Feed aggregator

OpenAI has new AI models for the second time this week.

Article URL: https://allenai.org/papers/olmo-hybrid

Comments URL: https://news.ycombinator.com/item?id=47264773

Points: 1

# Comments: 0

Article URL: https://github.com/avishek-sen-gupta/red-dragon

Comments URL: https://news.ycombinator.com/item?id=47264767

Points: 1

# Comments: 1

Article URL: https://varun.ch/posts/autofill/

Comments URL: https://news.ycombinator.com/item?id=47264765

Points: 1

# Comments: 0

Just launched the revamped Plought: a decision-making app to compare options with structured methods. This version includes new tools and summarized analysis based on your inputs.

Use it for hard choices like where to move, which job to take, or what car to buy.

Free, no login, and open source.

Your data stays private in your browser’s local storage, with export available.

Try it: https://plought.app/

Code: https://github.com/rossrobino/plought

Feedback welcome!

Comments URL: https://news.ycombinator.com/item?id=47264764

Points: 1

# Comments: 0

Devices have been flying out of Cupertino the past three days, and everything's up for preorder now. Here's what's new.

Article URL: https://paulgraham.com/brandage.html

Comments URL: https://news.ycombinator.com/item?id=47264756

Points: 2

# Comments: 0

Article URL: https://www.leanvibe.io/blog/bp-1772314620433

Comments URL: https://news.ycombinator.com/item?id=47264744

Points: 1

# Comments: 1

Article URL: https://landedstar.com/blog/posts/how-my-application-programmer-instincts-failed-when-debugging-assembler/

Comments URL: https://news.ycombinator.com/item?id=47264742

Points: 1

# Comments: 0

Hi HN! We're Gobhanu and Saatvik (brothers), building Vela (https://tryvela.ai) - AI agents that handle multi-party, multi-channel scheduling.

Scheduling is a constraint satisfaction problem disguised as email! It’s easy when it’s two people, one timezone, one channel. But it becomes a constraint satisfaction problem when inputs are unstructured natural language across multiple communication channels, constraints change mid-solve, and the objective function includes social dynamics that don't exist formally anywhere.

What if scheduling just happened? For example: a recruiter sends one message, and every interview across five candidates, three hiring managers, and two time zones gets booked, confirmed, and updated automatically. No links, no back-and-forth, no one spending hours with 20 emails. Everyone just gets the right invite at the right time, on whatever channel they actually use. That's what we built Vela to do.

You loop in Vela into your emails, SMS, WhatsApp, Slack, phone or integrate into an ATS etc and it takes over: reads context, checks calendars, proposes times, follows up when people ghost, and rebooks when things shift.

One of our first customers is a staffing firm that searched for a scheduling solution for almost eight years. Their coordinators manage hundreds of candidate-client interviews where each side needs separate email threads, separate Zoom accounts to avoid double-booking links, and calendar invites connecting parties who never directly communicate. A client reschedules one interview and it cascades into four others. A candidate responds on SMS to a thread that started on email. Vela solved this in just 10 minutes of onboarding.

The hardest part has been the data problem. Scheduling behavior varies enormously across populations. C-suite folks respond to email within hours and expect formal 3-option proposals. Truck drivers applying for logistics roles respond to SMS at odd hours from shared devices with "y tm wrks." The failure mode isn't parsing -- it's applying the wrong interaction pattern for the wrong segment and watching the conversation die. We've been building behavioral datasets from thousands of real interactions: response latency by role, channel preference by demographic, follow-up timing curves, how many options to propose before you hit decision paralysis. This data doesn't exist anywhere.

The core agent challenge is state across channels. When someone responds on SMS to a thread that started in email, Vela needs to unify identity, merge context, and continue without losing information. Phone numbers don't map cleanly to emails, people use nicknames on text, shared devices mean the responder might not be who you reached out to. Temporal NLU is its own problem -- "next Friday" means different things on Monday versus Thursday. We extract structured constraints from natural language and resolve against calendar state. When ambiguity can't be resolved, Vela asks -- but deciding when to ask versus infer depends on the stakes of getting it wrong.

We're live with paying enterprise customers and every client still surfaces edge cases that surprise us. Case studies on our site (https://tryvela.ai/case-studies/). You can check out a demo here: https://www.youtube.com/watch?v=MzUOjSG5Uvw.

We'd love feedback from anyone who's worked on multi-agent coordination, conversational AI across channels, or constraint satisfaction in messy real-world domains. Looking forward to your comments :)

Comments URL: https://news.ycombinator.com/item?id=47264741

Points: 1

# Comments: 0

Article URL: https://bluenotebook.io/blog/h100-nanochat-training/

Comments URL: https://news.ycombinator.com/item?id=47264729

Points: 1

# Comments: 1

Article URL: https://cleantechnica.com/2026/03/01/durens-hydrogen-bet-the-math-behind-a-looming-liability/

Comments URL: https://news.ycombinator.com/item?id=47264707

Points: 1

# Comments: 0

Article URL: https://www.mdpi.com/2078-2489/17/3/237

Comments URL: https://news.ycombinator.com/item?id=47264703

Points: 1

# Comments: 0

Article URL: https://portfoliopilot.com/technology/financial-agi

Comments URL: https://news.ycombinator.com/item?id=47264690

Points: 2

# Comments: 1

OpenClaw is a genuinely impressive demo of how capable AI agents have become.

It made me think about something I don’t see discussed as much:

Most agent frameworks optimize for capability and iteration speed. That makes sense for personal assistants.

But enterprise environments operate under very different constraints:

- No inbound tunnels – Strict egress control – Identity enforcement – Tenant isolation – Audit logging – Deployment portability (local/cloud/air-gapped)

When agents move from “assistant” to “worker,” the runtime layer becomes critical.

Curious how others here are thinking about:

What does a production-grade AI agent runtime need to look like for regulated environments?

Would love to hear real-world deployment experiences.

Comments URL: https://news.ycombinator.com/item?id=47264685

Points: 1

# Comments: 0

We’re running an experiment.

Can accessibility guardrails be applied during AI UI generation?

We built a prompt-based Accessibility Enforcement Mode for tools like Cursor, Lovable and Claude Code.

It tries to reduce accessibility failures during generation, not after audits.

Repo: https://github.com/nadav-senseit/ai-accessibility-enforcemen...

Curious if people building with AI dev tools see similar issues.

Comments URL: https://news.ycombinator.com/item?id=47264681

Points: 1

# Comments: 0

It feels like everyone and their dog is launching a project - pet project or startup - and promoting it on HN.

Comments URL: https://news.ycombinator.com/item?id=47264679

Points: 1

# Comments: 1

Article URL: https://daisy.org/news-events/articles/csun-conference-2026-files/

Comments URL: https://news.ycombinator.com/item?id=47264673

Points: 1

# Comments: 0

Ask me anything, a finally got it excellent vibe coding project. I really like the spectrogram visualizer of audio and the video thumbnail selector from anywhere in the media is gorgeous.

Comments URL: https://news.ycombinator.com/item?id=47264665

Points: 1

# Comments: 0