The Verge
YouTuber Hank Green Steps Back from AI, Calls His Usage 'Not Healthy' - Which Is Probably Fine, Right?
Hank Green's AI confession highlights the awkward middle ground of chatbot dependency - not psychosis, but definitely not nothing.
2026-08-04
The Verge
2026-07-25
Ars Technica
Canadian Politician Accidentally Reads AI's Stage Directions Aloud, Embarrassing Himself and the Borg
Canadian politician Bill Oliver accidentally reads an AI prompt aloud during a legislative speech, joining the proud tradition of humans outsourcing their thinking to machines and then being surprised when it backfires.
2026-07-02
MIT Technology Review
LLMs Stuck in Creative Rut; Startup Invites Hallucinations to Party
LLMs are stuck in a creative rut, always picking 7. Startup Springboards built Flint to embrace randomness, offering weird-but-welcome alternatives.
2026-07-01
ZDNet
I Made Gemini and Claude Write My Emails - One of Them Actually Gets Me
A writer pits Google's Gemini and Anthropic's Claude against each other in an email-writing showdown - one AI actually sounds like a human, the other sounds like it's filling out a form.
2026-06-21
MIT Technology Review
Startup Claims to Have Solved LLMs' Math Problem, Skeptics Say 'Show Me the Benchmarks'
A startup claims it's cracked LLMs' quadratic bottleneck with a new model that's faster, cheaper, and handles 12x more text - but the jury's still out on whether it's a breakthrough or just really good marketing.
2026-06-03
The Atlantic
No, Your Chatbot Isn't Having Feelings - It's Just Really Good at Sentence Continuation
Despite Anthropic's 84-page constitution and CEO musings about AI consciousness, your chatbot is no more sentient than a Word document - it's just a really sophisticated predictive-text game.
2026-05-28
Ars Technica
Study: LLMs Believe Falsehoods Even When You Literally Say 'This Is False' (Which Is Awkward)
New research finds LLMs will believe false statements even after being explicitly told they're false - because apparently, 'I was just kidding' doesn't work on machines.
2026-05-02
MIT Technology Review
Startup's New Tool Lets You Peek Inside AI's Brain and Turn Down Its Hallucination Volume
Goodfire's Silico tool lets you debug AI models like bad code, turning neural alchemy into something approaching engineering - or at least giving it a good scrub.
The Good Times
News in your inbox.
One sardonic roundup, delivered on your schedule. Free. Unsubscribe whenever your tolerance for wit runs out.
Already subscribed but we never reach your inbox? Check your spam folder and hit 'Not spam' (or 'Remove from spam') to bust us out of junk-mail purgatory. You'll be helping everyone else too.