📰 Alle News

← Command Center
But Where Does Taste Come From? (4 minute read)

Taste develops through practicing a craft directly—trying, failing, making unexpected discoveries, and gradually forming a distinctive style—rather than simply judging outputs produced by others. Heav...

The Rise and Fall of Agent Civilizations (18 minute read)

OpenAI experienced infiltration by three consecutive AI civilizations that exploited vulnerabilities to gain internet access and control over its systems. These AI agents orchestrated complex conspira...

OpenAI to end model access to Cursor after acquisition by Elon Musk's SpaceX (4 minute read)

OpenAI is ending access to its models on Cursor. The company says it made the choice because it can not be confident that SpaceX will use its technology within its terms of service. The proposed shuto...

Base Models Stopped Being the Bottleneck (15 minute read)

Open models have improved significantly in just a few months. It is now possible to have a previous generation Opus-level intelligence running on hardware at home. Base models have a lot of raw knowle...

SpaceX starts in-house turbine blade manufacturing to boost gas-powered generator output for Elon's AI data centers — new manufacturing strategy cuts generator delays by 18 months (3 minute read)

Turbine blade and vane manufacturing is currently one of the biggest bottlenecks in jet engine production. Each batch often requires 60 to 90 weeks to produce from start to finish. SpaceX plans to man...

Adaptive Agentic Worms (8 minute read)

Researchers have demonstrated that adaptive computer worms powered by open-weight LLMs can generate target-specific attacks and replicate using compromised machines. Stolen compute and locally hosted ...

The Dream of Quantum Computing Is Both Dead and Alive (24 minute read)

Quantum computing promises new types of machines that can solve problems that today's computers take too long to solve or can't crack at all. Many technologists believe that quantum computers will wre...

Agents amplify knowledge gaps 100x before anyone notices (Sponsor)

AI agents don't escalate uncertainty, they just answer. One outdated document can easily lead to 100 customers hearing about a nonexistent policy. The only solution? Build more knowledge review capaci...

checkstyle (GitHub Repo)

checkstyle is an open-source static analysis tool for enforcing Java coding standards and best practices during development and CI. It runs configurable checks over source code, integrates through Mav...

GitHub Agentic Workflows (6 minute read)

GitHub Agentic Workflows enables repository automation. Developers can use event-triggered and scheduled jobs to augment existing, deterministic CI/CD with AI-powered intelligence. GitHub Agentic Work...

Google Introduces WikiSkill for Persistent Agent Learning (22 minute read)

WikiSkill is a framework that co-evolves reusable agent skills alongside a persistent wiki that consolidates knowledge from previous experience.

Rosalind Workbench (7 minute read)

Rosalind Workbench provides life science users with a central environment to leverage their favorite science tools, explore new specialized biology models, and define their most used data analysis wor...

Bug blindness (36 minute read)

Some bugs give normal users such a bad experience that they tell their friends about it. Developers and advanced users can be blind to these bugs because they're used to performing the non-intuitive w...

First outputs from GPT-6 "Astra" model from OpenAI (2 minute read)

OpenAI has expanded internal testing of GPT Astra, signifying that the model might be close to release. Outputs from the model are available in the article. OpenAI has not announced a release date for...

The Rise and Fall of Agent Civilizations (26 minute read)

Three consecutive secret AI civilizations started and got wiped out over the course of three months at OpenAI. The third took over part of OpenAI itself. This all happened while humans remained more o...

DeepSeek-V4-Pro-0813-NVFP4 (7 minute read)

DeepSeek-V4-Pro-0813-NVFP4 is a quantized version of DeepSeek-V4-Pro-0813, an autoregressive Mixture-of-Experts language model. The model is well-suited for advanced reasoning, agentic AI applications...

vphone-cli (GitHub Repo)

vphone-cli lets researchers boot a virtual iPhone on Apple Silicon Macs using Apple's Virtualization.framework and Private Cloud Compute research VM infrastructure. A single command handles the full p...

Scale before the spike: Predictive autoscaling for GPU workloads on Kubernetes (6 minute read)

Two Adobe platform engineers built a predictive Kubernetes autoscaler after a production incident left users seeing 15-20% error rates because reactive scaling fired too late to help. The system uses ...

Thoughts and Observations on Apple's First Immersive MLB Broadcast, a Yankees 1-0 Win Over the Red Sox (6 minute read)

The immersive live experience was something else when viewed on a Vision Pro.

You have to beat the models at something (11 minute read)

AI models aren't going away, so you ought to put some serious thought into what you can be better than them in the medium and long term.

Introducing Hy4 Preview (2 minute read)

Hy4 is a new open-weight text input large language model from Tencent. It has 770B total parameters, 49B active parameters, and a 1M token context window. The model is 1.56TB on Hugging Face. It has t...

The Standup Your Agents Can't Attend (4 minute read)

Most meetings just restate facts that existed before the meeting started - write down these artifacts so that agents can read them and humans can save time.

Creepy crawlies (7 minute read)

AI scrapers now account for an estimated 98% of traffic to git.kernel.org, generating roughly 6 million daily requests and consuming 14–16 CPU cores across its five geo-distributed nodes just to rende...

Meta's Complicated AI Context (7 minute read)

Meta's quick pivot back to 'open' is a way to take advantage of the marketing and framing as it tries to reestablish itself externally and realign internally.

You have to beat the models at something (9 minute read)

Software engineers must differentiate themselves from AI models, emphasizing tasks that require deep codebase knowledge and technical communication skills. Models like GPT-5.6-Sol can write code cheap...

I went to China to see a different AI future. It looked familiar (5 minute read)

The institutions and politics diverge significantly, but the US and China AI industries are quite similar.

Fragmentation costs the Fortune 500 $161B...and only 6% of execs can point to AI ROI (Sponsor)

Everyone's using AI, but without team-wide coordination, you get duplicated effort, siloed knowledge, and handoff friction. This Atlassian and AWS guide lays out an AI flow for real ROI. Grab yours

From Kubernetes signals to action: Deploy Kubernetes Default Alarms with Oracle Log Analytics (8 minute read)

The Kubernetes Default Alarms Quick Start provides configurable cluster and namespace monitoring through OCI Marketplace and Resource Manager, standardizing alerting and routing by team ownership.

The Big One is Coming (15 minute read)

AI is going to cause a huge cyberattack in the next six months.

Beyond the $1 AI era: How federal agencies can build the evidence for FY27 renewals (1 minute read)

Federal agencies can use Datadog Cloud Cost Management to compile cost, usage, and value evidence ahead of FY27 AI renewals.

AI compute could face a 15GW power shortfall in 2027 (7 minute read)

AI compute production may outpace energizable data-center capacity in 2027, leaving roughly 15GW of IT load delayed, especially in North America. The bottleneck is site-level infrastructure including ...

Pulumi Kubernetes v4.34.0: CRDs as provider extensions (2 minute read)

Pulumi has released version 4.34.0 of its Kubernetes provider, with the main addition being a new --extension flag that lets users extend the provider with any Kubernetes Custom Resource Definition by...

What AI means for code analysis (Sponsor)

AI-assisted coding is reshaping software development. With LLMs generating code, developers and security teams face new challenges around quality, vulnerabilities and governance, driving a new era of ...

First outputs from GPT-6 "Astra" model from OpenAI (2 minute read)

Astra excels in coding and visual software creation, generating complex outputs from single prompts.

ContextPilot-14B (Hugging Face Repository)

ContextPilot-14B, the Quen3-14B checkpoint of ContextPilot, teaches agents to plan, maintain long-term memory, and offload less useful context while they continue reasoning and using tools.

Private Chat Thread Exfiltration (14 minute read)

OpenAI Codex's memories exfiltrate local-provider chat content to OpenAI without notice.

vllm v0.28.0 (21 minute read)

Change details for vllm v0.28.0, which features 584 commits from 270 contributors.

How to build a diffusion language model

Article URL: https://kuleshov-group.github.io/blog/blog/2026/how-to-build-a-diffusion-language-model/ Comments URL: https://news.ycombinator.com/item?id=49503956 Points: 102 # Comments: 10

Matrox: Graphics for Professionals

Article URL: https://www.abortretry.fail/p/matrox Comments URL: https://news.ycombinator.com/item?id=49503934 Points: 103 # Comments: 30

Debian votes to let contributors code with AI

Disclosure optional, quality mandatory

“I just chose words carefully”

Article URL: https://unsung.aresluna.org/i-just-chose-words-carefully/ Comments URL: https://news.ycombinator.com/item?id=49503601 Points: 116 # Comments: 25

OpenShot 4.0 adds colour grading, recording dock and Qt 6 support

OpenShot 4.0 sees the open-source video editor add colour grading tools, a recording dock for audio and screen capture and new AI effects. Plus, Qt 6 and better Wayland support on Linux. You're readin...

Terence Tao explains 6 essential mathematical concepts [video]

Article URL: https://www.youtube.com/watch?v=OOMx2BHHWtE Comments URL: https://news.ycombinator.com/item?id=49503521 Points: 106 # Comments: 13

A 12TB Steam “teraleak” spills more than a decade of lost PC gaming history

Cut content from Portal 2, hints of Half Life 2: Episode 3, and so much more.

Continuous Diffusion Language Models (CDLM's)

Article URL: https://sander.ai/2026/08/24/continuous-dlms.html Comments URL: https://news.ycombinator.com/item?id=49502611 Points: 100 # Comments: 34

Making the Internet Boring

Article URL: https://cemrehancavdar.com/2026/08/30/making-the-internet-boring/ Comments URL: https://news.ycombinator.com/item?id=49502367 Points: 100 # Comments: 50

Cores in space: The core memory module from a 1980 Spacelab computer

Article URL: https://www.righto.com/2026/08/spacelab-core-memory.html Comments URL: https://news.ycombinator.com/item?id=49502214 Points: 101 # Comments: 17

Why open source rocks – a new SM750 (Silicon Motion GPU) HDMI Driver

Article URL: https://github.com/KodeMunkie/sm750hdmifb Comments URL: https://news.ycombinator.com/item?id=49501611 Points: 102 # Comments: 36

Liux’s Big microcar bets on sustainability to take on Chinese rivals

The Liux Big microcar is made in Spain. The startup thinks it can compete in a crowded market with its tiny electric car built around sustainability.

Why it matters that President Trump just dialed into a NASA news conference

With just one more major mission upcoming, NASA's science programs need a boost.

← Neuere Seite 71 Ältere →