Videos

Claude Code's new mods feature allows users to extend the desktop app with custom plugins created using natural...

Meticulous provides exhaustive front-end verification by recording real user flows and diffing screenshots on every...

A browser agent with 99% per-step accuracy achieves only 36% overall success on a 100-step task because cost...

Claude Code's new mod system lets developers watch, change, or block events in the terminal, enabling dynamic model...

Kimchi's harness automatically selects the cheapest LLM per task while maintaining output quality, achieving 2.5x...

NVIDIA is releasing a 64 GB DGX Spark at $4,999, halving the VRAM of the original 128 GB model while keeping the...

Dave Ablar, a 10-year Python veteran, now writes zero code by hand — everything goes through Cursor or Claude Code....

DeepSeek's DeepGEMM-Ascend, announced September 30, 2026, ports a CUDA kernel library to Huawei Ascend chips while...

Switching matrix operations from Nvidia GPUs to Huawei's Ascend requires far more than a chip swap — the data...

Exa's semantic search is designed to replace the ten blue links of Google with context-rich highlights that are...

Vercel's CEO reveals his personal agentic engineering workflow built around FX, a 6MB CLI that boots as fast as a...

Small open Jev image models can automate decision nodes in RPA workflows by answering yes/no questions from raw...

MCP's main problem—context bloat from loading all tools upfront—is a flaw in agent harness design, not the protocol...

A community port of Splash Inco's engine roughly doubles inference speed on 2021 M1 Max Macs for the Quinn 3.8 27B...

The AI industry's push for memory features reveals that the real product is continuity, not model intelligence....

Semi Analysis’s ClusterMAX 3.0 reveals that most NeoClouds have terrible cybersecurity, with issues like outdated...

Gemini 4 Argon's 77.9% on Deep SWE would put it ahead of all other models, but the benchmark is public and...

Recursive Language Models (RLMs) are a harness design where the only tool is code, enabling compositionality and...

OpenAI and Synopsys are developing GPT-Synopsys, an AI agent that iteratively modifies chip designs, runs...

Opus 5.5 via Abacus AI can build a functional animated app in one shot using a three-step system: provide design...

Alibaba's Zhenru 900 beats Nvidia's H200 on paper in exactly two published specs, 216GB memory versus 141GB and a...

Codex's new ultrafast mode delivers real speed, but only small tasks justify its usage cost. It runs at up to 300...

Claude Sonnet 5.5 outperformed OpenAI's GPT-6.1 Sol across all four real-world tests, with the largest gaps in...

Generic AI output is almost always a clarity and destination problem, not a model problem: the presenter's five-move...

Darwin 27B RSI, a model from Final Bench and Vidraft, claims a 5.24-point gain on GPQA Diamond through recursive...

The 'latent shopper problem' is the key challenge for enterprise AI: users often don't know what they're looking...

Gemini 4 Argon can output a million tokens in a single response, roughly eight times more than other frontier...

Gemini for Argon scored 77.9% on the Deepsw SWE benchmark, beating reported scores from OpenAI and Anthropic, but...

Google's MVP satellite, carrying four TPUs and powered by 1 kW of solar, can only operate in 15-minute bursts...

ChatGPT Spaces is OpenAI's new project-hub feature that organizes chats, pages, and AI agents like Codex into...

Jev, a system one decision model, is 20-200x faster and 40-1000x cheaper than LLMs for making decisions, making it...

ChatGPT's Dots ($100/month) and GrokBot ($20/month) are model-locked personal assistants that prioritize ease of use...

OpenAI's computer use is now faster than average humans at most tasks, with the next frontier being superhuman...

OpenAI's new GPT-6.1 Soul model cuts the cost of a 10M/2M token job from $200 to $40, and OpenAI claims it matches...

OpenAI's Dots feels like a rushed response to Meta's Muse, lacking the polished interface and user-friendly features...

AI coding assistants are making code feel more maintainable while simultaneously eroding engineers' confidence in...

The Chief AI Officer role is still a moving target, varying by company type and AI maturity, but at WSO2 it breaks...

Agent-generated code passes tests far more often than it deserves to merge: SWE-bench's 'good enough to merge'...

For builders, OpenAI DevDay's real news is pricing: GPT-6.1 Soul is claimed to deliver near-Astra quality for...

Band's core claim is that today's agents are in 'digital solitary confinement': MCP and A2A let agents call tools or...

Model listings show only weight file size, not the full memory bill: KV cache grows with context and can match the...

DeepSeek's DSEC platform sustains over 380,000 concurrent AI agent sandboxes with most using minimal CPU, thanks to...

vLLM's paged attention mechanism, which treats KV cache like virtual memory pages, directly solved the 60-80% memory...

OpenAI released GPT6.1 Soul, a model that matches the performance of its frontier Astra model at a fraction of the...

OpenAI's GPT-6.1 Sol is priced at $2 per million input and $10 per million output tokens, which the company says...

ChatGPT Spaces, a new feature announced at OpenAI DevDay, aims to compete with Google Docs by embedding AI agents...

OpenAI's DevDay introduced Dots, a new personal AI assistant that works across text, phone, email, and Slack,...

Postgres 19 is removing several features before release — property graphs, online checksums, FOR PORTION OF,...

Claude Opus 5.5 can build a functional iOS app, landing page, and motion graphics video from a single prompt, with...