Videos

Grockbot is a platform for building coordinated AI agents, and the presenter shares five templates he uses to run...

Higsfield's API can be integrated into a custom app to generate images and videos, as demonstrated by a developer...

Claude Fable 5.1, combined with image and video generation from Hicksfield, can now rebuild an animated app like...

Jev, a model from Typesafe AI, is built exclusively for fast, cheap classification tasks and outputs only structured...

Anthropic's CEO Dario Amodei predicted that a one-person billion-dollar business could be created this year using...

Open source AI coding agents with local models can handle routine changes like adding a CSV export, but the lack of...

Prism ML's Bonsai 2 27B is a heavily quantized version of Qwen 3.8 27B that fits on a single GPU with 12-16 GB VRAM,...

Grok Bot's new voice mode lets users speak with individual AI agents, delegate tasks, and use computer use across...

Higgsfield launched an API for its video and image generation models, making it possible to use a pay-per-usage...

A 12GB GPU can run a real-time voice agent if processing stages are overlapped and memory is carefully budgeted. The...

The app market is seven times more crowded than four years ago (14,700 new subscription apps per month in Jan 2026...

David Ondrej describes running his business with GPT-6 Astra as an autonomous agency where AI agents handle tasks...

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

AI workloads are shifting from large transfers to small messages, making latency critical. TCP and RDMA suffer from...

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

Evolving a personal AI second brain into a team brain requires keeping personal agents separate and centralizing...

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

DeepSeek V4.1 Flash uses a Causal Encoder-Decoder architecture that halves prefill compute by processing input...

Sparkx 2.5 achieves a 1M token context window in a 4B parameter model by using 27 local attention layers (512 token...

AI underwriting is emerging as a new risk layer for AI adoption, with AIU raising $40 million to build insurance...

Multiscale indexing beats fixed chunking: AI21 found no single chunk size wins across queries, with an oracle...

Reinforcement learning can make search agents roughly 20 times faster and about 100 times cheaper than frontier...

Legora migrated from Elasticsearch and Postgres to Turbopuffer for project search and legal research, achieving an...

Docusign and NVIDIA built a purpose-built small VLM (Neotron Parse, ~850M parameters) that extracts complex tables...

Non-code knowledge work, unlike programming, lacks durable cues and well-defined tasks, making it significantly...

Agents almost never use llms.txt despite nearly 50% of websites publishing it—they go straight to docs or homepages...

Exa is building a search engine designed specifically for AI agents, arguing that AI systems will soon perform a...

BM25, a 30-year-old lexical scoring function, is making a strong comeback in agentic search because modern LLMs can...

Pinecone's Nexus is a persistent, specialized knowledge layer for enterprise agents, designed to solve the problem...

A GPU running a cold database query loses to a CPU because 97% of the time is spent moving data, not computing....

The RTX 5090 carries a $3,400 premium over MSRP due to FOMO and export demand, making it the worst value for local...

Grokbot can be wired to its own email inbox via Agent Mail and webhooks, letting it wake up and act on incoming...

Edge Zero's 2.9 GB memory claim for running Qwen 3.6 35B is technically accurate but misleading: it measures what...

Boris Cherny's viral advice to delete all Claude Code skills and rules is really about removing over-constraining...

Amit Desai's OUCH framework (Outcome User Cost Heuristic) shows that optimizing system decisions—act, confirm, or...

Voice AI failures are not random glitches but breakdowns in specific linguistic layers — sounds, words, interaction,...

EliseAI's cascaded voice agent architecture uses a streaming speculative transcriber (layering fast Flux on accurate...

Anthropic CEO Dario Amodei's essay calling for a slowdown in frontier AI development has united OpenAI's Sam Altman...

The real battle in production voice AI agents isn't the LLM—it's the transcript. State-of-the-art transcription...

Using GPT-6 Astra and the Seedance 2.5 (as well as GPT-image 2.5) integration with Hicksfield, a user can replicate...

Voice agents are moving from demos to production at scale, but reliability remains the top blocker, with even a 1%...

Tolan is a voice-first AI companion that routes every conversation turn through a small classifier called the tone...

Full-duplex speech-to-speech models solve the walkie-talkie problem of turn-taking, but they pay a fundamental...

Voice agents don't have to talk back. Charlie Guo from OpenAI argues speech is just one of three voice-agent modes,...

Google DeepMind is building a single native speech-to-speech model for Gemini that handles streaming translation,...

Safe Rust can still hand untrusted text to a shell because memory safety and application security answer different...

Using GPT-6 Astra and the Higsfield plugin, a user can generate editable After Effects animations from natural...

A lightweight screen control skill for coding agents, built entirely with CLI commands and no additional tooling,...

The video showcases seven GitHub repos and tools for use with ChatGPT and Claude, including a real-time global...

Anthropic's latest threat report reveals that seven Chinese labs, led by Alibaba, ran an industrial-scale...