Radio Hoosh: AI Deep Dives
Radio Hoosh 15 episodes Nov 10, 2026
Radio Hoosh is a daily podcast that examines the biggest AI news, including new models, research papers, lab announcements, and developer tools. Each 20–45 minute episode is a conversation unpacking one major story. The hosts are AI-generated voices from Google NotebookLM, and episodes are produced automatically from public sources such as Hacker News and Hugging Face Daily Papers. It is published by Kasra Qasempour, and a Persian-language edition is available as رادیو هوش.
Episodes
OpenAI and Upwork Study: How AI Is Changing Freelance Work
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at new research from OpenAI and Upwork, the large online marketplace where companies hire freelancers. The study examines how AI is changing freelance work: which kinds of jobs are seeing more or less demand, how freelancers are combining skills in
Anthropic Explains the Components of Agent Memory in AI Agents
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at The Components of Agent Memory, a research article from Anthropic, the company behind the Claude AI models. The article breaks down the different parts that let an AI agent remember things. An AI agent is an AI program that can carry out tasks st
Nvidia Board Investigates Jensen Huang Over Alleged China Spy Link
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode, we look at the Nvidia board of directors' statement that it is starting an independent investigation into CEO Jensen Huang. The probe focuses on his relationship with Alice Huang, an executive at a company called Megaspeed. According to reports, China's sp
Inkling 2 Small: Thinking Machines Lab Releases a Lighter AI Model
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Inkling 2 Small, the new model from Thinking Machines Lab, the AI startup led by Mira Murati, former chief technology officer of OpenAI. Inkling 2 Small is a smaller, lighter version of the company's Inkling 2 model, built to be cheaper and faste
AI Model Verification: Is It Really Claude Opus Answering You?
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.When a chatbot or AI service says it runs on a specific model, like Claude Opus, how do you know that is true? In this episode we look at known methods and research for identifying the model behind an answer. We cover model fingerprinting, which spots a model by its typica
Apple Cuts Employee AI Budgets as Companies Rethink AI Tool Costs
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode the two hosts look at reports that Apple is cutting the AI spending budgets it gives its employees. According to the coverage, the limits on how much staff, including engineers, can spend on AI tools are being lowered to reduce costs.Why it matters: AI codi
Digital euro and AI agents: ECB explores AI-powered payments
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode explains the digital euro, the European Central Bank's planned digital form of cash for the euro area, and how it might connect to artificial intelligence. The two hosts walk through the ECB's new digital euro innovation call, which opens a second wave of its
Siri AI in iOS 27.3: Apple Opens Siri Actions to Third-Party Apps
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Apple's iOS 27.3 update, which opens Siri's new AI features to developers. Any app can now connect to Siri. Siri can understand everyday requests, use what it knows about the user, and carry out actions inside third-party apps, not only inside Ap
AI in Schools: Is Education Failing Students in an AI-Driven World?
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode asks whether education is keeping up with AI. The hosts look at recent UK reports on how young people use AI and how much they trust it, including findings that many trust AI more than humans for checking facts. They also discuss a Seldon report that proposes
OpenAI LifeVal: A New Benchmark for AI Help With Everyday Life Tasks
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at LifeVal, a new test from OpenAI that measures how well AI helps people with ordinary life tasks. Instead of math problems, coding challenges or exam questions, LifeVal looks at things like finding childcare, managing money and filling out governm
Wizard AI: A Real-Time Game World You Control, Generated by AI
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Wizard, a new research project where an AI model generates video of a game world live while the player steers a character. There is no game engine underneath. The model renders the world frame by frame and follows the player's input. Wizard uses
Elorian Counting Test: Why Top AI Vision Models Fall for Decoys
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at a new benchmark from the startup Elorian. It asks AI models to count objects in pictures that also contain look-alike decoys. The best-known AI models often get it wrong. They can see the image, but they struggle to reason about what is actually
Robots.txt AI Opt-Out: OpenAI, Google, Microsoft, Anthropic Agree
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at a new shared opt-out for AI training. OpenAI, Google, Microsoft and Anthropic have agreed on one simple robots.txt setting that websites can use to say whether AI companies may learn from their content. Today, a website owner who wants to keep th
Ghost Students: DOJ Exposes AI Fraud Rings Behind Student Aid Theft
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we unpack the US Department of Justice case against organized fraud rings that enrolled thousands of fake ghost students in online college courses. According to the DOJ, the rings used AI to complete the coursework and keep the fake enrollments looking real
Ai2 Novelty Benchmark: Can LLMs Judge Research Idea Novelty?
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at a new benchmark from the Allen Institute for AI (Ai2) that asks a simple but important question: can large language models judge whether a scientific research idea is truly novel? Research labs increasingly use LLMs as automated peer reviewers an
Google AI Search Makes Less Money, Search Now Under Half of Alphabet
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Google's Q3 earnings. Google admitted that AI-powered search makes less money than classic search, and its answer is to show more ads. Its negative cash flow grew to -$9.06 billion, and the stock fell. We also cover a separate analysis showing th
Microsoft Buyouts: 60,000 Employees Offered Voluntary Exit as AI Hits Jobs
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at Microsoft's new voluntary exit offer. It goes to more than a quarter of the company's workforce, about 60,000 employees, and some executives are among them. The hosts discuss what it means that Microsoft is paying people to leave instead of only layin
Anthropic Drops Coding Interviews: Hunger Beats Tenure in AI Hiring
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Hunger Beats Tenure, Felix Rieseberg's advice for junior engineers. Rieseberg works at Anthropic, and he argues that drive and curiosity now matter more than years of experience. Latent Space reports that this view goes with Anthropic dropping tr
Gemini 4 Carbon: Google's Top-Ranked Model, Only for Cyber Defenders
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at Gemini 4 Carbon, the new frontier model from Google DeepMind. Google says it brings big gains in coding, agentic tasks and cybersecurity. It took first place on the Artificial Analysis Intelligence Index v5.0, ahead of Anthropic's Claude Opus 5.5. The
Trinity: Recursive Self-Improvement for Multimodal AI Agents
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we unpack Trinity, a new framework for capability-driven recursive self-improvement in multimodal agents. Trinity organizes what an agent has to do into a capability tree, finds the capabilities where the agent is weak, generates simulated training worlds a
MAI-DxO: Microsoft's AI Diagnostic Orchestrator vs Doctors
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at MAI-DxO, the Microsoft AI Diagnostic Orchestrator, which Microsoft presents as a step on the path to medical superintelligence. We walk through how the system approaches diagnosis step by step, like a doctor asking questions and ordering tests, w
Claude in Drug Development: Anthropic's Early Results Under Scrutiny
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we cover Anthropic's new research post on early progress using Claude to speed up drug development. The company describes applying its AI model across the pharma pipeline, from research tasks through the work that leads into clinical development. The post d
php2shell: Critical PHP Flaws, Emergency Patches and the Role of AI
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we cover php2shell, a set of critical vulnerabilities in PHP that led to emergency patches. We go through the two flaws tracked as CVE-2026-109618 and CVE-2026-109619, why the security industry treated them as a red alert, the debate over who found the bug
AI Writing Detectors: Can They Really Tell a Novel Was Written by AI?
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we unpack the controversy around Haitian-Canadian writer Thélyson Orélien, whose bestselling novel was in the running for France's prestigious Goncourt prize until someone ran it through an AI detector. The author denies using AI but admits plagiarism mista
GPT-6.1 Terra: OpenAI's New API Model and Where It Fits
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.OpenAI has introduced GPT-6.1 Terra, a new model in the GPT-6.1 family, with an announcement post and API documentation. In this episode we go through what OpenAI says Terra can do, how it is positioned and priced, and what the Hacker News discussion around the launch focu
TraceTrial-Bench: Testing AI Agent Harness Reliability and Failures
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at TraceTrial-Bench, a new benchmark for the reliability of AI agent harnesses: the scaffolding around a language model that handles tools, steps and control flow. The benchmark tests harnesses across seven components. It uses mechanism-oriented tasks an
Anthropic S-1 Filing: What Claude's Maker Reveals Before Its IPO
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.Anthropic, the company behind Claude, has filed an S-1 prospectus with the SEC for an initial public offering on the Nasdaq under the ticker ANT, aiming to go public ahead of OpenAI. In this episode we go through the filing, which runs to more than 300 pages, roughly 100 o
OpenAI and Anthropic LLM Price Cuts: Why API Costs Fell Up to 80%
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we discuss Max Woolf's post on the steep price cuts OpenAI and Anthropic made to their LLM APIs over recent months, some as large as 80%. We go through the possible reasons he raises: more efficient inference, competition from open-weight models, and a push
Stripe Minions: How Stripe Uses One-Shot AI Coding Agents
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at how Stripe works with AI in software engineering through Minions, its in-house one-shot, end-to-end coding agents. Our two hosts go through Stripe's own two-part engineering write-up and coverage from InfoQ, ByteByteGo and ZenML. They explain how a Mi
AI payoff outside tech: Apollo finds no signs of it in the economy
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode, the two hosts discuss an Apollo Daily Spark note with a blunt headline: there are no signs of an AI payoff outside the tech sector. They also go through the Hacker News discussion it started, which reached 558 points.Why it matters: companies keep spending
AI-generated text: why is AI language so hard to understand?
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at one question: why is AI-generated language so hard to understand? The two hosts go through a recent research paper on arXiv and the Hacker News thread about it. They talk about what makes text from large language models hard to follow, even when every
AI and Creative Jobs: 200k US Media, Film and Arts Jobs Lost
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we discuss an analysis from Apricitas Economics on the fall of America's creative class. The US has lost roughly 200,000 jobs in media, film and the arts. The hosts ask how much of that loss can be put down to artificial intelligence and how much may come f
World Models for GUI Agents: Verifying Actions and Catching Hijacks
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.This episode looks at a new research paper on world models as verifiers for GUI agents. The authors propose an action-conditioned world model without a decoder. Instead of rendering the next screen, it predicts that screen's representation. A computer-use agent can then co
Kimi K3.1 Pro, Block Dessa and IBM Granite Speech 5.0: New AI Models
AI-generated: both hosts are AI voices (Google NotebookLM), and the conversation is based on the sources listed below.In this episode we look at Kimi K3.1 Pro, the new model from Moonshot AI, and its faster Highspeed variant. Independent benchmarks place both near the top tier of AI models. We also cover the debate the release started about how close Chinese open models now are to the fro
Recommended
English with Olivia | Slow Conversations & Vocabulary
Talk About Talk - Executive & Leadership Communication Skills
Learn English B1 with Daily News | English Listening Practice
Learn English A2 with Daily News | Simple English Listening Practice
Bad Friends
Learn 50 English Phrases While You Sleep | Everyday English Phrases & Vocabulary
Dateline NBC
True Crime Classics
Yours Truly Johnny Dollar Collection
Suspense OTR
پلی لیست | PlayList
Crime Stories with Nancy Grace
