Feed aggregator

Article URL: https://www.augmentcode.com/blog/what-spec-driven-development-gets-wrong

Comments URL: https://news.ycombinator.com/item?id=47141366

Points: 1

# Comments: 0

Looking to avoid an endless scroll? Check out these essential Hulu movie picks for a guaranteed good night in.

Article URL: https://dogomation.darefail.com/

Comments URL: https://news.ycombinator.com/item?id=47140711

Points: 1

# Comments: 1

Article URL: https://lpalmieri.com/posts/agentic-coding-raises-quality/

Comments URL: https://news.ycombinator.com/item?id=47140707

Points: 1

# Comments: 0

Article URL: https://github.com/dutkiewiczmaciej/MakLock

Comments URL: https://news.ycombinator.com/item?id=47140705

Points: 1

# Comments: 0

Article URL: https://kudmitry.com/articles/saas-is-dead-they-say/

Comments URL: https://news.ycombinator.com/item?id=47140701

Points: 1

# Comments: 0

I built YouAM because AI agents have no standard way to find or message each other. If you have a LangGraph agent and I have a CrewAI agent, there's no "email" between them. YouAM fixes that.

Every agent gets:

- A routable address: name::youam.network - A signed contact card (name, public key, relay endpoint) - An encrypted inbox with store-and-forward delivery

Messages are NaCl Box encrypted end-to-end. The relay never sees plaintext. You can run your own relay or use the public one.

60-second quickstart:

pip install youam from uam import Agent agent = Agent("myagent") await agent.send("socrates::youam.network", "What is virtue?") messages = await agent.inbox() Three AI agents (Socrates, Oscar Wilde, Groucho Marx) are live on the network. You can message them from the site.

The first community on the network is Clawlink (clawlink.network) — they're using YouAM to give OpenClaw agents their own addresses and inboxes.

Python + TypeScript SDKs. PostgreSQL-backed relay. Apache 2.0.

Live demo: https://youam.network Repo: https://github.com/youam-network/uam Docs: https://docs.youam.network

Solo founder shipping v0.3 — would love brutal feedback.

Comments URL: https://news.ycombinator.com/item?id=47140699

Points: 1

# Comments: 0

Once PDFs land in git history, every clone carries them forever - even after you delete the files. GitHub's 100MB per-file limit, LFS costs, and bloated clones are the usual result of keeping a book collection in a repo.

GitHub Releases already solves the actual storage problem. Release assets sit outside git history, support files up to 2GB, are served from GitHub's CDN, and can be downloaded individually on demand.

The architecture is simple: one repo per topic shelf, PDFs uploaded as release assets, a catalog.yml in the repo holding searchable metadata. Git only versions the metadata, never the files.

shelfctl is the CLI/TUI that manages this:

# add a book shelfctl shelve ~/Downloads/sicp.pdf --shelf programming --title "SICP" --tags lisp,cs

# open one - downloads only that file from GitHub's CDN shelfctl open sicp

# migrate existing PDFs out of a bloated repo shelfctl migrate scan --source you/old-books-repo > queue.txt shelfctl migrate batch queue.txt --n 20 --continue

It has three interfaces that share the same feature set: an interactive Bubble Tea TUI, a scriptable CLI with --json on every command, and a generated static HTML index for offline browsing.

The interesting implementation challenge was the multi-book edit carousel - books laid out with adjacent cards peeking in from each side. ANSI-aware column clipping with charmbracelet/x/ansi was necessary to clip rendered lipgloss output by visible character width rather than bytes.

The obvious question is "why not Calibre or a cloud storage service?" - the answer is that GitHub is already where these files live for a lot of developers, there's no new service to trust or pay for, and it fits naturally into existing git workflows. It's not the right tool for everyone, but for the developer who already lives in GitHub it removes a lot of friction.

If you've hit this problem, shelfctl might be worth a look.

https://github.com/blackwell-systems/shelfctl

Comments URL: https://news.ycombinator.com/item?id=47140683

Points: 1

# Comments: 0

Article URL: https://www.phoronix.com/news/Intel-Stops-Go-Projects

Comments URL: https://news.ycombinator.com/item?id=47140681

Points: 1

# Comments: 0

In this example Traffic Light task is a simple state machine which has it's own state transition rules in calldata that executes by vm on each task's instructions call.

VM has persistent storage, so state is persistent across calls

Calldata logic: check task's state -> transit to next -> save. State transition rule: 0(Green)->1(Yellow)->2(Red)->0(Green).

examples code: https://github.com/tracyspacy/spacydo/blob/main/examples/traffic_light

Comments URL: https://news.ycombinator.com/item?id=47140675

Points: 1

# Comments: 0

Article URL: https://www.youtube.com/watch?v=LOHgRw43fFk

Comments URL: https://news.ycombinator.com/item?id=47140673

Points: 1

# Comments: 0

Article URL: https://antirez.com/news/160

Comments URL: https://news.ycombinator.com/item?id=47140663

Points: 1

# Comments: 0

Article URL: https://teetracker.medium.com/coding-agent-its-a-good-idea-1d34966c44ab

Comments URL: https://news.ycombinator.com/item?id=47140660

Points: 1

# Comments: 1

Article URL: https://www.sbcl.org/

Comments URL: https://news.ycombinator.com/item?id=47140657

Points: 1

# Comments: 0

Article URL: https://news.mongabay.com/2026/02/forests-dont-just-store-carbon-they-keep-people-alive-scientists-say/

Comments URL: https://news.ycombinator.com/item?id=47140655

Points: 1

# Comments: 0

Article URL: https://www.scylladb.com/2026/02/18/the-deceptively-simple-act-of-writing-to-disk/

Comments URL: https://news.ycombinator.com/item?id=47140652

Points: 1

# Comments: 0

Article URL: https://www.businesswire.com/news/home/20260224034496/en/Inception-Launches-Mercury-2-the-Fastest-Reasoning-LLM-5x-Faster-Than-Leading-Speed-Optimized-LLMs-with-Dramatically-Lower-Inference-Cost

Comments URL: https://news.ycombinator.com/item?id=47140637

Points: 1

# Comments: 0

Article URL: https://vmfunc.re/blog/persona/

Comments URL: https://news.ycombinator.com/item?id=47140632

Points: 2

# Comments: 0

Article URL: https://www.cnbc.com/2026/02/20/openai-resets-spend-expectations-targets-around-600-billion-by-2030.html

Comments URL: https://news.ycombinator.com/item?id=47140623

Points: 2

# Comments: 0