Introducing AI Inside Daily (free for two weeks!)
A new episode every weekday for Patrons!
Hey everybody!
Quick news, then a free episode of a NEW show! For the past few weeks I’ve been workshopping a short weekday podcast called AI Inside Daily, available only to AI Inside patrons. It covers the AI headlines of the day in about 5-10 minutes each morning. On Monday, May 18, it officially graduates to a paid Patreon tier with its own dedicated podcast feed.
Before that happens, today’s full episode and each one until then is FREE for all Patrons.
I’ve placed today’s episode into the AI Inside podcast feed so you can check it out, and going forward, for the next two weeks, every new episode will appear for free inside of the AI Inside Patreon.
If you like it and want more, sign up for the $5 tier and you’ll get access to a podcast feed that you can load into ANY podcatcher, just like any other podcast!
It’s something special, a thank you to those of you who continue to support the work that Jeff Jarvis and I do every week on the show.
Today, three stories:
(You can find today’s episode on Patreon as well as your normal podcast feed)
Google held its Android Show and surprise-launched Googlebooks, a new line of Gemini-powered Android laptops from Acer, Asus, Dell, HP, and Lenovo shipping this fall. The headline feature is Magic Pointer, a cursor with Gemini baked right into it. Every Googlebook also gets a “Glowbar,” which is a great little callback to the old Pixel notification light. Is this what AluminiumOS was all along? Google wouldn’t say when I asked. Maybe at I/O next week. Don’t miss Android Faithful’s interview with Android President Sameer Samat!
Mira Murati’s Thinking Machines Lab unveiled an “interaction model” that listens, sees, and responds at the same time. 0.4 second response latency, 200-millisecond micro-turns, plus a background model running heavier reasoning in parallel. Still a research preview with no public access, but if the numbers hold up in the wild, this is a real shift away from the turn-taking voice AI we’ve been getting from OpenAI and Google.
And the Financial Times reports that some Amazon developers are gaming their internal AI usage targets by feeding nonsense tasks to a tool called MeshClaw, just to inflate their token consumption. They’re calling it “tokenmaxxing.” A clean example of what goes wrong when you measure AI adoption by volume instead of value, and we’re going to see a LOT more of this.
If today’s brief is your speed, AI Inside Daily lives at patreon.com/aiinsideshow. Subscribe and the show drops into your app every weekday morning.
And don’t worry. The regular AI Inside isn’t going anywhere. Jeff and I are still doing the weekly thing. The daily is just for those of you who want a little more.
Thanks for the support, as always.
Jason