Now Playing
GPT-5.5 Hallucinates 52% Less, Mythos Restricted & Tech's 142K Layoffs
Audio is available on Spreaker — see link below.
OpenAI claims GPT-5.5 cuts hallucinations by 52%, Anthropic quietly restricts a model deemed too dangerous, and 142,000 tech workers are being traded for $700B in AI capex. Today's briefing covers the honesty race, copyright war, and a species-level warning from DeepMind's CEO.
Audio is available on Spreaker — see link below.
Two major AI labs are claiming measurable breakthroughs in honesty this week, and for the first time, at least one of them is putting a number on it. OpenAI's GPT-5.5 Instant is now the default model inside ChatGPT, and the company says it produces fifty-two point five percent fewer hallucinations than its predecessor on high-stakes prompts across medical, legal, and financial domains.
The more striking development from Anthropic isn't Opus four point eight. It's what they decided not to release.
One hundred and forty-two thousand U.S. tech workers have been laid off in the first five months of this year. That's up thirty-three percent compared to the same period last year.
The labor story is sharpest among younger workers. Developer employment for workers under twenty-six has dropped twenty percent since twenty twenty-four.
CNN became the first television network to file a copyright lawsuit against an AI company, naming Perplexity as the defendant. It follows failed licensing negotiations in twenty twenty-five and joins a growing legal track that includes suits from the New York Times and the Chicago Tribune.
DeepMind's CEO Demis Hassabis told a Stanford audience that AI is advancing ten times faster than the Industrial Revolution, with what he called little margin for error over the next decade. He called the current moment a species-level transition and pushed for international regulation and independent model evaluations within five to ten years.
The two signals worth holding onto from today: honesty benchmarks need independent verification before they mean what the labs say they mean, and the Mythos situation is an unresolved proof point on whether the industry has governance structures equal to its capabilities. Both will have answers.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.