It doesn't feel like speech recognition has been improving at the same rate as other generative AI. It had a big jump up to about 6% WER a year or two ago, but it seems to have plateaued. Am I just using the wrong model? Or is human level error rate, some kind of limit, which I estimate to be about 5%.
My M1 8GB Air did great before Tahoe; even medium complexity Xcode projects ran fine on it with other apps running. Since I made the mistake of upgrading it to Tahoe, it’s too painful to work in those projects.
It chugs if I launch a node server yes but that's an outlying use case for an 8gb air.
AI is so good these days I am using the laptop for quick changes more often, as I just push every change. I rarely need to fiddle. The general experience of using my desktop and laptop are converging.
I developed Node.js applications 10 years ago on a 2009 MacBook Pro with 5GB of RAM, and it was only a little tight on memory. 8GB _should_ be enough for moderate complexity development, but everything has become more and more memory-hungry with time.
> It chugs if I launch a node server yes but that's an outlying use case for an 8gb air.
May I ask if you have many 3rd party apps installed? What apps do you usually keep open at a time? Because 8GB should be more than fine for a node server.
Is it an Apple Silicon Air, or an older Intel model?
Tahoe is a massive regression in my personal experience (16GB here). So many random bugs and menu bar pop-up slowdowns (how is the system menu bar this unresponsive?).
Spotlight has gotten so bad, I can literally count the time it takes between typing the app name and the result showing up in the dropdown. Ended up switching Spotlight to Tuna.
Oh my god, yes. Spotlight on Tahoe is a joke. Why will it so often not display any results at all, even for system apps like Safari or Terminal? You’d think those would be in an always available cache guaranteed to always show up instantly? So many questions.
> Why will it so often not display any results at all, even for system apps like Safari or Terminal?
I've experienced this too, even after giving spotlight multiple shots months apart. For your sanity, I say just stop using spotlight. Don't let Apple steal your valuable waking hours with their crap QA.
I’ve had issues like that in the past, had to clear the cache for the spotlight index or something like that. Fixed it right away. Not sure if you’re facing the same issue.
For some reason Alfred is also a *lot* slower on Tahoe. I can often wait a whole second after pressing the shortcut before the bar appears whereas on Sequoia it was instant.
I'm pretty old and was the same as you for about five years, but now I just tick anything, much like the young adults. If they want my info, they can have it. I've not heard a convincing explanation why I, personally, should care
My experience is different. I was diagnosed with ADHD as a kid, but I never took medicine. I took amphetamines recreationally at uni, and the sense of calm and focus and relaxation it gives you is quite material. I think it stimulates the part of the brain that enables you to focus and relax.
I did briefly see a psychiatrist and took amphetamines under his care, but ultimately decided that I didn't want the hassle of medical supervision, so preferred non-medical treatments.
But the experience made clear to me that meds do have a role. I suppose I use coffee as a mild stimulant. Although it's so mild, I'd never really think of it in that way.
I'm definitely cheap when it comes to software costs, and I'm guilty of paying for big bills of mud.
This is completely different though. In the past I'd always go back to paid SaaS when the mud broke apart.
Not true for the past 18 months for me, I am developing features, maintaining code, all via natural language to a much higher standard than I've ever had in my business before.
A low bar, perhaps, but it's definitely not the same.
reply