I know it was hypothetical, but @ismh86’s example of relying on AI search or summarization of Materials Safety Data Sheets would be pretty concerning. The MSDS basically exists to tell you how this chemical is going to kill you. I don’t think AI is reliable enough for something like that at this point.
I’m sure there are lots of less life-or-death possibilities for companies using AI to help understand their internal documentation out there.
A discussion about Macs, iPads, native apps, web apps, server based automation, and AI, etc. 5 Stars!
It was a wonderful episode! I really enjoyed it. Love all the nerdy stuff Federico does! Now I have to get MPU Pro to see what he does with Notion. 
I’m gonna be using AI to help me figure out how to make my AI do what Federico does with his AI… Loved the episode!
I agree my ideal iPad would be a folding iPad.
Also wanted to say that jump desktop is the best solution to making an iPad more “Mac-like” 
I use JD on my iPad to access some computers at work. I don’t prefer it to using my MacBook to access them, but when I have to or I know it will just be a specific time frame, I can get significant work done with a keyboard and mouse on my iPad, remote accessing a Mac.
I always wondered whether Federico was born in N America or his parents were from there and spoke English at home etc. I did not expect that it was self taught! Kudos and new found respect! And little or no accent either. Bravo
I forget where Federico has told this story, probably in several podcasts at this point, but I believe he learned English initially because he wanted to play certain video games without waiting until they were localized to Italian so he’d import the English versions.
But Chris, I think the point was that they were discussing using AI on your own materials. There’s much less chance of hallucinations when the GPTs are told to get data from specific materials.
Overall, this was the area where I found this episode to be mind-blowing. I kept rewinding to be able to listen to Federico again as he explained the types of things he was doing with AI on his own data. His ideas opened up a whole new world for me about ways that AI will be useful to me. I loved hearing him say that he finds generative AI to be almost useless but the natural language processing to be so powerful. This was true Mac PowerUsers stuff!
Ironically, I have found generative AI, when used on my content I have created first, to be extremely useful.
I do agree, and preach, than OTHER AI tools then LLM/Generative stuff is immensely more powerful and useful.
e.g. Adobe is killing it with “Remove tools” and now the compositing “Harmonize” added to Photoshop and I have been a big fan of the “AI Voice processing” in Premiere Pro to automate basic talking head/voice over clean-up.
I like Frederico and I subscribe to his stuff, but - like all of us - many of his opinions are driven as much by his emotions, but he may not even realise it.
If he tried to explore generative AI, he’d find it useful, but he’s chosen to not do that (citing moral reasons, originally, if I remember correctly) so of course it’s useless for him.
That said, he may well change his opinion in a few months.
Exploring is part of his job - he goes off on his adventures, learns stuff, entertains us all, and it saves us a lot of hassle (and money).