Listen to today’s podcast: https://www.youtube.com/channel/UC-nqwUyvLDEvs7bV985k-gQ
AI Daily Podcast 12/27/2025
Welcome back! Today’s podcast episode was created from the following stories: a mix of breakthroughs in AI infrastructure, shifting talent and money, and how these tools are shaping everyday work and life.
TurboDiffusion: 100–200× Acceleration for Video Diffusion Models
Author: Not specified | Date: Not provided | Source: GitHub repository
TurboDiffusion is an open-source framework that claims 100–200× end-to-end speedups for video diffusion on a single RTX 5090, using SageAttention/SLA for faster attention and rCM for timestep distillation. It ships quantized and unquantized checkpoints for T2V and I2V models at 480p–720p, plus ComfyUI integration and training code. While the paper and checkpoints are still evolving, the promise of near-real-time video generation could reshape creative workflows.
More than 50 people became AI billionaires in 2025
Author: Mike Pearl | Date: December 26, 2025 | Source: Gizmodo
Forbes tallied more than 50 new AI billionaires this year, driven largely by enterprise tools, data labeling, and cost-cutting “agent” platforms, with a few splashier consumer exceptions like ElevenLabs. The piece highlights headline-making moments (from Lucy Guo’s brief youngest self-made woman billionaire status to 22-year-old founders at Mercor) alongside stark economic context: affordability challenges and concentrated consumer spending. It’s a snapshot of how AI’s wealth boom is colliding with broader inequality.
Codex vs. Claude Code (Today)
Author: Joe Fabisevich | Date: December 22, 2025 | Source: build.ms
Fabisevich contrasts OpenAI’s Codex and Anthropic’s Claude Code as two superhuman coding companions with different working styles. Codex shines for hands-off, long-running tasks that yield high-quality results with less steering, while Claude Code excels for engineers who prefer interactive tooling, knobs to turn, and Plan Mode. The takeaway: pick the agent that matches your workflow, then pressure-test the other for a week.
Building an AI agent inside a 7-year old Rails application
Author: Catalin Ionescu | Date: December 26, 2025 | Source: catalinonescu.com
Ionescu shows how to add an AI agent to a legacy, multi-tenant Rails monolith using RubyLLM tools gated by Pundit policies and Algolia—without relaxing strict authorization. By encoding data access into function calls, the LLM can retrieve only what the user is allowed to see; GPT-4o offered the best balance of speed and correctness. It’s a pragmatic blueprint for shipping agents in regulated, real-world systems.
12 executives, researchers, others who left OpenAI in 2025 — mostly to Meta Superintelligence Lab
Author: Katherine Li | Date: December 26, 2025 | Source: Business Insider
OpenAI saw a wave of high-profile exits in 2025, with at least seven research scientists decamping to Meta’s Superintelligence Lab and additional leadership departures across the year. The churn follows 2024’s restructuring and leaves Sam Altman as one of two remaining members of the original founding team. It underscores the intensifying talent war—and the strategic bets—behind next-gen AI.
The rise of the AI wingman
Author: Henry Chandonnet | Date: December 26, 2025 | Source: Business Insider
Singles are increasingly turning to AI for profile polish, opening lines, and texting advice, fueling startups like Rizz and Roast while big apps roll out their own AI helpers. Alongside enthusiasm comes skepticism—concerns about “chatfishing,” AI-heavy conversations, and altered photos—yet surveys show a growing share of daters already using AI. The dating playbook is changing, and platforms are racing to capture the AI coach in your pocket.
Mark Cuban said there’s a compelling reason for new grads to join small businesses instead of big companies
Author: Aditi Bharade | Date: December 26, 2025 | Source: Business Insider
Mark Cuban urges new grads to target SMBs, where they can drive immediate impact by deploying AI agents to optimize processes big companies already have covered. With agentic AI adoption rising rapidly and a tight entry-level job market, smaller firms may offer higher leverage and more flexibility. For job seekers, it’s a practical way to turn AI fluency into day-one value.
