Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A developer on Hacker News is questioning the effectiveness of Spec Driven Development (SDD) with LLM coding agents, citing issues with code style and hygiene despite using tools like OpenSpec. The developer seeks advice…
The traditional coding interview is dead as AI tools make cheating trivial, forcing a shift from hiring coders to hiring engineers who can design algorithms, review code, and debug. The author proposes five new pillars f…
A price war is erupting in the AI industry as companies face soaring inference costs, with startups like Swan AI paying $113,000 monthly for a four-person team. Businesses are dynamically routing queries across multiple …
A developer outlines a practical playbook for AI fluency among software engineers, emphasizing that effective AI use goes beyond prompt engineering to include clear context, iterative refinement, and control over quality…
LMCache introduces a novel KV cache optimization layer to accelerate LLM inference, enabling faster local deployment on consumer hardware. AllenAI releases olmo-eval, a workbench for evaluating open language models throu…
A developer argues that the era of explicit prompting for AI is ending, replaced by autonomous agents like loopers and robovacs that operate without human input. The post describes a shift from manual prompt engineering …
Google has started rolling out Search agents in AI Mode for AI Ultra subscribers, enabling 24/7 background monitoring of web content to deliver synthesized updates on user-specified topics, such as new apartment listings…
Meta's AI unit is in chaos as executives and employees disagree on strategy, including whether to keep AI models open-source or shift to proprietary systems. The company is moving 7,000 employees into AI roles amid restr…
Google has unveiled DiffusionGemma, an experimental open model that generates text up to 4x faster than traditional auto-regressive models by processing entire blocks of text simultaneously using diffusion techniques. Th…
Five AI models and one human fan were tasked with predicting every group-stage scoreline of the 2026 World Cup before kickoff. Claude Sonnet 4.6 from Anthropic scored the highest with 9 points, while GPT-5.5 and Gemini 3…
A new study reveals that large language models generate recurring fictional character ensembles, such as Elena Vasquez and Marcus Chen, who appear as experts across thousands of AI-generated documents. The research docum…
NVIDIA achieved leading agentic coding performance on the first agentic AI benchmark, AA-AgentPerf, delivering up to 20x better performance than previous generations. The benchmark, created by Artificial Analysis, measur…
A developer introduced Mechanically-Falsifiable Independent Control (MFIC), a discipline for enforced tests and runtime checks that prevent AI agents and human developers from being confidently and silently wrong. MFIC r…
A developer argues that the developer community must prioritize building large language models free from corporate and state control, warning that companies like Anthropic propose licensing regimes that would create a ne…
Google Research announced Gemini-SQL2, a text-to-SQL capability powered by Gemini 3.1 Pro, which achieved 80.04% execution accuracy on the BIRD Single-Model Leaderboard, surpassing its predecessor Gemini-SQL and competit…
NVIDIA's Blackwell Ultra NVL72 platform achieved top performance on the first agentic AI benchmark, AgentPerf from Artificial Analysis, running up to 20 times more agents per megawatt than the previous Hopper architectur…
RuntimeWire's weekly report shows a viral spike in readership after a head-to-head comparison of DeepSeek V4 Pro and GPT 5.5 Pro went viral on Reddit and Hacker News, driving weekly reads from 4,500 to 28,015, a 523% inc…
Hacker News user franze created a thread for AI-coded games, showcasing projects like sphere chess, Squishy and Friends, Strait Sweeper, Bello, and 3D Tik Tak Toe, built using models like Claude and ChatGPT. The thread h…
Samsung Electronics is lifting its three-year ban on external generative AI services, allowing employees in its DX Division to use ChatGPT, Gemini, and Claude. The reversal follows Chairman Lee Jae-yong's order for an AI…
Rocket Close, a Detroit-based title agency and appraisal management company within Rocket Companies, launched Supercharger, an agentic AI solution built with AWS to optimize title operations. The system uses Strands Agen…