iOS 27 beta 1โฆ Speech synthesis still struggles to pronounce the word live appropriately based on context. Even small open-source models are able to get this right.
@BrianRoemmele I used both of those portable phones, the black one shown in the car also had a leather carrying bag and a blank phone book to write contact info in. The keypad tones were freaking loud.
@MissMi1973 I had a similar experience. I set the model to Opus 4.6, medium thinking effort, and it automatically switched to 4.7 when I hit send. Went from 0% to 47% usage burned with one request on my pro plan. ๐ฅบ
Anthropic folks can see when people use the F-bomb. So, I figured Iโd give it a shot๐.. I'm never this mean and I usually just start a new session and ask it not to guess.
@BrianRoemmele Holy smokes! This is 36 seconds packed full of so many triggering memories.. I almost sneezed just thinking about the carpet freshener trapped underneath those thick fibers. LOL
Iโm probably the only weirdo that was curious to know what @MaartenGr recorded in his article โA Visual Guide to Gemma 4โ; https://t.co/nESIFARlNY
I made a Mac app that takes a spectogram and converts it to audio. Then I had Gemma 4 e4b tell me about it.
@dominikmartn Nice work! For those that are struggling to get a similar experience and have strict practices against using private APIs in your org, have a look at WWDC25 session: https://t.co/YDHag8C98Y You wonโt get the control Domโs repo provides but youโll at least have something.
@dominikmartn Nice work! For those that are struggling to get a similar experience and have strict practices against using private APIs in your org, have a look at WWDC25 session: https://t.co/YDHag8C98Y You wonโt get the control Domโs repo provides but youโll at least have something.