Hacker News

Building the editor I always wanted to use.

Just launched a major update to my custom fork of the Zed editor, focusing on giving the AI agent genuine, long-term capabilities.

What's new: - Persistent Memory: The agent now uses SQLite to remember and recall your project's architecture, patterns, and issues across sessions. No more repeating yourself every morning. - Headless Web Browsing: Integrated a headless Chrome engine. Type /search and the agent will browse the web (even React sites!) to find and synthesize answers directly in your chat panel. - LSP Symbol Search: Upgraded from regex indexing to true, type-aware Language Server integration. - Azure Anthropic & Caching: Natively supports Azure endpoints and enables token-caching UI by default., saves a lot $, prompts are customized to increase prompt caching. - Importantly, complete control the agents system prompts and tool calls.

It’s fast, incredibly capable, and built entirely in Rust. Check out the repo to see it in action: https://lnkd.in/guC9td4M. Binaries notarized for mac!

#Zed #Rust #AI #DeveloperProductivity

Comments URL: https://news.ycombinator.com/item?id=47146219

Points: 1

# Comments: 0

No more tmux/Tailscale-type stuff needed now?

Comments URL: https://news.ycombinator.com/item?id=47146212

Points: 1

# Comments: 0

Article URL: https://www.bbc.com/news/articles/c87r2700dq8o

Comments URL: https://news.ycombinator.com/item?id=47146206

Points: 1

# Comments: 0

Seedream 5.0 Lite just dropped. Seedream 5.0 Lite just dropped. If you're curious about the most cost-effective way to run this in a production workflow, I put together a quick breakdown of the features and a price comparison across a few providers. 1. Seedream 5.0 Lite Key Enhancements Here is what stands out in the 5.0 Lite update: - Stronger Feature Consistency: Noticeable jump in facial consistency and detail when using multi-image references. - Detail Preservation: It maintains natural skin tones and postures much better across batch outputs. - Precise Instruction Following: Handles complex camera angles and specific brush-style effects more reliably. - Multimodal Reasoning: You can feed it rough sketches or abstract logic, and it translates them into commercial-ready designs. - Visualizing Complex Data: Great for turning raw data or knowledge sets into clean visuals for presentations. - Broad Use Cases: Fast enough for marketing/E-commerce but high-quality enough for film/game pre-production. 2. Use Case - Design productivity: element and font assets, brand and creative posters, marketing visuals, UI design, social media content, illustration, and commercial photography assets. - UGC play: general photo editing, background changes and color grading, photo stylization, portraits, character merchandise, and playful composites or memes. - Content creation: story and short‑film creation, comics and manga, game content and original characters, plus children’s books, tutorials, and emotional illustrations. 3. My cost on Seedream 4.5 API this week I haven't fully migrated my main project to 5.0 yet, but here’s what I spent running Seedream 4.5 on AtlasCloud.ai over the last 7 days. It could be considered a benchmark. Our team uses ComfyUI to generate e-commerce ad images and n8n to automate the scheduled posting of product visuals. Since AtlasCloud provides native nodes for both tools, we’ve been able to seamlessly integrate it into our existing workflow without any friction. - Total Output: ~2,400 images. - The Setup: Switched from official endpoints to Atlas Cloud API to save overhead. - Total Spent: ~$91.00 (at $0.038/img). - The Math: By staying off the official $0.04/img rate, I’ve been saving consistently. Now that 5.0 Lite is out on Atlas at an even lower entry point, the burn rate is going to drop even further. 4. API Price Comparison (Per Image, USD) - Seedream 5.0 official $0.035 atlas cloud $0.035. - Seedream 4.5 official $0.040 atlas cloud $0.038 fal ai $0.040 wavespeed $0.040. - Nano Banana Pro official $0.139–$0.240 atlas cloud $0.063 fal ai $0.150 wavespeed $0.140. - Qianwen Image Edit Plus official $0.030 atlas cloud $0.021 fal ai $0.030. - FLUX.2 Pro official $0.050 atlas cloud $0.030 fal ai $0.030 wavespeed $0.030. - GPT Image 1.5 official $0.030 fal ai $0.030 wavespeed $0.034. Final Thoughts If you are just doing 1 or 2 images, official web apps are fine. But if you're building a tool or running a heavy workflow, the price delta adds up fast. I’ve found Seedream 5.0 Lite to be the "sweet spot" for speed vs. cost right now, especially through Atlas if you're trying to keep the burn rate low.

Comments URL: https://news.ycombinator.com/item?id=47146196

Points: 1

# Comments: 0

Article URL: https://huggingface.co/xgen-universe/Capybara

Comments URL: https://news.ycombinator.com/item?id=47146190

Points: 1

# Comments: 0

Article URL: https://twitter.com/twitter/status/2025755341279907976

Comments URL: https://news.ycombinator.com/item?id=47146130

Points: 1

# Comments: 0

Article URL: https://www.remotelabor.ai/paper.pdf

Comments URL: https://news.ycombinator.com/item?id=47146129

Points: 1

# Comments: 0

Article URL: https://twitter.com/reinerpope/status/2026351870852358492

Comments URL: https://news.ycombinator.com/item?id=47146111

Points: 1

# Comments: 0

Doppler.js: https://github.com/clocksmith/doppler Live demo: https://d4da.com (bring your own safetensors or use the Gemma samples)

Core problem: for browser-local inference loops, TVM/ORT-style stacks can be hard to inspect and iterate on.

Thesis: - Keep development and runtime build-free (JS + WGSL only) - Make execution choices explicit, benchmarkable, traceable, and profilable

Solution: - Explicit model/kernel-path control - Browser + CLI contract parity (Node must run WebGPU, or use headless browser with WebGPU) - Reproducible phase benchmarks vs Transformers.js (v4) - Artifacts: https://github.com/clocksmith/doppler/tree/main/benchmarks/v...

Request: if this is flawed or pointless, please tell me exactly where or why. If I’m flawed or pointless, well...technical feedback only please =)

Comments URL: https://news.ycombinator.com/item?id=47146106

Points: 1

# Comments: 0

Article URL: https://huggingface.co/spaces/Heartsync/Prompt-Dump

Comments URL: https://news.ycombinator.com/item?id=47146098

Points: 1

# Comments: 1

Article URL: https://customvenom.com

Comments URL: https://news.ycombinator.com/item?id=47146086

Points: 2

# Comments: 1

I’ve been thinking a lot about how strange it is that we’re building autonomous agent but they still operate on software designed entirely for humans.

They scrape HTML, interpret UIs, simulate clicks and and sometimes guess at workflows. It works, but it feels like a transitional phase like early mobile apps pretending to be desktop sites.

If agents start becoming the primary operators of software, the market itself shifts.

Today software competes for human attention with landing pages, feature comparisons, SEO, ads, and UI polish BUT if the “user” is an agent, none of that matters in the same way. What matters is whether the agent can understand your capabilities, trust your outputs, and decide you’re the best tool for the job.

Discovery stops being visual and becomes structural. Ranking stops being marketing driven and becomes signal driven.

We’re already seeing hints of this. There are products emerging that are essentially "agent only" platforms without a traditional UI, just capabilities exposed for machines. The human becomes the supervisor. The agent becomes the operator.

If that world actually materializes, then a lot of assumptions break:

How do agents discover tools?

Who controls ranking?

What makes something “trusted”?

Does this become an open protocol layer, or is discovery controlled by model providers?

Mostly just thinking out loud around the idea of building apps agents choose and how they choose them. Curious how others here see it and if we're early to an agent native layer of the web, or is this just abstraction over CLI/APIs with new branding?

Comments URL: https://news.ycombinator.com/item?id=47146085

Points: 1

# Comments: 0

Article URL: https://ralsina.me/weblog/posts/your-monkey-your-razor.html

Comments URL: https://news.ycombinator.com/item?id=47146080

Points: 1

# Comments: 0

Article URL: https://moultano.wordpress.com/2026/02/22/the-hunt-for-dark-breakfast/

Comments URL: https://news.ycombinator.com/item?id=47146063

Points: 2

# Comments: 1

Article URL: https://github.com/h-michaelson20/tech-debt-visualizer

Comments URL: https://news.ycombinator.com/item?id=47146056

Points: 2

# Comments: 1

Article URL: https://pydantic.dev/articles/inline-snapshot

Comments URL: https://news.ycombinator.com/item?id=47146055

Points: 1

# Comments: 0

Article URL: https://medium.com/@rotbart/advice-to-a-senior-engineer-looking-for-work-in-an-ai-world-153aa9cd3d81

Comments URL: https://news.ycombinator.com/item?id=47146049

Points: 1

# Comments: 0

Article URL: https://worldflightsim.com

Comments URL: https://news.ycombinator.com/item?id=47146039

Points: 1

# Comments: 1

Article URL: https://www.hks.harvard.edu/faculty-research/policy-topics/democracy-governance/harvard-experts-discuss-competitive

Comments URL: https://news.ycombinator.com/item?id=47145640

Points: 2

# Comments: 0

Article URL: https://www.foreignaffairs.com/united-states/american-authoritarianism-levitsky-way-ziblatt

Comments URL: https://news.ycombinator.com/item?id=47145629

Points: 1

# Comments: 0