Hacker Newsnew | past | comments | ask | show | jobs | submit | LogHouse's commentslogin

My use case (not very sexy) is process documentation. I take a basic outline, run it through GPT, manually edit some things here and there, boom, done. 10x productivity in that regard.


IT/CS pays well which brings in a lot of competition to the market.


For what it's worth, four years old seems too young for a phone or an AI companion. Parenting is an active role that shouldn't be outsourced.


It’s not sending every word to the cloud. I think you must invoke the AI features. Am I wrong?


I understood that it will have the full context of the data on your phone, in order to be ,,useful”.

We are yet to see if that means only the data you’ve invoked ai features for, or totality of your emails, notes, messages, transcripts of your audio, etc.


From the presentation it sounds like the on-device model determines what portion of the local index is sent to the cloud as context, but is designed for none of that index to be stored in the cloud.

So (as I understand it) something like "What time does my Mom's flight arrive?" could read your email and contacts to find the flight on-device, but necessarily has to send the flight information and only the flight information to answer the arrival time.


Strong agree here. Features are cool, but I value screen real estate and simplicity. Plus, the gpt app works fine for me. I don’t need it built into other things yet.


I know it’s simple but I am so pleased with this. Now if they fixed the mouse acceleration curve and the scaling issue for monitors, I’d be at peace.


It's the only OS that offers a good acceleration curve for my trackballs, been dying with the lack of configuration for the curve under wayland :(


Possible good news: recent versions of libinput support completely custom acceleration curves! Though I don't know if any wayland compositors expose configuration options for it yet.

https://wayland.freedesktop.org/libinput/doc/latest/pointer-...


What's the scaling issue?


Not GP, but macOS doesn't have fractional scaling, which means that running at non-integer scales is inefficient. For example:

A 3840x2160 monitor run at 1x means a framebuffer of 3840x2160.

At 2x, the effective resolution is 1920x1080, and this also uses a framebuffer of 3840x2160.

At 1.5x, the effective resolution is 2560x1440. This is implemented by rendering at 4x (5160x2880) and then downscaling that.

At 1.25x, the effective resolution is 3072x1728. This is also implemented by rendering at 4x (6144x3456) and downscaling that.

The difference between all of these is quite noticeable.

Edit: Apparently Apple has been shipping its laptops with non-integer scaling for a while, which is interesting.


Non-integer scaling is essentially impossible to do well unless you give up and turn all your UI design into web pages. Otherwise you get pixel cracks and other issues from finding all the places things are/aren't rounded properly.

The laptop displays work by scaling down 2x/3x too.


Windows does it just fine.


WinUI essentially webpage


I don't even mean the modern stuff. Even fairly old Windows apps can do fractional scaling just fine. WPF does it without a hitch, and it dates back to 2006. Even many WinForms apps can do it, and then of course Qt etc.

But even disregarding all that - if such scaling can be done without problems with webpages, then clearly there's no technical impediment to doing the same elsewhere. It's not like HTML is magic.


The worst part is that third party solutions (eg BetterDisplay) work. So Apple could find this if they wanted with ease. No clue why they’re so stubborn on it


Are you sure it works in a different way? I always thought Betterdisplay still showed you fractional scaling by downscaling from a much higher resolution. Well, I'd say I can still see the blurriness in the display when I pick a custom high dpi resolutnion from the list.



I wondered the same, but frankly, what other options are there?


It's going to be easy to substitute in their own LLM behind the API in the future. None of the branding or platform is controlled by OpenAI.


> Windows is the king of business productivity; I've never seen someone with a mac do what I qualify as real business work using any applications outside of the Adobe ecosystem

I agree but I feel that’s a very unpopular statement. There’s also counter examples. The OpenAI teams run on Macs, and it’s VERY hard to say they’re not productive. But in general, if you use excel or outlook, yah, there’s no way to replace that with a Mac. Even if, as you said, the apps are available.


In this context, “feedback” means reputation compromising news cycles.


It's so weird to me that a company like Microsoft would care that much about "reputation". Everyone basically hates them already. Many of the most successful companies in the US are widely (if not universally) hated by the people who pay them. Nobody loves comcast, or exxonmobil, or centurylink, or EA, or equifax, or facebook. People feel trapped and unable to avoid paying some companies or using their products no matter how much they hate them. How many people have paid Microsoft for their OS at all? How much money would they really lose if they ignored the bad press? How many grandmas would start downloading linux?

I'm glad Microsoft is making changes, but I wonder how much is out of fear for their reputation and how much is just to try and comfort people and get the news to stop talking about it so that everyone doesn't just disable it as soon as it rolls out.


So much of the youth use nothing but mobile OSs and Microsoft lost that - they’re not even a player! Macs have been selling more each year. Valve/Proton is bringing gaming to Linux. Microsoft is losing the non-corp consumers. The features are aimed at non-corp consumers.


But will UBO v2 be updated after manifest 3 goes live?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: