Episodi

  • The Latest in AI — X Evening Edition (October 26, 2026)

    The Latest in AI — X Evening Edition (October 26, 2026)
    Oct 27 2026
    20 new developments in AI spanning frontier research, infrastructure financing, and local AI execution: Cloudflare’s rapid patch of a GLM-5.3 KV-cache leak, Broadcom’s $63B SPV airline-style compute financing, OpenAI’s Lean-verified math breakthrough and GPT-6.1 Terra release, CodeRabbit deploying GPT-6.1 Sol for autonomous security loops, Apple’s iOS 27.2 Siri overhaul, Tencent’s open-source Hy-SimulMT-30B speech translation, and new benchmarks running ternary 30B models and 400 tok/s clusters on legacy GPUs. Show Notes / Topics Covered: Infrastructure & Financing: Wall Street banks structure Broadcom’s $63B compute facility as aircraft-style leases; Verizon anchors enterprise AI fabrics; and helium joins the US Critical Materials list as a vital fab dependency. Security & Incident Response: Cloudflare exposes and patches a GLM-5.3 multi-tenant KV-cache leak in 15 hours; CodeRabbit shifts security from static scanners to live interactive fix loops with GPT-6.1 Sol; Greptile gives GPT-6 Astra browser tools for live bug hunts. Frontier Models & Math: OpenAI refutes the Exponential Time Hypothesis with Lean formal verification, debuts Codex Ultrafast for infrastructure debugging, and launches GPT-6.1 Terra to halve mid-tier pricing. Local & Edge Compute: Hobbyists hit 400 tok/s on triple V100s; ternary Nemotron 30B unlocks 368k context on 12GB GPUs; Embedl runs Qwen3.8-27B at 40 tok/s on Jetson Thor; and Hob launches an offline desktop agent running Gemma 4 and gpt-oss privately.
    Mostra di più Mostra meno
    33 min
  • Anthropic's IPO Filing, Meta's Financing Plans, and AI-RAN

    Anthropic's IPO Filing, Meta's Financing Plans, and AI-RAN
    Oct 26 2026
    We examine the reality behind AI's massive capital demands, breaking down Anthropic's proposed IPO filing numbers and Meta's expansive multi-stage financing plan. Plus, a look at recent telecom milestones from Nvidia, Nokia, and mobile operators testing AI-RAN infrastructure.
    Mostra di più Mostra meno
    28 min
  • The $95 Billion Paradox: Why AI Gets Cheaper as the Stakes Get Bigger | Latest in AI — October 26, 2026 (Morning Edition)

    The $95 Billion Paradox: Why AI Gets Cheaper as the Stakes Get Bigger | Latest in AI — October 26, 2026 (Morning Edition)
    Oct 26 2026
    Dante Martin’s AI voice clone and Jenn Xeno trace the money, hardware and trust constraints behind 17 AI developments: Anthropic’s IPO filing, Meta’s proposed $95B financing plan, the economics of cheaper AI, TSMC and European AI infrastructure, AKV agent caching, Tencent translation, live 5G optimization, Toyota robotics, autonomous driving, AI hurricane forecasts, Amazon Bee, cybersecurity and AI copyright. Source links and editorial corrections are documented in the accompanying production record.
    Mostra di più Mostra meno
    56 min
  • Latest in AI — Oct 25, 2026 (Midday Edition)

    Latest in AI — Oct 25, 2026 (Midday Edition)
    Oct 25 2026
    Dante Martin and Jenn Xeno discuss how physical infrastructure constraints, soaring data center delay penalties, and specialized insurance policies are challenging the AI boom.
    Mostra di più Mostra meno
    27 min
  • Latest in AI — Oct 25, 2026 (Morning Edition)

    Latest in AI — Oct 25, 2026 (Morning Edition)
    Oct 25 2026
    NVIDIA’s open-source Flora engine cuts robot video-labeling costs from about $90 to 72 cents per hour, while specialized models focus on the physical details robots need to act. The episode also examines security challenges as AI-generated code scales, a 33-billion-parameter model’s agentic coding performance, and AI moving onto everyday devices—from voice synthesizers to smart TVs.
    Mostra di più Mostra meno
    26 min
  • Latest in AI — Oct 24, 2026 (Evening Edition)

    Latest in AI — Oct 24, 2026 (Evening Edition)
    Oct 25 2026
    On the October 24, 2026 Evening Edition, Dante Martin’s Voice Clone and Jenn Xeno explore how cheaper AI is shifting the power to build—from OpenAI’s reported deal with Replit to AI-assisted software development, rigorous automated testing, Postgres bug discoveries, and the new zero-copy language Goose.
    Mostra di più Mostra meno
    28 min
  • Latest in AI — Oct 24, 2026 (Afternoon Edition)

    Latest in AI — Oct 24, 2026 (Afternoon Edition)
    Oct 24 2026
    In this Afternoon Edition, we explore AI’s move from chat windows into the physical world: Dream Labs’ world action model for unfamiliar robots and Axon’s AI-enabled body cameras. We also examine the limits, risks, and human oversight shaping AI as it enters high-stakes environments.
    Mostra di più Mostra meno
    33 min
  • Latest in AI — Oct 24, 2026 (Morning Edition)

    Latest in AI — Oct 24, 2026 (Morning Edition)
    Oct 24 2026
    Three local-AI breakthroughs challenge the idea that frontier capabilities require massive data centers: a 35B model aces coding tasks on 16GB of VRAM, Qwen 3 27B matches a cloud flagship on a Korean document-retrieval benchmark, and a new quantization approach squeezes a huge model onto a 12GB GPU. The episode explores how smaller models and hardware optimizations could bring powerful AI to everyday computers.
    Mostra di più Mostra meno
    33 min