Videos

Meta's Muse Wonder models are competitive frontier models that score number one on the Deep Suite coding benchmark...

Higgsfield Ad Studio can generate 20 static ads and video variations from a single brand kit in minutes, but the...

Learning to build software is still worth it, but the focus shifts from writing code to 'intent engineering' –...

An AI gateway that uses identity-based networking can decouple LLMs, data connectors, interfaces, and sandboxes,...

AI can now decompile and recreate software from scratch using verifiable loops, reducing years-long decompilation...

The presenter builds a complete meeting notes system in Notion that uses AI to automate prep, transcription, action...

The central tension in AI-assisted development is that generation is cheap but evaluation is expensive, and the...

The video argues that the most effective AI builders focus on 'harness engineering' — building wrappers around...

WorkOS built an internal platform, Atlas, with a CLI called wow that lets any employee build, deploy, and share...

C2C (cache-to-cache) passes a sharer model's KV cache directly to a receiver model via a trained fuser, skipping...

A 300B parameter model can run on a 128GB mini PC using aggressive quantization to ~2-3 bits per weight, sparse...

Higgsfield's AI Influencer Studio allows creating two distinct AI characters from scratch and transferring their...

A builder is running Kimi K2, a 2.8 trillion parameter model, on 16 Nvidia DGX Spark boxes at about 21 tokens per...

VS Code moved from monthly to weekly releases by using AI to automate code review, triage, and error fixing, with...

Traversal's causal machine learning engine claims to achieve 80%+ root cause accuracy for high-severity incidents...

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

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

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

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

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

GEPA is a reflective optimization technique that achieves large performance gains from very few samples by using...

MiMo-V3 introduces two techniques—KV bridging and KV reuse—that shrink a million-token KV cache from 12.09 GB to...

Unreal Labs' Unreal Agent claims up to 40% lower inference cost than Codex on Terminal Bench 4.0 by changing the...

Jev is a new AI that makes thousands of fast, cheap decisions using structured JSON inputs, distinct from text...

Robotics has made negligible progress for 70 years because it has always been approached as a hardware problem...

Jev is a new class of model — system one / large programmable — designed for code consumption, not chat, optimized...

Jev is a decision-only model from Type Safe that outputs a fixed-choice answer with a confidence score instead of...

A 1998 StarCraft: Brood War benchmark called Brood War Bench shows frontier AI models lose badly in real-time...

Jev AI is a new model optimized for micro-decisions — yes/no, choice, or score — that claims to be 400x cheaper and...

Claude's new interface merges chat and co-work modes and adds Docs and Slides, making it a more self-contained...

A three-person security team used an AI model, Claude, to chain four known bugs into a full compromise of OpenAI's...

Anthropic merged Claude's chat and co-work modes into a single interface that automatically routes requests,...

Jev, a system one model from Typesafe, claims up to 200x speedup on decision tasks by selecting from predefined...

Codex's scheduled tasks eat your usage limit because they run inside chat threads. The workaround is to build...

A Claude command center that pulls real-time data from your connected apps is now buildable in about 15 minutes...

Cognition's SWE-2 is a post-trained version of Kimi K3 that improves agentic coding, matching GPT-4.5 on DeepSWE...

By combining ChatGPT 6 Astra's computer-use ability with Blender and Higsfield, a single prompt generated a fully...

Feeding an AI tool like Base44 thousands of existing UI screens from sites like Stripe and Apple lets it generate a...

AI models do not think like humans; they lack the abductive reasoning that humans use to invent explanations from...

GPT-6 Astra's long-context recall jumped from 73.8% to 96.3%, and its design judgment and data analysis have...

GPT-6 (Astra) can generate fully playable games like Fall Guys and Sim City from single prompts, and demonstrates...

The looped model Hugging (3.5B parameters, 32 laps) doubles its science score by running more laps, but its silent...

This video is a satirical fictional announcement of 'GPT-6 Astra' from OpenAI, parodying the hype cycle of AI model...

Cerebras CS4 runs GPT-4 at over 4,400 tokens per second, and the upcoming CS5 targets 10,000 TPS for medium models...

Claude Code's new /design command generates three design variants in parallel, with an inline editor for typography...

GLM 5.3 Flash, revealed as the mystery 'Ox-Alpha' model, delivers near-frontier intelligence at roughly 2–3% of the...

AI has made implementation free and fast for everyone, but the value has shifted to deciding what to build and why....

The presenter argues that Notion's unique structure—shared databases, skills libraries, and MCP connections—makes it...

Building a court-defensible AI fraud-detection system for the California DFPI forced the team to stop treating the...

Xiaomi's X-Ring O100 chip achieves 1.22 TB/s memory bandwidth by using wafer-on-wafer hybrid bonding at a 1.4 micron...