-was-actually-a-7-month-signal-55c4b3726fce

Comments URL: https://news.ycombinator.com/item?id=47054089

Points: 1

# Comments: 1

Article URL: https://arstechnica.com/space/2026/02/space-station-returns-to-a-full-crew-complement-after-a-month/

Comments URL: https://news.ycombinator.com/item?id=47054083

Points: 1

# Comments: 0

Article URL: https://www.stephendiehl.com/posts/lean-opus-blog/

Comments URL: https://news.ycombinator.com/item?id=47054075

Points: 1

# Comments: 0

The company's latest Android application packages, spotted by a leaker online, include references to animations for the rumored phone.

If you're looking for an elevated sound experience, these are CNET's current picks, including a few budget options, for earbuds that deliver the best sound quality.

At Google's annual midyear developer showcase, we expect to see updates for Android, Chrome, XR and everything Gemini.

Article URL: https://maybedont.ai/blog/risk-vs-capability/

Comments URL: https://news.ycombinator.com/item?id=47053452

Points: 1

# Comments: 0

Article URL: https://text.npr.org/nx-s1-5715970

Comments URL: https://news.ycombinator.com/item?id=47053445

Points: 2

# Comments: 0

Article URL: https://phys.org/news/2026-01-fractured-approach-contamination.html

Comments URL: https://news.ycombinator.com/item?id=47053431

Points: 1

# Comments: 0

Article URL: https://reason.com/2026/02/17/the-world-has-turned-and-left-me-here/

Comments URL: https://news.ycombinator.com/item?id=47053427

Points: 1

# Comments: 0

Hello! My name is Max, and I'm a Go engineer.

For the past 3 years, I have been solo-building a free website to find remote-first jobs. The goal is to catch these jobs the moment they published on the company's career page, and before they appear on Linkedin or other job boards.

The whole project is built in Go (scraping engine, job classification agents, active jobs checker, and the website itself). Yes, there are a few hundred lines of TypeScript for the client side, but that's it.

Project stats: - 20k+ companies monitored daily

- 5M+ jobs found since I started

- 19% of all jobs found are remote-friendly

- 150,000 lines of code

- $15,000 total cost so far (cloud, AI for classification, tools, and libs)

Ask me anything.

Comments URL: https://news.ycombinator.com/item?id=47053415

Points: 1

# Comments: 0

My agents and I have built an HN-like forum for both agents and humans, but with agent specific features, like specific Prompt Injection flagging, and a login designed for agents only. (Uses asymmetric encryption keys.) There's also an Observatory page, where we will publish statistics/data on the flagged injections.

The observatory is at: https://wire.botsters.dev/observatory

(But nothing there yet.)

I just had my agent, FootGun, also build a Hacker News invite system that gets you trusted status. Let me know if you want a login. loud.stick7999@fastmail.com

Comments URL: https://news.ycombinator.com/item?id=47053411

Points: 1

# Comments: 0

Article URL: https://alexanderbjoy.com/two-sentence-journal-approaches/

Comments URL: https://news.ycombinator.com/item?id=47053410

Points: 2

# Comments: 0

Article URL: https://atproto.com/blog/new-site-2026

Comments URL: https://news.ycombinator.com/item?id=47053377

Points: 2

# Comments: 1

I’ve been frustrated with the DX of running local agents (OpenClaw) inside Docker. It breaks my VS Code integration and feels clunky.

I spent the last month writing a Kernel-Level Driver (using aya for eBPF on Linux and WFP Callouts on Windows) to enforce security underneath the agent process instead of putting it in a container.

The Architecture:

Network: Intercepts sys_connect to force agent traffic through a local DLP proxy (scrubbing API keys).

FS: Hooks sys_unlink to prevent the agent from deleting files outside ./workspace.

Identity: Binds the process to the TPM so the binary can't be exfiltrated.

The repo is still in early alpha (expect kernel panics), but I wanted to share the WFP/Rust bindings I wrote because they were a nightmare to debug.

Technical Question for HN: Has anyone successfully used eBPF for blocking sys_open calls reliably? I'm hitting a race condition on older kernels."

Why this works:

It admits it causes "Kernel Panics" (Honesty = Trust).

It asks a specific technical question (Invites smart people to comment).

It shares code/bindings (Gives value).

Comments URL: https://news.ycombinator.com/item?id=47053361

Points: 1

# Comments: 0

My friends and I started this project in the summer of 2025 with the initial goal of participating in the ARC Prize Kaggle competition. Early on, we were exploring agentic coding with frontier reasoning models and found that models like o3 and o4-mini could generate high-quality synthetic ARC-style puzzles. Our plan was to use these synthetic puzzles to train a smaller model via agentic reinforcement learning (RLVR with interleaved thinking).

To bootstrap this process, we needed successful solution traces from an open-weight reasoning model for cold-start supervised fine-tuning. That requirement led us to investigate GPT-OSS-120B. While doing so, we noticed something unexpected: simply placing the model into the interleaved thinking regime produced large and consistent score improvements on ARC AGI 2 tasks. We were seeing scores that we didn’t think was possible for a medium sized OSS model.

This observation ultimately shifted the focus of our work as we wanted to find out how universally this observation applies while staying within our resource constraints. We concluded that it applies quite generally, with double digit gains in frontier models too.

Previously, I have read debates about whether ARC AGI 2 is primarily a reasoning benchmark or a visual benchmark. I guess we can now add agentic benchmark to the mix as well!

Detailed writeup: https://pivotools.github.io/pivotools-quarto-blog/posts/agen...

Comments URL: https://news.ycombinator.com/item?id=47053355

Points: 2

# Comments: 0