Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. GPT-6 Sol and Luna(openai.com)
    339comments
  2. Claude Opus 5.5(anthropic.com)
    601comments
  3. OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005(cryptocellar.org)
    330comments
  4. Claude Opus 5.5 Intelligence, Performance and Price Analysis (Max)(artificialanalysis.ai)
    43comments
  5. WordPress: Unauthenticated path traversal leading to conditional RCE(github.com/wordpress)
    43comments
  6. Pentagon: Palantir AI Overreliance Led to Strike Killing 123 Iranian Children(bloomberg.com)
    53comments
  7. SAML: A Fractal of Bad Design(trailofbits.com)
    2comments
  8. Unreal Agent(unreallabs.ai)
    13comments
  9. There's a high chance of devices being sold with GrapheneOS preinstalled in 2027(grapheneos.social)
    71comments
  10. OpenAI is well positioned to fast-follow Jev(arcturus-labs.com)
    156comments
  11. 16-bit Intel 8088 chip (c. 1985)(allpoetry.com)
    11comments
  12. Launch HN: Coverage Cat (YC S22) – Umbrella insurance via your personal agent(coveragecat.com)
    20comments
  13. Writing Rust code that's fast by asking agents to make the code faster(minimaxir.com)
    37comments
  14. The human brain is two separate organs, study finds(stanford.edu)
    discuss
  15. Apple has added persistent 'ads' to iOS, and it's driving users crazy(techradar.com)
    362comments
  16. George Lucas Returns to Earth, Bearing Gifts(commonedge.org)
    10comments
  17. Obscura: The first VPN that can't log your activity(obscura.com)
    discuss
  18. Solitaire Alone Together(solitairealonetogether.com)
    25comments
  19. Show HN: Drop – A rootless Linux sandbox with gVisor support(droprun.sh)
    44comments
  20. Show HN: AI·rete·RAG – a Rete rule engine decides, RAG explains why(ai-rete-rag.com)
    2comments
  21. Can gzip be a language model?(nathan.rs)
    135comments
  22. A study of sequence weighting at scale(janestreet.com)
    discuss
  23. One Minute Park(oneminutepark.tv)
    5comments
  24. AMD's random number generator can't generate a 0?(flatassembler.net)
    171comments
  25. Explaining to business people why building software is still hard(manager.dev)
    3comments
  26. Truman World(trumanworld.live)
    31comments
  27. The Economics of Open-Weight Inference(ornn.com)
    25comments
  28. Relativistic raytracing(publish.obsidian.md)
    5comments
  29. Show HN: InstinctFlash – Run 5B world-action models in real time on Jetson Thor(github.com/general-instinct)
    2comments
  30. Training a model to identify AI-generated web content from structure alone(arxiv.org)
    2comments

Unreal Agent

26 pointsby 1h agounreallabs.ai
10 comments
48m agoHN ↗

I don't really understand how async tool calls translate to token savings

It says that it removes tokens wasted while a model is waiting on synchronous tool calls. What tokens exactly is Pi using when waiting?

23m agoHN ↗

I think a running agent periodically checks if a process it started has finished.

33m agoHN ↗

I think this is what I’ve been waiting for! Live interactions with models are fundamentally asynchronous! This will be great for interactivity.

What a good time for harness design. Just while OpenCode is growing up a bit and focusing on their harness. I’m delighted to see people focusing on good general solid harness design principles.

Someone make an OpenCode API (the best general agent end UI API I know) compatible server for it!

Sucks that I haven’t upgraded my client to V2 yet and I don’t really want to build on an outdated API . ..

25m agoHN ↗

The headline graph is kind of bizarre.

For some reason they're comparing their harness running on Astra xhigh to Codex with Astra max?

---

Also worth noting that OpenAI just added support for async tool calling to their harness, which isn't 1:1 with this approach, but is slowly ramping up in being able to provide something similar.

A big part of why Codex uses so many tokens is that it basically hot loops on polling tasks it starts for... absolutely no good reason: https://www.reddit.com/r/codex/comments/1wdlp7q/weve_discove...

I fixed it on my fork of Codex too, also back in Jan/Feb – I keep this patch rebased, for anyone who wants it: https://github.com/tekacs/codex/commit/9ffcf8db9078eae43d411...

It results in token savings similar in scale to those displayed here by Unreal.

---

My harness has used a slightly fancier version of the approach that Unreal is using since ~Feb, and... it definitely works excellently, but it's also assuredly smoother with Astra and other recent models that are more aware of async tool calling.

22m agoHN ↗

Honestly the fact that we're still modeling agent harnesses like chat and strapping them into shell sessions instead of building async actor systems with sandboxed OS functionality access actors is bananas to me. So much easy wins can be had just by building on right abstractions... Hopefully will get enough time to play with this idea soon on my own.

21m agoHN ↗

Sounds like a trademark issue when Epic ships a wildly popular Unreal Engine

19m agoHN ↗

But I don't think you can trademark a generic word like Unreal...

10m agoHN ↗

Mildly disappointed that this has nothing to do with Unreal Engine, the popular game engine.

9m agoHN ↗

omp.sh does this way better by just allowing structural toolcall execution in eval with python/js.