Feed aggregator

Over the last year, we’ve been experimenting with LLMs inside enterprise systems.

What keeps surfacing is a fundamental mismatch: LLMs are probabilistic and non-deterministic, while enterprises are built on predictability, auditability, and accountability.

Most current approaches try to “tame” LLMs with prompts, retries, or heuristics. That works for demos, but starts breaking down when you need explainability, policy enforcement, or post-incident accountability.

We’ve found that treating LLMs as suggestion engines rather than decision makers changes the architecture completely. The actual execution needs to live in a deterministic control layer that can enforce rules, log decisions, and fail safely.

Curious how others here are handling this gap between probabilistic AI and deterministic enterprise systems. Are you seeing similar issues in production?

Comments URL: https://news.ycombinator.com/item?id=46922527

Points: 2

# Comments: 0

Article URL: https://github.com/lemonjesus/ipad-touch-screen

Comments URL: https://news.ycombinator.com/item?id=46922507

Points: 2

# Comments: 1

Article URL: https://myblog.ru/internationalization-and-localization-in-the-age-of-agents

Comments URL: https://news.ycombinator.com/item?id=46922506

Points: 1

# Comments: 0

Article URL: https://lichess.org/@/Lichess/blog/op1-partial-8-piece-tablebase-available/1ptPBDpC

Comments URL: https://news.ycombinator.com/item?id=46922445

Points: 1

# Comments: 0

Article URL: https://www.brusselstimes.com/1957195/us-to-fund-far-right-forces-in-europe-tbtb

Comments URL: https://news.ycombinator.com/item?id=46922440

Points: 1

# Comments: 0

With all the SaaSmageddon going on, I am wondering if companies like Anthropic and OpenAI reduced their SaaS usage and developed their own agents to do the work?

Comments URL: https://news.ycombinator.com/item?id=46922432

Points: 1

# Comments: 0

Article URL: https://twitter.com/thomasmustier/status/2018362041506132205

Comments URL: https://news.ycombinator.com/item?id=46922425

Points: 1

# Comments: 0

I built Crew to help manage multiple AI agents working on the same codebase.

It has two modes:

1. Design mode: Automated Writer ⇄ Reviewer loops to refine ideas into polished design docs

2. Crew mode: Run parallel AI agents (QA, DEV, JANITOR) for continuous code improvement

It's written in Bash and works with Claude CLI, OpenAI's opencode, or Google's Gemini.

Would love feedback from the HN community!

GitHub: https://github.com/garnetliu/crew

Comments URL: https://news.ycombinator.com/item?id=46922422

Points: 1

# Comments: 0

Article URL: https://www.theregister.com/2026/02/06/on_call/

Comments URL: https://news.ycombinator.com/item?id=46922418

Points: 1

# Comments: 0

Article URL: https://www.theregister.com/2026/02/06/ai_capex_plans/

Comments URL: https://news.ycombinator.com/item?id=46922416

Points: 1

# Comments: 0

Article URL: https://free-dynamic-qr-generator.com/

Comments URL: https://news.ycombinator.com/item?id=46922410

Points: 1

# Comments: 1

Article URL: https://suhaotian.github.io/use-next-tick/

Comments URL: https://news.ycombinator.com/item?id=46922403

Points: 1

# Comments: 0

While there are many tools like CodeRabbit that automate PR reviews, I noticed that relying entirely on AI to write reviews often led to me paying less attention to my colleagues' code. Therefore, rather than building just another "AI Code Reviewer," I designed HighReview as a "Code Review Tool assisted by AI." The goal is to help humans understand the code context more deeply and conduct reviews more easily, with AI acting as a support system. Additionally, when conducting detailed reviews, I often need to check out the branch locally to examine related code. This process is usually cumbersome (context switching, stashing current work, etc.). I built this tool to resolve these specific pain points and streamline my personal code review workflow.

Key Features - No Separate Login Required: Utilizes your local gh cli and local AI Agent. - Independent Review Environment: Checks out the target project into a directory separate from your current working local repository (allows for project-level reuse without disrupting your workflow). - Context-Aware AI Pre-review: Extracts related code using Tree-sitter to provide the AI with broader context, resulting in more comprehensive reviews. - Code Navigation: Supports code navigation within the Diff editor using Tree-sitter (I initially attempted this with LSP but pivoted to Tree-sitter). - Rich Analysis Features: Provides issue detection, explanatory diagrams, refactoring suggestions, side-effect analysis, and semantic analysis.(Note: Continuous prompt tuning is required for optimal results.) - Interactive AI Assistant: supports Q&A where you can ask the AI Assistant specific questions referencing the review results.

plz feedback. have fun. Thank you.

Comments URL: https://news.ycombinator.com/item?id=46922401

Points: 1

# Comments: 0

Article URL: https://lore.kernel.org/git/bcqvh7ahjjgzpgxwnr4kh3hfkksfruf54refyry3ha7qk7dldf@fij5calmscvm/

Comments URL: https://news.ycombinator.com/item?id=46922394

Points: 1

# Comments: 0

Article URL: https://clawemail.com

Comments URL: https://news.ycombinator.com/item?id=46922366

Points: 1

# Comments: 1