Models I’d be most excited to see:
* Llama 3.2 8b-level model with sub-100ms latency (OpenRouter: Groq’s latency is 390ms)
* Better, faster models for <32GB local RAM
* Sonnet 3.5, but cheaper & better multilingual
* The obvious: new frontier models, reliability across the board
@aliceisplaying I just need anthropic to let me back into my account so I can buy API credits (was spuriously banned two months ago, got allegedly reinstated multiple times, still blocked lol)
@kostdn@apple_inte You could at least make the argument the base A12 devices have much lower multicore performance than A12X/Z. Plus the base iPad 8 started with only 32GB storage.
@SukritGanesh@StatisticUrban That still leaves a lot of room for misinterpretation. Most AI things need a light touch, but this kind of ad needs to be totally banned, no exceptions
@Liandrium Semantically notifications fit with the time and the ringer indication, and Siri is supposed to live in the dynamic island now, which itself makes semantic sense as the dynamic island is supposed to be a “smart” area
@merlinaudio_@luciascarlet As much as I think 95% of “child safety” is unnecessary and tramples on the rights of adults (and children tbh!) I didn‘t mind. They‘re the only people with a reasonable, private, optional approach to the ”problem” and they need to get politicians on board with it
@23Aaron_@mweinbach Oh yeah they were explicit: “iPhone Air, iPhone 17 Pro, iPhone 17 Pro Max, iPad (M4) or later with at least 12GB of unified memory, and Mac (M3) or later with at least 12GB of unified memory”.
@DolphinMossad Honestly not sure she’ll even be that bad. Bad, yeah, but the big worries with her are dictatorship and corruption and the various presidents carrying out Castillo’s term did pretty well in that regard… At least she might finish her term if she wins.
@MMinnym45@constans "The establishment of the chaplainship to Congress is a palpable violation of equal rights, as well as of Constitutional principles" - James Madison
@sarahookr Of course human data entry is subject to human error as well. And the greedy synthesis - have AI do it and humans verify - will make the humans lose interest and defeat the whole purpose.