I find it so hard to see anyone using this (for now) - making a reservation using OpenTable as they did in the demo literally takes all of a single minute once you're logged in, and don't you need to be for Operator to book, purchase, etc. things on your behalf?

It's so good. It was refreshing to see something that unique after so long having recycled content in the theaters.

If only this became a strong enough driving factor for a meaningful percentage of people in the US to consider Android alternatives, maybe we could get some competition back in the phone market

I wonder if I’d be any less frustrated about not having Japanese-quality 7-Elevens here in the States if I had a Buc-ees nearby

Even after it’s out of beta, I’m sure they’ll continue to lean heavily on a small “AI may occasionally be incorrect” disclaimer written somewhere.

This is a good example of why I think the AI search products (SearchGPT, Perplexity) should have been focused on way earlier than they were - the average person uses LLMs a lot like Google search and that added accuracy is important (also real-time information)

Not saying this is a good thing, but surely that will change eventually right? I’m sure there will be a point in the near future when LLMs can scrape data and write compelling articles from it that don’t sound like AI. Curious how the economics of the work will change then.

I’d be curious to hear his answer too, but if I had to guess it’s just a combination of the common factors including control over feeds, not deprioritizing both links & news, a preference against another Meta platform, etc.

Had a similar experience at my workplace using Rspack with the swc loader for TypeScript files - it’s been a huge improvement over Webpack

I want some of these AI features to give me the option to set chats and calls as ephemeral, meaning they don't store logs and persist the chat history. Whether simply doing a single query to search for something or having a short audio call, I don't always want to go back & manage/delete everything.

Since these primarily just seem like features, did they choose to not simply include them in Arc due to a desire to eliminate tech debt and start from scratch, or to preserve Arc for its existing user base?

The Discover feed definitely needs work. I find that it adds quite a lot of totally irrelevant or engagement bait posts and doesn’t seem to change regardless of how much I use the “see less” option

The argument from the right would probably be something like: people waited until Black Friday to buy things at a discount that they needed but couldn’t get previously due to inflation - causing misleading sales figures driven more by needs than wants

Understandable, I didn’t feel particularly bothered by Alien Romulus in that regard though since it was basically its own isolated movie without needing to have seen the others

This tracks with the recent leaks about a more LLM-like Siri in development as well, potentially eliminating the need for external models. Still strange though considering how they used ChatGPT as a marketing tool - wonder if they’ll choose to buoy OpenAI or dump them

Also, am I to understand that, if you failed to click the little “Redeem” button next to this message on the item’s page before purchasing, you’d lose out completely on $90?

This feature request always came off to me as a scapegoat or catch-all for users’ general dissatisfaction. Is not needing to do a single swipe over to the following feed when the app opens really going to change the entire user experience?

Without commenting on the broader Google breakup, they shouldn’t get to sleepwalk into winning the AI search race simply from existing market share. AI Overviews hasn’t been a very good feature so far, and I’m glad they have pressure from a few sides to compete more directly.