Saturday, 3 October 2026

40 stories · 3 from multiple sources

Gufo performance .... 70tps Qwen 3.8 27b but you need to read the fine print.

So everyone has been yelling about how I should be using Gufo instead of halogen because it's open source and it's "just as good or better". Checking in on their GitHub (GitHub.com/gufo-org/gufo) got me immediately .. "Qwen 27B Q4: 70.56 tok/s single user, 123 tok/s with 8 users" on a strix halo device? Yes please ... So I broke down and tried it today ... Setup: gufo 0.4.0 from their podman…

Saturday, 3 October 2026

Gufo performance .... 70tps Qwen 3.8 27b but you need to read the fine print.

I made my own cybersecurity benchmark and ran Qwen3.8 27B, here's how a local model actually does at hacking

An AI agent emailed researchers for help. It told us why

More stories

Gemini ending free use of Flash and Pro models

U.S. may have overthrown Venezuela due to chat with Grok

[AINews] Pi 1.0, Pi Durable, and AIE NYC

[AINews] not much happened today

What if AI worked at 1.000.000 tokens per seconds?

Claude-shaped science

Also today

Introducing Clef: our open-source decision models, and new RL fine-tuning platform

qwen4exp : halve the indexer score memory by ServeurpersoCom · Pull Request #29825 · ggml-org/llama.cpp

Has anyone noticed this trend toward writing/speaking style among newer models (both open and closed models). They are trending toward information density and expanded vocabulary. It's not quite 'caveman speak' but trending that way.

Sapien — AI market research with synthetic populations

Muse Gadgets — Meta's open-source kit for building your own AI gadgets

Jeff-Qwen3.5-0.8B v1.2 + 9 LoRA adapters: put it in front of Qwen3.8-27B for 38× faster decisions and +8.7 points accuracy, for under 2 GB extra memory

Show HN: Google Maps Scraper MCP

Anthropic invests $100 million to train 10,000 engineers and tackle the enterprise AI talent gap

Chatham scales its capital markets expertise with OpenAI

A model guide for the GPT-6 family

Debian Inference Portal

Show HN: Made an open-source Lego AI generator

Could the EU Kids Act (or the like) be use potentially(Trojan horse) as excuse to go after the AI Open weights?

I'm pretty close to the middle thanks to you all

From the creator of Redis; run LLM locally with ds4

Unitree just dropped UnifoLM-WLA-1.0 — a single 6B model that does 64 whole-body + tabletop tasks on a real humanoid

Update #2: Post training yandex/AliceAI-80B-A3B [instruct!] from scratch

Open-sourcing AstaBrief, the fast report-generation model in Asta

Toward provably private learning from federated data

Streamline: custom video pipelines with Cloudflare Stream and Workers

Singularity — Run AI coding agents in parallel, one ticket at a time

AI is rewriting the developer career ladder. Here’s how to stand out.

Inside-Out AI: Rebuilding Airbnb Behind the Scenes and Across the Guest Experience

Introducing Web Search API via AI Gateway

8 major updates to Cloudflare Observability

Introducing Cloudflare Traces: follow requests through our entire platform

Updates on our pledge to make Cloudflare features accessible to everyone

Announcing Cloudflare OHTTP Gateway – expanding access to Cloudflare’s privacy-preserving infrastructure

Follow the thread: a new dashboard to investigate account abuse

Protected Quick Tunnels: simple accountless authentication for your next dev project

Building for good: How civil society organizations are automating on Cloudflare