Hardware Podcasts

AI Detection with Max Spero

Oxide and Friends · 3 Sept 2026 · 1h 01m · Episode 20

What this episode covers

Max Spero, founder and CEO of Pangram, joins Bryan Cantrill and Adam Leventhal to discuss the technology and societal implications of AI detection. They analyze the recent Stanley Druckenmiller Wall Street Journal op-ed as a watershed moment for AI-assisted writing and explain how Pangram's classifier model, Pangram 4, achieves low false positive rates through 'synthetic mirroring' training data. The conversation covers the use of AI detection in hiring processes, the threat of 'humanizer' tools, and the philosophical impact of LLMs on truth and thought.

Most relevant to

early-stage companies

Topics

Mentioned

Companies
Pangram, Goodfire, Oxide Computer Company, Oxide, Yelp, Amazon, LinkedIn, Reddit
Products
Pangram 4, Claude, ChatGPT, GPT-4, Shell Game
Tools
Silico, Turnitin, Winston AI
Also
Stanley Druckenmiller, Wall Street Journal, Evan Ratliff

Related episodes from other shows

Other shows on the same ground, closest first.

  • How I Built This · 58m

    ICYMI... HIBT Lab! OpenAI: Sam Altman

    Sam Altman, co-founder of OpenAI and former president of Y Combinator, discusses the evolution of artificial intelligence from narrow tasks to the pursuit of Artificial General Intelligence (AGI). He details OpenAI's structural shift from a nonprofit to a hybrid model to secure necessary capital, the strategic decision to keep powerful models like GPT-3 behind an API for safety control, and his vision for AI as a ubiquitous, helpful companion rather than just a job-replacing tool.

    AI & machine learningStarting a product company
    7 chapters

  • How I Built This · 57m

    HIBT Lab! OpenAI: Sam Altman

    Sam Altman, co-founder and CEO of OpenAI, discusses his journey from a teenage entrepreneur with Looped to president of Y Combinator and finally to leading the development of artificial general intelligence (AGI). He explains the strategic pivot of OpenAI from a nonprofit to a hybrid nonprofit/for-profit structure to secure the massive capital required for compute and talent, while maintaining safety incentives like profit caps. Altman details the technical challenges of training models like GPT-3 and DALL-E, the risks of disinformation, and his vision for AI becoming an ubiquitous, helpful layer in daily life rather than just a job-replacing tool.

    AI & machine learningStarting a product company
    7 chapters

  • IoT For All Podcast · 20m

    Large Language Models and How To Use Them | Cyrano.ai's Scott Sandland | Internet of Things Podcast

    Scott Sandland, Founder and CEO of Cyrano.ai, explains that Large Language Models (LLMs) are essentially advanced autocomplete systems that predict the next word based on massive datasets, making natural language the new interface for computing. He argues that while LLMs accelerate execution, the 'move fast and break things' philosophy is dangerous for AI due to unpredictable ethical outcomes and biases, necessitating a parallel 'ethics roadmap' alongside technical development. Sandland highlights Cyrano.ai's proprietary API, which uses linguistic analysis to assess empathy and relationship value in contexts like sales and mental health, rather than relying on generic sentiment analysis.

    AI & machine learningSecurity & privacy
    6 chapters

  • Built 2 Scale · 1h 01m

    Live for Your Drive

    Hosts Matt Perrott and Scott Wilcox discuss the economic shift from 'token maxing' to LLM routing as open-source models like Kimi close the gap on frontier labs. They explore the emerging HR challenges of employee token spend, the mobile nature of voice-based AI work, and BuildPass's upcoming Apple Watch integration for site defect walks. The episode also covers a 45-minute hackathon that produced 'Tinder for Talent' and the future of AI glasses in construction.

    AI & machine learningConnected products & IoT
    7 chapters

  • Embedded.fm · 1h 08m

    502: Chat, J'ai Peté!

    Elecia White and Christopher White discuss the TV adaptation of Martha Wells' 'Murderbot Diaries' and the deceptive nature of Large Language Models (LLMs) like ChatGPT, Claude, and Gemini. They highlight specific failures of AI, including lawyers citing non-existent legal precedents and GitHub Copilot agents generating endless, looping pull requests. The conversation also covers control systems, contrasting theoretical 'fancy math' like Kalman filters with empirical tuning, and mentions the Segway patent US7370713B1.

    AI & machine learningEmbedded firmware
    6 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Oxide and Friends. The episode itself belongs to its publisher.