Live data from Hacker News

Apple unveils new accessibility features

apple.com

61–70 of 400 posts

Re: Apple unveils new accessibility features

#61

This looks like a genuinely useful application of LLMs. I wish more companies focused on how they can help humans instead of replacing us or squeezing us as hard as possible in the name of productivity.

Aren't the LLM-based features of this announcement catch-up features? Describing the contents of the screen is something Gemini has been doing on Pixel phones for a while. It's a fairly obvious use case for a multimodal AI.

My one hope is that this eventually becomes widespread enough to stop alt text scolds.

Re: Apple unveils new accessibility features

#64
One thing Apple really needs to get right is speech to text transcription. They've nailed accessibility in so many ways and yet it feels like they're a decade behind on properly transcribing voices. At least half a decade.

Input on the iPhone is so dreadful nowadays. Their palm rejection is definitely worse than before, so mistyping is more frequent. Their text-correction algorithm for typing is worse than before, and it frequently makes incorrect corrections to words that I don't notice, because they change words a few words back from where I typed. And STT hasn't improved. On top of that, my fingers are tired of the phone form factor. Please make the iphone not a chore to use, apple.

Re: Apple unveils new accessibility features

#65

I have difficulty trusting this. There are plenty of videos online of LLMs making up stuff like "I just ate a hot dog, is there mustard around my mouth?" "No, everything is clean" while there is a big yellow stain om the guy's face

The problem is using a language model to assess images.

Probably 80% of "LLM's are below expectation" complaints (from the general population) involves some form of image analyses.

Image tokenization is hard because unlike language tokenization, where every token is extremely dense with meaning, image tokens tends to be meaningless or irrelevant but are processed all the same.

Give an SOTA LLM a picture of toothpicks and ask it to move one to make a square, and it will probably struggle and fumble it. But give a mid-size LLM from 2 years ago the same problem in verbal form, and it will nail it almost every time.

That takeaway is, do everything you can to avoid having the LLM need to rely on images for the answer.

Re: Apple unveils new accessibility features

#66
post #48

Earlier quoted context omitted.

Blind people can't change video speed? The control is available right there.

Yes, the audio speed can be adjusted. Whether that control you see visually is actually accessible to a blind user is a different matter entirely. Further, it maxes out at 2x, but a blind person would typically screen read at the equivalent of 3-6x.

Huh, 2x is low even sometimes for sighted people.

Related, it seems like YouTube recently paywalled speed increase beyond 2x. Another way in which it's not cheap to lose sight, I guess.

Re: Apple unveils new accessibility features

#67
post #6

Fun fact: This video was made accessible to sighted people because no blind person would ever listen to voice at that speed. Honestly if you ever observe a blind person using computers you'd impressed how they can listen to audio at unimaginable speeds.

I did IT for a community Center way back in the day and the director was blind. I was blown away by how fast his screen reader read things out to him - completely incomprehensible to me - and his efficiency with keyboard shortcuts would put even vim/emacs elitists to shame.

Re: Apple unveils new accessibility features

#68
post #20
post #7

Earlier quoted context omitted.

It’s still useful to get the information instantly and verify it later. Arguably asking someone you trust to read the number for you might be a better idea than calling the company. Not everybody has that option though.

And not everybody wants to use that option all the time. Asking a human makes you feel dependent more than using a tool does.

Aaaand the logistics of making that call to the company to confirm the amount on the bill can get awkward. IVR and hold-time hell just to get a human to have to explain your predicament as to why you're asking for such a mundane piece of information that was in fifty other touchpoints that you couldn't access as quickly or easily.

(I'm also picturing the poor CSR at the other end of the phone wading through hundreds and hundreds of call logs over the years for simple requests and managers up above screaming 'why is this guy calling us all the damn time costing us money'...)

Re: Apple unveils new accessibility features

#69

Earlier quoted context omitted.

https://youtu.be/wKISPePFrIs?si=ahGfFp0U7-pTU9w6&t=43 my go to example of this is this talk by Saqib Shaikh (a blind software engineer at Microsoft) giving a talk about Visual Studio. Link is timestamped

I think it takes quite a lot of practice to reach this speed. It's not rare among blind developers, but I think it still takes a lot of work to get there. Pretty impressive! I wish more people would watch videos like this just because having a realistic idea of how blind people do certain tasks can help you move from pity or even compassion to a more productive kind of understanding. I think sometimes when you haven'…

If you are dedicated, few weeks to few months of usage with regular ramp up. You should be careful with adjusting which symbols are read though and sometimes the programing languages matters because different symbols have different significance for understanding the code.
Post reply on HN