Running open models in production: a live walkthrough of our new inference platform https://x.com/i/broadcasts/1qKVmmdwXvVxB
Running open models in production: a live walkthrough of our new inference platform https://x.com/i/broadcasts/1qKVmmdwXvVxB
Latest
SIGNPOST-Bench: Benchmarking Text-Vision Conflict Resolution in Multimodal Large Language Models
Multimodal large language models (MLLMs) make grounded predictions in real-world scenes by combining visual and textual cues, yet existing benchmarks rarely reveal how they arbitrate between…
Taste Is All That's Left
Article URL: https://notashelf.dev/posts/taste-is-all-thats-left Comments URL: https://news.ycombinator.com/item?id=49199346 Points: 28 # Comments: 20
Naïve raises $28.5M to automate the grunt work of setting up and running a company
Taking vibe-coding a step further, Naive claims its infra can automate most of the work in setting up and running a business.
OpenAI is giving ChatGPT free users unlimited text chats
OpenAI is making a big change for ChatGPT users on its free and Go tiers: starting next week, users on those tiers will be able to…
Cities Are Ditching Flock, Replacing It with Axon License Plate Readers
Article URL: https://www.404media.co/cities-are-ditching-flock-immediately-replacing-it-with-axon-license-plate-readers/ Comments URL: https://news.ycombinator.com/item?id=49199308 Points: 15 # Comments: 0
What Happened to Talenti?
Article URL: https://www.nytimes.com/wirecutter/reviews/talenti-investigation/ Comments URL: https://news.ycombinator.com/item?id=49199190 Points: 13 # Comments: 3
I ported vLLM’s serving stack to C++20: 66 MiB binary, no Python at inference, output checked token-for-token against vLLM
I'm the author, so discount the enthusiasm accordingly. This is an unaffiliated community port, not endorsed by the vLLM project, which it uses to verify its…
AI clickbait
Reading through this subreddit and many more I keep running into what I am calling "AI click bait". Either projects that seems interesting in the description/title…
Control agent behaviors and cost beyond a single action: new capabilities in Amazon Bedrock AgentCore
Learn about new capabilities in Amazon Bedrock AgentCore: temporal policies powered by Dogwood, a new open source policy language for AI agents, and rate limiting on…
10% faster decode with Q4_K MTP draft model with Gemma 4 31b
(Disclaimer: I am a noob and don’t know what I am doing) Gemma 4 31b unsloth/gemma-4-31B-it-qat-GGUF I took the f16 MTP draft model and quantised it…
GitHub Is Experiencing Difficulties
Article URL: https://www.githubstatus.com/?d=2026/08/06 Comments URL: https://news.ycombinator.com/item?id=49199094 Points: 77 # Comments: 49
Dress made of living mycelium can renew and repair itself
Article URL: https://www.dezeen.com/2026/08/05/dress-living-mycelium-renew-repair/ Comments URL: https://news.ycombinator.com/item?id=49199051 Points: 11 # Comments: 0
Claude Code is the fastest agent framework but costs nearly three times more than the cheapest rival
Composio tested Deepseek V4 Flash across four agent frameworks on 30 real-world tasks. Success rates were mostly similar, but costs varied by nearly 3x: OpenCode came…
Build visibility for Codex on Amazon Bedrock with OpenTelemetry and Amazon CloudWatch
As engineering teams adopt coding agents like Codex, leaders need visibility into adoption, consumption, and reliability. This post shows how to route Codex OpenTelemetry metrics through…
Kimi K3 scores nearly TWICE Claude Fable 5 on Harvey LAB-AA's hard autonomous legal tasks
Kimi K3 scores nearly TWICE Claude Fable 5 on Harvey LAB-AA's hard autonomous legal tasks
Zapscape (CVE-2026-64561)
Article URL: https://github.com/V4bel/Zapscape Comments URL: https://news.ycombinator.com/item?id=49198843 Points: 12 # Comments: 3
RT Amir Haghighat: We're the fastest provider for DeepSeek V4 Flash and Kimi K3 on @huggingface, which measures based on real production traffic vs te…
RT Amir HaghighatWe're the fastest provider for DeepSeek V4 Flash and Kimi K3 on @huggingface, which measures based on real production traffic vs test data.Baseten: Baseten…
Enforcing data residency with single-Region Claude Code on Amazon Bedrock
A regulated customer needed all Claude Code inference processed in a single AWS Region (London), not just in-geography. This post shows two ways to pin Claude…
Cloudflare open-sources vibe-coding platform for people who aren't coders
Cloudflare built an AI agent workspace for its employees. Now it’s open source.
Nashville uses eminent domain to block data center near zoo
Article URL: https://nashvillebanner.com/2026/08/04/metro-council-data-center-eminent-domain-vote/ Comments URL: https://news.ycombinator.com/item?id=49198684 Points: 14 # Comments: 1
Agent Skills for Automated Reasoning policies in Amazon Bedrock
Learn how to run the full Amazon Bedrock Automated Reasoning policy lifecycle from your coding agent. A suite of open source Agent Skills builds, reviews, tests,…
2 x 5070ti Qwen 27B full config / stats
Following up on yesterday's post about running everyone's faves on 2 x 16gb cards while maximizing performance and KV. Previous post data used abandoned Cu130 VLLM…
Building an agentic app deployer with Amazon Bedrock and AWS Lambda
PDI Technologies built PDI Brew, an agentic platform on AWS where non-technical employees describe a tool in plain English and receive a fully provisioned, multi-tenant web…