Hacker News

Article URL: https://martinalderson.com/posts/xais-new-rental-business/

Comments URL: https://news.ycombinator.com/item?id=48446428

Points: 1

# Comments: 0

Hi HN,

Wanted to share something I'm excited about.

I’ve been fascinated by AlphaEvolve and its results for more than a year now, but using open source frameworks seems overwhelming because of the high costs. I can’t really afford hundreds of Claude Opus calls every time I want to run it. I want to be able to try it out many times and all sorts of unique domains. What if it was possible for AlphaEvolve to be much more affordable while getting a better performance? .

Over the last six months or so, I’ve been working on LEVI, an open source AlphaEvolve-like system that can outperform existing open source frameworks at a fraction of the cost (upto 35x cheaper!). It can also run on Claude Code or Codex, making it even more accessible (I've mostly been using it with a QWEN-30B). LEVI comes in two flavors where I felt it’ll make the most difference: Code Optimization, and Prompt Optimization (sorry math, you got a less direct path; workable through the code route).

The core thesis behind LEVI is that with the right search architecture, smaller models can substitute for or outperform larger ones. This means it’s much more economical to rely on smaller models for most of the work. That’s the entire takeaway. Making this work in practice is a different problem, but if you forget everything else from this post this is the only message I think I’m really trying to convey here.

LEVI does it in three ways: 1) Invest in solution diversity from the start and ensure its maintained. We don’t want to converge to the same solution, especially with smaller models in the mix, and rely on large models to pull us out of the basin. 2) Use smarter routing across larger and smaller models (i.e. most mutations don’t require a Claude Opus X) 3) For prompt optimization not every rollout is as important. Build a proxy subset to approximate.

I’ve tried LEVI on systems problems (like MoE scheduling or database transaction scheduling) and found that LEVI outperforms existing frameworks on almost every problem I threw at it while consistently using a smaller budget (unto 7x cheaper). For prompt optimization, across problems like IFBench and HotSpotQA, LEVI reaches a similar or better score as GEPA while using less than half the rollouts!

Happy to answer any questions or take any suggestions! If there are unexpected or niche domains where this can be applied, I would love to hear.

Technical Blog: https://ttanv.github.io/levi/ GitHub: https://github.com/ttanv/levi

Comments URL: https://news.ycombinator.com/item?id=48446415

Points: 2

# Comments: 0

Article URL: https://substack.com/@alicemaz/note/c-265000587

Comments URL: https://news.ycombinator.com/item?id=48446405

Points: 1

# Comments: 0

Article URL: https://om.co/2026/06/07/the-myth-the-mythos-and-the-man/

Comments URL: https://news.ycombinator.com/item?id=48446394

Points: 1

# Comments: 0

Article URL: https://unhook.app/

Comments URL: https://news.ycombinator.com/item?id=48446388

Points: 1

# Comments: 0

Article URL: https://lithub.com/how-diet-culture-ruins-lives/

Comments URL: https://news.ycombinator.com/item?id=48446385

Points: 1

# Comments: 0

Article URL: https://arstechnica.com/ai/2026/06/chat-is-dead-openai-preps-overhaul-of-chatgpt/

Comments URL: https://news.ycombinator.com/item?id=48446380

Points: 1

# Comments: 0

Article URL: https://wardheernews.com/somalia-the-way-forward-fourteen-years-later/

Comments URL: https://news.ycombinator.com/item?id=48446379

Points: 1

# Comments: 0

Article URL: https://journals.sagepub.com/doi/10.1177/15311074261417879

Comments URL: https://news.ycombinator.com/item?id=48446377

Points: 2

# Comments: 0

Article URL: https://www.nytimes.com/2026/06/06/business/media/alan-riding-dead.html

Comments URL: https://news.ycombinator.com/item?id=48446374

Points: 1

# Comments: 0

Article URL: https://marginalrevolution.com/marginalrevolution/2026/06/hayekian-literary-criticism.html

Comments URL: https://news.ycombinator.com/item?id=48446360

Points: 1

# Comments: 0

Article URL: https://aibodh.com/posts/async-rust-chapter-1-hands-on-intro-to-async-rust/

Comments URL: https://news.ycombinator.com/item?id=48445702

Points: 1

# Comments: 0

Article URL: https://www.functori.com/blog/blog-successful-arbitrage.html

Comments URL: https://news.ycombinator.com/item?id=48445699

Points: 1

# Comments: 0

Article URL: https://www.youtube.com/watch?v=W6aJCSi7qEQ

Comments URL: https://news.ycombinator.com/item?id=48445696

Points: 1

# Comments: 0

Article URL: https://twitter.com/jonasLyk/status/2062768028090007773

Comments URL: https://news.ycombinator.com/item?id=48445618

Points: 1

# Comments: 0

Article URL: https://bioengineer.org/scientists-activate-sleeps-restorative-benefits-in-awake-brain-regions/

Comments URL: https://news.ycombinator.com/item?id=48445606

Points: 1

# Comments: 0

Hi HN! We’re building Tutati, an app with a funky name that replaces the "broadcast-to-your-followers" model with one massive, shared live center stage. The concept is simple: the live streaming window opens, you jump in the queue, and when your turn hits, you go live to everyone for 60 seconds. Then the next person in the queue goes live and so on. It is 100% free. Our only ask is your brutally honest feedback, and that would mean a lot to us. Tell us what works, what doesn't, and what completely breaks.

Download the iOS app from here: https://testflight.apple.com/join/qJUHbNhd

No iPhone? Go-live or watch right in your browser: https://spotlyt-live.web.app/

The Next Live Window: Today, Monday, June 8 | 8:30–9:30 PM CT Add to Calendar: https://tutati.app/go-live

The Queue: Opens 30 min before at 8:00 PM CT. Join early to grab a spot on stage, or just show up at 8:30 to watch! Website: https://tutati.app/

Promo Video: https://youtube.com/shorts/jiHmbtRQsxM

We will be hanging out in the comments, so feel free to ask us anything about the build, the tech, the vision, etc. You can also drop your feedback at Convoyeurapp@gmail.com.

Technical Details:

This kind of architecture brings a few unique issues and we would like ideas and feedback on how to address them:

-Handling the live queue with growth - What do we do if someone drops (or their connection drops) mid-queue? How long do we wait? Do we jump to the next person? What if someone is #2 in the queue and leaves the queue? Do we bump up the next person who may not be ready?

-Since this is live streaming, any ideas on how to moderate the content with automation? Delay the live streaming for a few seconds to have time to moderate? We picked a direction and built it but your feedback here would be invaluable.

Comments URL: https://news.ycombinator.com/item?id=48445583

Points: 2

# Comments: 0

Article URL: https://batchphotoedit.com/

Comments URL: https://news.ycombinator.com/item?id=48445580

Points: 1

# Comments: 0

Article URL: https://www.cbc.ca/news/canada/facebook-overseas-alberta-separtism-9.7223966

Comments URL: https://news.ycombinator.com/item?id=48445562

Points: 2

# Comments: 1

Last week, Anthropic released https://github.com/anthropics/defending-code-reference-harne..., a reference harness for autonomous vulnerability discovery that uses Claude Code agents to find, verify, and patch memory-safety bugs. I wanted to use it but I only have access to GitHub Copilot.

This is a port of that harness to the GitHub Copilot CLI. PORTING-PLAN.md covers the decisions made to map the handful of features that work differently between Claude Code and the Copilot CLI.

The result is a working reference for anyone who wants to build autonomous security agents on Copilot, tracking Anthropic's approach as closely as possible.

Feedback welcome!

Comments URL: https://news.ycombinator.com/item?id=48445559

Points: 1

# Comments: 0