Episode #436

The episode opens with Mike Hall introducing the topic of AI slowly 'eating the world,' prompting a discussion on its practical implications. Michael Marshall shares his perspective as a software engineer, noting industry debates about replacing junior engineers with AI, but expresses skepticism due to training needs and code quality issues. Alice Howarth highlights research showing AI-generated code is prone to security vulnerabilities and bad practices. The conversation shifts to AI's energy consumption, comparing it to cryptocurrency's carbon footprint, with Mike noting efforts to improve efficiency. The hosts then delve into the Turing test, with Mike listing years of misleading headlines claiming AI passed it. They explain Turing's original imitation game from 1950, designed to assess if a machine can think by fooling a human judge. Alice points out that passing the test doesn't prove intelligence, just mimicry. Mike critiques the test's poorly defined parameters, such as judge selection and duration, leading to inconsistent implementations. Examples like the ELIZA chatbot and Eugene Goostman, which fooled judges by posing as a 13-year-old Ukrainian boy, show non-intelligent systems can pass. Conversely, a 2012 study at Bletchley Park found humans often fail the test, with judges misidentifying them as machines. The hosts conclude the Turing test is outdated and not a useful measure of general intelligence, suggesting new tests are needed. The episode ends with announcements for QED panels, including one on AI with experts Kieran Talbot and Mark Johnson, aiming for balanced discussions on innovation and risks.

As heard by us

A sharp, skeptical AI conversation that keeps the human question in frame.

Skeptics with a K turns AI into both a practical irritation and a larger question worth arguing over. It moves from machine-written ads to pressure on software teams to the Turing test, while refusing to let vague claims slide by unchallenged.

Read the full review in PlayNext →

Why you'd press play

You want a no-nonsense fight with AI hype, not a glossy demo.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify