Hacker News Front Page

Article URL: https://edjefferson.com/localwiki/

Comments URL: https://news.ycombinator.com/item?id=50202285

Points: 963

# Comments: 187

Article URL: https://www.anthropic.com/research/formal-verification

Comments URL: https://news.ycombinator.com/item?id=50202217

Points: 420

# Comments: 258

Hey HN, we recently open sourced our Consumer Agent Testing Tool to help teams iterate against proprietary assistants like Muse and Instinct. We built this to reduce our own frustrations with building an MCP/API for these agents to consume.

The problem: testing against these tools is hard! Agent memory + account statefulness make reproducibility, in the ways required for repeatable/reliable outcomes, between agents and the APIs/MCPs they interact with very challenging.

The solution: CATT lets devs test their services against a surrogate Muse or Instinct agent (not affiliated with either of these companies). Testers define starting conditions, approvals, and success conditions for their service's state. CATT then records the full interaction (incl. tool calls, responses) so it can diff what happened from what the agent reports. Runs can be tested in local builds or via CI, with or without LLM backing.

CATT is designed to emulate the important parts of personal assistant interactions. Its got:

- Tool calls and command-line tool access - Simulated approvals for service actions - Outcomes tracking and comparison - Optional MCP profile based on an unofficial Muse template

There are limitations to how much we know about different assistant specific behaviors so the tool labels its assumptions as `documented`, `unofficial`, `assumed` or `unknown` for developer review. We'll do our best to keep the tool updated as new insights about the internal workings of these assistants come to light. It is an area of active research and interest, a good example is this recent report on how Instinct uses Exa for web search [0].

Are you building a connector or MCP? If yes, we'd like to hear from you! Have you had assistants behave incoherently, use stale information, or procalim success without evidence? What else should we add tests for?

The best way to get started is...to point your coding assistant at the repo and ask it to integrate the tool with your CI or local testing pipeline!

Hope it's helpful!

[0]: https://www.sunbeam.ai/thoughts/instinct-is-using-exa-for-we...

Comments URL: https://news.ycombinator.com/item?id=50201860

Points: 4

# Comments: 0

Article URL: https://www.wolframcollatzmahler.com/rule30/index.html

Comments URL: https://news.ycombinator.com/item?id=50201842

Points: 15

# Comments: 3

Article URL: https://www.lesswrong.com/posts/WHEQps5PK8LrrcTAb/how-does-fable-5-1-hide-messages-in-images

Comments URL: https://news.ycombinator.com/item?id=50201445

Points: 177

# Comments: 56

Article URL: https://train.samismith.com/

Comments URL: https://news.ycombinator.com/item?id=50201414

Points: 107

# Comments: 26

Article URL: https://www.vanityfair.com/story/jared-kushner-hacked-phone-secret-negotiations

Comments URL: https://news.ycombinator.com/item?id=50201259

Points: 315

# Comments: 126

Title is a bit of an exaggeration but the question is more: if the frontier models are so good, why even sell them as your primary money maker?

Couldn't OpenAI or Anthropic make a lot more money by just starting companies themselves that leverage the models and building their own products?

Is selling tokens really the best strategy?

Especially with commodification and prices declining wouldn't that indicate there is more money to be made on the application-side of the models rather than the utility side

Comments URL: https://news.ycombinator.com/item?id=50200997

Points: 404

# Comments: 439

Article URL: https://resources.arc.net/hc/en-us/articles/20498285812375-Release-Notes

Comments URL: https://news.ycombinator.com/item?id=50200947

Points: 9

# Comments: 6

Article URL: https://developer.chrome.com/blog/html-in-canvas-origin-trial

Comments URL: https://news.ycombinator.com/item?id=50200164

Points: 110

# Comments: 49

Article URL: https://mastodon.catgirl.cloud/@49016/117475976434068995

Comments URL: https://news.ycombinator.com/item?id=50199757

Points: 128

# Comments: 37

Article URL: https://github.com/maxilevi/project-hedra/tree/pre-ai

Comments URL: https://news.ycombinator.com/item?id=50199389

Points: 213

# Comments: 41

Article URL: https://3d.camilleworld.com/play/

Comments URL: https://news.ycombinator.com/item?id=50199349

Points: 11

# Comments: 2

Article URL: https://www.typeui.sh/micro-interactions

Comments URL: https://news.ycombinator.com/item?id=50199290

Points: 67

# Comments: 37

Article URL: https://jamesyu.org/award-winning-ai-literature/

Comments URL: https://news.ycombinator.com/item?id=50199232

Points: 48

# Comments: 51

Article URL: https://arxiv.org/abs/2610.27251

Comments URL: https://news.ycombinator.com/item?id=50199230

Points: 128

# Comments: 74

Article URL: https://www.biograph.com/newsroom/biograph-completes-first-u.s.-study-with-caristo-diagnostics%E2%80%99-ai-powered-coronary-inflammation-detection-finding-risk-in-over-42-of-asymptomatic-patients

Comments URL: https://news.ycombinator.com/item?id=50198940

Points: 72

# Comments: 37

Article URL: https://github.com/rbartoli/agent-usage-guard

Comments URL: https://news.ycombinator.com/item?id=50198936

Points: 19

# Comments: 7

Article URL: https://github.com/NVIDIA/OpenShell

Comments URL: https://news.ycombinator.com/item?id=50198675

Points: 471

# Comments: 205

Article URL: https://bernsteinbear.com/blog/object-shapes/

Comments URL: https://news.ycombinator.com/item?id=50197692

Points: 11

# Comments: 1