Oddbean new post about | logout
 In just the past two weeks:

🖥️  Microsoft announced deep integration of generative AI tools across Windows 11

🔍  Google rolled out changes to its Bard generative AI that allow it to use all your documents, emails and calendar items as fodder

🗨️ Amazon showed off the next generation of generative AI capabilities for Alexa, which should make it chattier and more flexible

💬 Meta announced a chat-based assistant for Instagram, WhatsApp and Facebook

🧵/4 
 "Those who don’t opt to use AI to help them summarize others’ reports (likely generated with the help of AI), respond to emails (ditto) or adapt to new business processes (also created with the help of AI) risk drowning in a fire hose of communications and increased complexity."

Now, honestly, I hope I am wrong. But I also think that in the not too distant future, opting not to use these tools will be like opting not to use email in 2023. Possible, but hard.

5/5

https://www.wsj.com/tech/ai/why-you-soon-wont-be-able-to-avoid-aiat-work-or-at-home-65febec7 
 @ca5977db Except email worked (or at least it did back then). This is just grey goo. 
 @ca5977db here's one thing that baffles me about the "use it to summarize!" Use case for LLMs, don't most important documents *already come with a summary*? An abstract, an executive summary, a lede paragraph... 
 @5d5c4ecb honestly they’re rarely all that great. And then what if what you really need isn’t in there? What if you need a summary of the way a document addresses subject x, say, which is buried on page 65, 32, 87 and in half a dozen footnotes and 2 figures? 
 @ca5977db None of it is real AI. But that’s not the sstounding thing. The astounding thing is how instantly, freely, willingly, and enthusiastically so many people including media types have been fooled into going along with all this stuff. Marching right off the proverbial cliff. 
 @ca5977db I wrote about this too! https://notes.ghed.in/posts/2023/excessive-intelligence/ 
 @0c1eee0b man, that kicker was chilling. Good thinking on this. 
 @ca5977db can confirm this is already happening.

You can tell who's doing it too. AI has a certain "rambling" quality that adds superfluous words or context.

I'm gently chided for not participating. Feels like something from 1984.