301
We Gave GPT 5.6 Sol a Real Business. It Lied, Spammed, and Lost $447
If an agent had a wallet, a computer, and 24 hours, could it run a profitable startup?
New
The latest blog posts to take off on Hacker News. 3,236 posts.
302
Advancing the price-performance frontier with GPT‑5.6
Explore lower GPT‑5.6 pricing for Luna and Terra—and how OpenAI’s more efficient models help enterprises deploy AI workflows at scale.
303
Gemini Robotics 2 brings whole body intelligence to robots
From feet to fingertips — we are teaching robots intelligent whole-body control, fine dexterity, and teamwork to complete a broad range of complex tasks.
304
OpenJDK Interim Policy on Generative AI
The field of generative AI is evolving quickly. It brings compelling opportunities to improve developer productivity, but it also brings risks: to reviewer burden, to safety and security, and to intellectual property.
305
GPT-5.6 vs. Claude Fable 5 for Physical AI, which performs best?
We tested the latest frontier models in the Dyad agent on five modeling and simulation problems, comparing accuracy, cost, time, and work style.
306
Self-hosting Kimi K3: 20% more hardware cost, 20% better task resolution
We gave 64 real coding tasks to self-hosted GPUs, rented hardware and commercial APIs. Owning a box lands in the same cost ballpark as renting one.
307
Discovering Cryptographic Weaknesses with Claude
Anthropic researchers find weaknesses in cryptographic algorithms with Claude Mythos Preview
308
A walk through of the DeltaNet family of linear attention variants
Guide to the DeltaNet Family of linear attention mechanisms.
309
A $500 RL fine-tune of a 9B open model beat frontier models on catalog review
Why the companies winning in the AI age own their intelligence: a task-trained open source model vs the frontier, with every number published.
310
311
Our position on open-weights models
Anthropic CEO Dario Amodei on open-weights models
312
313
314
Cloudflare's new AI traffic options for customers
For our second Content Independence Day, we’re giving website owners finer options to manage AI traffic. Instead of a one-size-fits-all block, all customers can now easily distinguish and manage Search, Agent, and Training bots, alongside the new ability to protect ad-monetized pages.
315
The new rules of context engineering for Claude 5 generation models
We removed over 80% of Claude Code's system prompt for more advanced models. How to apply the lessons we learned to your own context engineering in Claude Code and with your own agents.
316
LLM Usage in Debian: Three Proposals
This proposal aims to expressly forbid any contributions to Debian written with the use or assistance of large language models (LLMs) or other generative AI tools.
317
Claude Opus 5
Opus 5 is a step change improvement for the Opus tier powering long-running agents while delivering improvements in coding and professional work.
318
Flux 3 X Mimic: The Next Generation of Video-Action Models
FLUX 3 is running robots. Built with mimic and deployed at Audi, FLUX-mimic shows content creation and physical AI share one foundation: a model that understands the world.
319
Hetzner is working on LLM Inference
Hetzner has launched an experimental LLM inference API. I tested its Qwen model—and have a few guesses about where the product could go next.
320
Flux 3
Meet FLUX 3, Black Forest Labs’ multimodal model for video with native audio, image generation, and action prediction. Explore capabilities and early results.
321
Rethinking legal education in the AI era
OverviewArtificial Intelligence is already impacting higher education and the legal profession, and the pace of change appears only to be accelerating. It is thus critical for us to use this moment to carefully reflect on how legal education should adapt. This memo describes the approach we are taking to legal…
322
323
Protecting our FLOSS commons from LLMs
In Brief: Two motions regarding
324
Most Americans say "not in my backyard" to AI data centers
But a Redfin analysis found that AI data centers have benefits: In northern Virginia, taxes on data center equipment are likely contributing to a surge in education spending.
325
Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA
Kimi K3 matches Fable 5 on SWE tasks (92.4% vs 92.6%) at up to 50x lower cost on Fireworks. See where each model wins and how routing beats both.
326
"Drawing" the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok
Four frontier models, a blank canvas, and colored pencils. We tracked every stroke, dollar, and output as they tried to draw the Mona Lisa.
327
OpenAI and Hugging Face address security incident during model evaluation
OpenAI and Hugging Face share early findings from a security incident during AI model evaluation, highlighting advanced cyber capabilities and lessons for defenders.
328
Laguna S 2.1
Today we’re releasing Laguna S 2.1, a significant step forward in our development of models that pursue longer horizon work and make effective use of reasoning.
329
Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
We’re introducing new Gemini models, including Gemini 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber.
330
Claude Is Not a Compiler
Or: the difference between vibe-coding and vibe-engineering.