Practical Year - Part 2 with Donny Wals - Empower Apps

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Leo Dion (host) (00:03):
all right.
Swift ui what are we onyear four and a half now?
No, it was 19.
It was

Donny Wals (guest) (00:09):
Yeah, I was say, I think five.
Yeah.

What where are we?

I haven't written UI kit, code like from scratch in
probably three years now, maybe four.
So sore at a good spot.
I think.
I think for a lot of the apps thatI work on, they're usually smaller.
If they're new switchwise is perfectly fine.
Like the new navigation model thatwe got, I think in I was 16, is

(00:34):
pretty good, like navigation path.
I don't know that everybodyloves it, but I think it's
much better than it was before.
There's still somecomponents like image picker.
I still drop down to UI kit for that.
I think social has a builtinone right now that I tried.
I didn't like it, so I decidedto do the bridging thing.
And I.
Yeah, in client projects, like existingclient projects, that's the only place

(00:57):
where I still do UI kit pretty much.

Are they are clients, are you seeing clients wanting
to move from UI kit to Swift ui?
Yeah, this is gettinga pain in the butt.
Can you just move it to Swift ui?
'cause we

Yeah.
I don't think I've seen any client thatsaid, oh no, we're fine with UI kit.
We intend to stay there.
They've all said We want tomove to Swift ui, we want new
stuff to be done in Swift July.
A lot of concerns aroundnavigation, 'cause a lot of
them will use like a coordinatorpattern or something like that.
And that seems to be like,it's not a fully figured out

(01:29):
problem with Swift UI yet.
There's some takes on it.
But yeah, UI Kit has aspecific way of doing that.
So if you.
If you have your UI kit app it canbe pretty much non-trivial to migrate
that into the switch UI world.
So a lot of people will keep their likeUI kit view controller and that just
presents a full screen SWIFT UI view.
That's a pattern that I'veseen a lot and still see today

When do you dip into UI kit you said?

for image picker is one that I did recently.

Okay.
Yeah, we talked recently in a previousepisode, I think it was with Nathan,
about contacts and event stuff thatstill needs UI kit and then like
for me it's mostly with bushel.
It's been app kit stuff doing,dealing with Windows directly.

Yeah.

I can only do with Ns, window delegate and things like

Yeah, I think for the Mac there's been
more rough edges than for iOS
Also hover API, so I dunno ifthose are fixed, but I dunno.
Hover used to be pretty bad.

the other, well what it seems to me is like iPad is well,
okay, iPhone is the first class citizen.
Obvi is number one, but iPadwould be number two, and then
like Mac is number three.
So if they do something with iPadthat involves Windows or stuff like
that, then the Mac will get it too.

Yeah.
It seems like iPad is likethe middle thing there, right?
Where iPad gets everything iPhonegets, and then they want to make
iPad better, and it's oh, mightas well throw this at the Mac too.

So speaking of, was there anything else
you wanted to mention with 50 y

no.

Speaking of platforms we had a new platform revealed two
years ago now, year and a half,

Yeah.

that was gonna revolutionize everything.
And we're talking about the Vision Pro.
You got one, right?

I don't.

You did it.
Oh my gosh.
So surprised.
Yeah.
Have you done anything with it?

I've tried it.
It's a cool device.
It feels very high-end,very high quality.
Sensors seem pretty good.
I think what it's mostly sufferingfrom is just, it's great tech,
but it's not necessarily doingsomething that a Quest cannot do.

Do you have a quest?

I don't have one.
I borrowed one from a friend fora little while, so I used it a
bit, but it, the Quest did feela little bit more clunky with
the controllers that one had.

so the quest, did you just play games on it?

Yeah, a little bit.
Yeah,

I mean, so like the Vision Pro differentiates itself when
it comes to like movies and work, right?

for sure.
Yeah.
And so that, that's where it'sreally good, I think for movies.
I don't see myself sitting on acouch and telling my wife like,
Hey, I'm gonna watch a movie.
You go enjoy something else.
And just putting on the vision,

Right, right.
Exactly.

and right, and like for airplanes, it's great,
but realistically, how much do youactually fly for most people, right?
So there's all these thingswhere it's really good.
I think for productivity andhaving a big screen while
working, it could be pretty cool.
But then again, for the 4,000 thatI would spend on vision, I can
probably get like a pretty, prettycool setup in the real world, right?

(04:44):
So, and that's where I keep bumpinginto with vision, it's like, it does
a lot of things really well, but itfeels like it, it is just not the one
thing that consumers would really need.
And I think Apple is hoping thatwith this iteration, developers
would pick it up and they wouldbuild that killer app, right?
Remember like when iPhones.
People would buy iPhones, Ithink just to play Angry Birds.

(05:06):
I know definitely that wasa selling point for me where
Angry Birds would be on it.
It would be so cool.
I don't think I think Apple's hopingthat vision would get like a handful
of apps where people would be like,and this is why you need vision.
And they didn't really get that.
It seems.

Do you think that's related to the Vision Pro?
Do you think it'srelated to the app store?

I think it's probably a bit

think the challenging of building for a device like this.

I think it's probably a bit of everything.
I think companies that are good atbuilding VR experiences kinda looked
at this and decided, no, we, we'dmuch rather keep developing for
the platforms that we already have.
I think developers were like, oh, thisis really cool, but it's expensive.
So there's only a handful ofdevelopers that were like,
I think this is really cool.
I'm gonna build stuff for this.

(05:53):
So they bought it and they built stuff.
But then most vision owners that I know,they're like probably eight outta 10.
They only really use it because it'ssitting in their office and they see
it and they're like, I should use that.
And then I put it on.
They're just like,okay, now what do I do?
And there's only a couple that arelike, I use it like multiple times a
week for work or for this or for that.

Right, right.
Yeah.
I mean, yeah,

So it's probably a mix of everything.
And I think the pricing too, right?
It's three and a

Oh, it's definitely price.
Yeah.

it's it's only attractive to people that are hoping
to make an app for it right now,I don't think and like super rich
people, but I don't think I wouldbe able to recommend to any of my
non-developer friends to get one at

Well, I mean, I wouldn't recommend my developer
friends to get one either.

Well, if they were interested in making an app for it.
But other than that,

that's, you're never gonna make the money back.
I mean, you better be doing itwith, as a hobby and not as a

Oh yeah.
Oh yeah, for sure.

Yeah.
Yeah.
Okay.
Fair enough.
Yeah, so like I have right heremy pro display XDR that I probably
paid, but the same as a Vision Pro.
And guess what?
I like actually use it every day andit has really good quality picture.
And like with the Vision Pro,it's just I know I'm gonna have
52 paper cuts, like dealing withit and just being like, I'm done.

(07:12):
I don't want to deal with this.
I'm just gonna go on my Mac.
So that's part of the problem toothat I saw was like, yeah, it's
a great experience when I do thedemo, but oh, the wifi doesn't work.
Or, oh, there's a problem withthe battery, or, oh, it's just,
I don't wanna deal with it.
I also too am a little bit.
I wouldn't say burnt, but I've gonethrough this with the Apple Watch, so

(07:32):
I'm like, yeah, like it's gonna take awhile even if it does become a success.
So I'd rather wait and see

Yeah, and I think Apple watch the nice
thing right now, especially,I think it came down in price.
It feels I don't know, butit's not super expensive.
Right.
An Apple watch, I think.

No, it's nowhere near as

so popular.
I think if they could make a visionheadset for a thousand to 1500, like
basically iPad pricing, I think a lotof people would be interested in it.

I mean, you still have to have the app market for it.

Yeah.
Yeah.
But at least then it would be withinreach for more people and Yeah,
it wouldn't be as hard of a sell

Yeah.
I guess for me to spend that kindof money, I better use it like every

for sure.
For sure.

That's my philosophy, but that's not the consumer market.
So who am I to judge?
Right.
All right, so let's talkabout the big thing this year.
Well, it's been a big thing forthe last few years, but Apple
finally dipped their toes into ai.
Where are, where's Donnie at with ai?
When can we expectthat practical AI book

A practical AI book.
I don't know if I would be theperson to write that to be honest,
but I started the year I've beenon a journey with AI this year.
I think I started the yearpretty skeptical, basically like
echoing what everybody else says.
Oh, but I can't even ask it to do this.
I very vividly remember a demo fromPaul Hudson two years ago what,

(09:01):
yeah, like almost two years agoat iOS g where he said, let's try
to use chat g PT to build a clock.
And he just told it like, write me aSwift app that this display a clock
with moving hands, like a detailedprompt and it kept like messing up.
I'm pretty sure the same demotoday would be much better, but

Jordy did that at

none of that.
That was Paul Hudson a deep ashe did something similar kind

No, Jordy did Jordy did it at Deep Dish and it was a success.
It wasn't a clock, itwas something else, but

because with the clock the messy thing was
that it kept messing up, like theangles of the arms and everything.
But I'm pretty sure that if PaulHudson did the same demo, like today
would be so much better already.
And I think for me, the thing to figureout, and the thing that I learned
throughout the year is, okay, so thisAI thing is only getting bigger, right?

(09:50):
And we can keep saying, oh,but it can't even build an app.
Oh, but it can't even do this.
Oh, but I have to do this.
So I was kinda like, so what can we do?
Right?
It's so easy to give itstuff that it can't do.
And then I started just asking for likesmall bits Hey I want to clip an image.
I have a UI image.
I wanna clip it to be a circle.
How do I do that?

(10:11):
And it just gave me smallsnippets, small bits and pieces,
and just put them into a project.
And then cursor came out.
And I took that and peoplewere saying like, it's so
great, you can do anything.
So I was just like, okay, let's see.
So I told it like, write me code tomake an API call to the most useful

(10:32):
stock market app that's out there.
I was like, well, you can usethis one, but you need an API key.
So I was like, I don't haveAPI keys, just gimme anything.
And it just used like some stock market,API, and it would just make a request.
So I'd use documentation, everything,and it actually worked right?
And then I told it like, okay,make a website out of this.
Allow me to do this,allow me to do that.
And I think I actually was able tomess with it for four to five hours,

(10:54):
adding features until something broke.
And I wasn't able to figure outwhat, 'cause my experiment was, I'm
not going to review the code, right?
I'm just going to tell it what I want.
Accept any suggestion.
But like being able to go over 45hours before it actually messed up
to a point where I couldn't fix it.
I think it was really cool and Imade it, I made a website, right?

(11:15):
It wasn't an app, so it wasn't somethingI was very comfortable with per se.
And so that's when I learned okay,so if we actually use an AI tool like
this, we can ask it for small bits.
Small pieces.
We still have to break down the problem.
We can ask it to help breakdown the problem if we want.
We're still breaking down the problem,we're still finding the pieces.
And so now I actuallyuse AI all day long.

(11:37):
I have copilots fromGitHub inside of Xcode.

Okay.

And so one of the complaints people have about that
is it can't do like whole apps orlike whole views or stuff like that.
And I actually like it that way.
I like just today I was writinglike a function to, to expand
a rectangle into a circle.
So I was like, okay, where's themiddle of the rectangle start?
Mid X equals.
It was like, oh, you have a rectanglemaybe when you use it's mid X position.

(12:02):
And then I hit enter and it was like.
Maybe you also wantto have mid midwives.
Yeah, I want mid y.
And then it just kept like hintingone or two lines at a time.
And that's super nice 'cause it's veryeasy to digest for me as a developer
and it just saves me keystrokes, right?
Like it's it's reading my mindin terms of what I want to do
and that's where I like it.

What do you think?
You said used copilot.
Did you try using the applegenerative AI stuff in Xcode?

So, so only the Swift Auto Complete thing,
which is complete garbage.
Like I didn't try it.
I think 62 is supposed to be better.
But for me it's I disabled it afterI, I wrote dot frame and it suggested
dot frame with colon space, space,space, space, space, biodegradable.
And I was like, that's not even valid.

(12:49):
Swift.
Like, why are you suggesting this?
This is ridiculous.
You're supposed to be, I.
Good at Swift.
That's when I was like, okay,this thing is getting in the
way more than it's being useful,

Yeah.

but like I said, 16 two is supposed to be better.
So maybe I haven't tried that.

okay.
Okay.
That's supposed to be outprobably by the time this
comes out, so yeah, we'll see.
How about AI on the apple and asfar as guess intense is the way of
getting at it or any of that stuff.
Have you looked into any of that?

Only at app intents, but I haven't had
a chance to play with it yet.
Mainly because I don't have I'm inthe EU right, in the Netherlands.
So we don't get Apple Intelligence yet.
There are ways around it.
I could probably get around it.
Main issue is that my iPhone15, is it that, yeah, it's 15.
The one that supports it.

has to be a pro.

Yeah, it's the 15 pro that I have, but it's my main device
and I tried getting Apple Intelligenceon my Mac where it is allowed, but
I had to put Siri in English andall of my HomeKit stuff is in Dutch.
So that kind of broke, likeall of, yeah, it, it broke
all of my voice commands.
I was like, okay, fine.
I'll,

there a country.
There's gotta be a country, likeone of the Caribbean islands.
A Bermuda Island wherethey're still their

yeah, so, so, so the issue is Apple limits this to us English

Oh, right, right.
Okay.
Not even like Canada.

I think you have to have, I think it's allowed in
Canada as a country, but it's stilllimited to devices with US English
language or something like that.

That's stupid.
We didn't talk about job market.
Should we cover thatbefore we close out?

Why not?

How's it been for you?

So the job market it's, I've never
fully relied on the job market.
Like I always made a point outta that.
So that's been good for me.
And I tend to favor long-termengagements with clients.
And so I had two of thosegoing into the year.
One I got out of around March withno urge to replace that client 'cause

(14:55):
I just had a baby at that point.
So I was like, you know what, it's nottoo bad having one, one fewer client.
And so towards the end of the year I wasthinking maybe I should start thinking
about replacing that client now.
I heard a lot of people sayoh, job market is not good
job market is not good.
Luckily I had two people reach outa couple weeks ago, but that said,

(15:18):
those are probably the only twopeople that I had in my inbox for
eight months or something like that.
So, so I can definitely tellthe job market's not great.
I do feel like it's somewhat likepicking up, but I'm wondering if it's
companies spending their budget, right?
Because it's end of the year, if youdon't spend your budget, you might

(15:39):
not get new budget for the new year.
So maybe they're like,

right, right.

Donny Wals (guest): we should spend that. (15:43):
undefined
So it's always hard to sayat the end of the year.
I think it's tough times overall.
I don't hear anybody sayingthat jobs are just there

Why do you think that is

Donny Wals (guest): I'm not exactly sure. (15:54):
undefined
I think a lot of companiesprobably spent too much money.
In general, like they have just ingeneral for the past many years, right?
So it's not something new likeall the fees, capital and stuff.
It seems like money is infinitein some tech companies.
I do feel like a lot of people arelooking forward to super high paying

(16:15):
jobs that you almost only get likeat the big five or six companies
that are from the us especially inthe Netherlands, there's like only
three or four of them that willpay you like 150 k plus and that's
where people want to be, right?
So I feel like a lot of smallercompanies are just struggling to
hire and they're not as visible.

(16:37):
So yeah.
Why is it?
I'm not exactly sure.
I think there's a lot of expectationsfrom developers in terms of pay.
I think there's probably also a lotof companies just trying to cut costs.
I wonder how much the politicallandscape in the US elections and
stuff, maybe companies were reluctantto spend around that because I do

(16:58):
feel like, that's always a timewhere a lot can change for companies.
So yeah, maybe they were reluctant tospend for a bit, but it's hard to say.
I dunno if you have any thoughtson that, but I find it hard to say

Okay.
Do you think it'sgonna get better in 25?

maybe, I think in 25 you'll see that AI
knowledge and AI skills will bemore important than they are today.
So I wonder if the kind of person thatpeople are looking for will change.

I mean, do you think AI is gonna continue getting
better and better, or do youthink there's a ceiling to it?

I think there will be a ceiling to it.
I don't think we've reached it yet.
Right.
We have these like super broadmodels that can do a lot of things.
I wonder if we're going to seelike more specific models or like
interfaces for more specific stuff,

That's a really good point.
Yeah.

come up things like rag, I forgot what it stands for, but
what it means is like inserting moreand more context into your LLM, right?
So when you have your code base, yougive it all the files, you teach it
like what's in your code base, and thenfrom there you're gonna ask questions.
I think it's only goingto get better at that.
Yeah.
See, I think

Yeah, I agree a hundred percent.
And I think there's too much of anemphasis on trying to build a sentia
general intelligence and not like anactual tool that's useful to people.

exactly.

I think what's the, that realization comes in, I think,
yeah, that, that's a really good point.
I hadn't thought of that.
Where like you build a morespecific model to do stuff.

And I, I always say I don't think AI is going to
come for our jobs as developers.
I do think if we ignore AI and wejust keep pretending that it's not
going to do anything for us for longenough, I think you'll be out of a
job because there's just going tobe like junior developers or other
people running circles around you.

To have developed not before the internet, but like
when it wasn't quite as useful.
And I feel like it's like theinternet where it's like now
you could just look up stuff.
You don't

Yeah.
Maybe.

figure it out on your own or look up a paper manual where
now it's like you can actually justask ai a question and it does its best.

it also reminds me of like other transitions, right?
Objective C to Swift, where therewere definitely objective C developers
saying, oh, but you should still knowhow objective C works under the hood.
Otherwise you can't bea great iOS developer.
They were right for a couple yearsand now it's just who cares what
objectives see Similarly UI kitto Swift ui, but oh, but all these
newcomers, they only learn Swift ui.

(19:35):
They're never going to be able to buildlike a great app five years later.
Who, who still touches UI Kit?
Like

Right,

a few people.
Like you can learn Swift UI onlyand I'm pretty sure AI is going to
be a similar story where people aresaying now oh, you don't need it, or
Oh, you still need to know exactlywhat Dakota is doing under the hood.
There's for sure value in that.
But I think a lot of companies alsojust care about getting stuff done
rather than, oh, right, there's abetter way to write this for Loop.

(20:02):
You should have used that.
I was like, yeah, maybe,but this one works too.
So yeah I think there's going to belike a lot of room for people to do ai
and I'm not saying it's not importantto learn how to go code really well.
I think that's extremely important,but I think people that solely
focus on that and not, and don't.
Look at how AI could be enrichingtheir workflow rather than,

(20:25):
fighting it all the time.
I think that's where people are wrong.

Yeah, I agree a hundred percent.
Donnie, thank you so muchfor coming on the show.
I really appreciate it.
Where can people find you online?

Also all the social networks these days, right?
Mastered on threads, blue sky X.
But if you actually wanna havea conversation with me, blue
Sky is where I'm mostly at these

Okay, sounds good.
You can find me on all the socialnetworks as well at Leo g Dion.
My company is Bright Digit.
If you really enjoyed thispodcast, please and subscribe
and give me a review or if you'rewatching this on YouTube as well.
I'm trying to think what elseI was supposed to mention.
If you haven't had a chance,try out Bushel should be

(21:08):
version two should be released.
So give that a try and I'mtrying to think what else I
have to mention on this episode.
Should always have thatin my notes, shouldn't I?
If there isn't, I'll probablyhave had it in the intro.
So thank you again and I look forwardto talking to you in the next episode.
Bye everybody.

Bye.
Thanks for having me, Leo.

All Episodes

Practical Year - Part 2 with Donny Wals

Episode Transcript

Popular Podcasts

Stuff You Should Know

Dateline NBC

On Purpose with Jay Shetty

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Practical Year - Part 2 with Donny Wals