0:00

and welcome to Python Bites,

0:02

where we deliver Python news

0:04

and headlines directly to your

0:06

earbuds. This is episode 429,

0:08

recorded April 21st, 2025. And

0:10

I am Brian Ockett. And I am

0:12

Michael Kennedy. And this episode is sponsored

0:14

by the folks at Posit Workbench. Thank

0:16

you, Posit. Also, listen to them later

0:18

in the show, of course. If you'd

0:20

like to connect with us, please do

0:22

so on Blue Sky or Mastodon. We

0:24

have all of those links in the

0:26

show notes. We'd love to hear from

0:28

you, love to hear new topics that

0:30

you think we might want to cover.

0:33

And if you'd like to listen to

0:35

us live, head on over to pythonbites .fm

0:37

slash live, usually Mondays at 10, Pacific

0:39

time. But you can also use that

0:41

link just to find the YouTube channel

0:43

to watch older episodes too. And

0:46

finally, you don't

0:48

need to write anything down while you're

0:50

listening because we'll just send it

0:52

to you. So head on over to

0:54

pythonbites .fm. Join the mailing list or

0:56

the newsletter list, and you'll just

0:58

get an email sent every week with

1:00

all the links. But it also has

1:02

backup information of what you need to

1:05

know to understand the story. So really

1:07

nice. And we don't use it

1:09

for spamming or anything like that. Michael, what

1:11

do you got for us? I might

1:13

have a problem. I like to self -host

1:15

stuff too much. It means

1:17

I end up with things that I

1:19

have to take care of rather than

1:21

just have a login somewhere. But I'm

1:23

going to make a recommendation nonetheless. So

1:26

if you find yourself

1:28

using Slack, I think

1:30

even Zoom maybe, if you're

1:32

using Jira, if you're using Notion,

1:34

and you would like that all in

1:37

one place rather than a bunch

1:39

of different apps that you pay for,

1:41

for free, for open source. self -hosted

1:43

or paid as hosted. If you

1:45

really want to do it that way,

1:47

there's a project or tool or

1:50

platform, whatever called Holi, H -U -L -Y,

1:52

an open source platform that serves as

1:54

an all in one replacement for

1:56

linear JIRA, Slack, and Notion. How cool

1:58

is that? Okay. Okay. So maybe

2:00

you want to have the place

2:02

to do chat conversations or you

2:05

want to store your documents or

2:07

you want to do project management

2:09

and sync it with your GitHub

2:11

issues. or do planning. So

2:13

all of this stuff is super

2:15

cool, I think, and even has video

2:17

meetings. So if you want to

2:19

also get rid of Microsoft Teams or

2:22

Zoom or whatever, it gives you

2:24

a nice private way to do all

2:26

that, right? And I don't know,

2:28

it just that really resonates with me

2:30

as just like, here's this cool open source

2:32

thing that we can do and we can

2:34

run it and we don't have to have

2:37

all these different services. And I know

2:39

about the others. I don't

2:41

know the pricing for Jira. I've never

2:43

paid for Jira, but slack is

2:45

out of control like slack is super

2:47

expensive for what you get from

2:49

it and things like that. And so

2:51

having this with a bunch of dashboards

2:53

and it even has a nice

2:55

self hosting option. So if you are

2:57

a person who does Docker compose,

2:59

guess what? You just Docker compose up dash

3:01

D you have your whole platform running, which

3:03

I think is pretty excellent. So that's how

3:05

a lot of these self hosted things are

3:07

shared and maintained. So you don't even have

3:10

to figure out how to put it into

3:12

Docker. It's already there and set up. So

3:14

you just run it and you're good to

3:16

go. Just make sure you do backups. Pretty

3:18

neat, I think. That's pretty cool. Yeah. And

3:20

it also syncs both ways with GitHub. So

3:22

if you've got GitHub issues, it

3:24

will sync with the issues. And if

3:26

you have GitHub projects, it'll sync with

3:28

those projects. So it's project management tools and

3:30

it's issue management, like it's JRO replacement

3:32

stuff is mirrored on GitHub. So not

3:34

everybody has to use it. Like you

3:36

can have internal people on this and

3:38

external people just perceiving it as GitHub.

3:40

I have GitHub issues. I was just

3:42

talking to my therapist with it about

3:44

it the other day. Yeah,

3:47

that's indeed. And John Other

3:49

says, this is why I love

3:51

the podcast. Thanks for the holy

3:53

recommendation and loves self -hosting as much

3:55

as I do. So also thanks,

3:57

John. Holy self -hosting Batman. Anyway, holy

4:00

self -hosting. Indeed. Yeah. The

4:03

problem is you can, like, well, now I've

4:05

got 12 apps to back up and maintain,

4:07

but it is super cool to be able to

4:09

say we don't have to worry about data

4:11

privacy. We don't have to worry about sharing

4:13

things or if those places get hacked or

4:15

if they change their business model or if they

4:17

go out of business, you know, you got

4:19

a self -hosted open source thing that you can

4:21

fork and just run. And there's something cool

4:23

about that. I wonder if anybody's got like,

4:25

because this sounds great, but I don't really

4:27

want to do that work. So I wonder if

4:30

there's a self hosting as a service service.

4:33

I think they're actually, I'm forgetting the name of

4:35

it right now, but yes, basically

4:37

is a self -hosting as a service. Yes,

4:39

it's amazing. Okay. What

4:41

an interesting idea. All right, that's

4:44

it for this one. Over to you.

4:46

Okay. Well, I am going to

4:48

talk about critical, what are

4:50

those called again? CVEs? Common. call,

4:52

I should have practiced this,

4:55

common vulnerability and exposures. So

4:57

CVEs, we're used to talking about

4:59

these when there's like really wide

5:01

scale attacks, but they kind

5:04

of happen all the time,

5:06

like vulnerability problems. And

5:08

this propped up this

5:10

last week rather urgently,

5:12

because the CVE system,

5:14

the entire system is

5:16

sort of built on

5:18

top of an a

5:20

non -for -profit called MITRE.

5:25

But there was a

5:27

contract with the US

5:29

government to maintain this database

5:31

of CVEs. However, and

5:33

it's been around for 25

5:35

years, it ran the

5:37

risk of possibly going away

5:40

because of all of the cost

5:42

cutting that our current lovely

5:44

administration is doing. It's

5:48

a waste. It's government waste, I tell

5:50

you, Ryan. Like what? And

5:52

it's not even like a lot of

5:54

people involved in this are volunteers anyway. That's

5:56

just nuts. Anyway,

5:58

so a 25 year old

6:00

CVE program ran the risk

6:02

of going away. There was a

6:04

letter that came out from

6:07

the vice president of MITRE gave

6:09

notice of potential halt of

6:11

operations. And apparently they

6:13

had been worried about this

6:15

for some time. So

6:17

in order to deal with

6:19

this and possibly make

6:22

sure that we don't have

6:24

this risk in the

6:26

future, there is now a

6:28

CVE foundation. So there's

6:30

been an announcement as of

6:32

April 16th that the CVE

6:35

foundation has formally been formally

6:37

established to ensure long -term

6:39

viability, stability, and independence

6:41

of the CVE program. This

6:43

is really cool. It isn't

6:45

something they just suddenly did.

6:47

They've been thinking about it

6:49

for a while and planning

6:51

it. There's an

6:53

announcement at thecvefoundation .org.

6:56

They're going to release information about

6:59

the transition, what is structured

7:01

as the transition planning opportunities for

7:03

involvement later, but right now

7:05

there's just an announcement. This is

7:07

pretty crazy that we would

7:09

run the risk of losing this.

7:12

This is how we talk

7:14

about vulnerabilities. But

7:16

apparently, there was an announcement also,

7:18

I couldn't find the link to it,

7:20

that they did not lose funding.

7:23

So it's okay for now, but it's

7:25

still the for now part. So

7:27

the foundation wants to make sure that

7:29

it's not just for now. I

7:31

still think it's something that we should

7:33

fund as a government, but you

7:35

know, is what it is, I guess.

7:38

I think we should fund it.

7:40

But also, I feel like maybe

7:42

that should be more of like

7:44

kind of in the style of

7:46

Python or Mozilla Foundation or, you

7:48

know, that it's tied to the

7:50

US government rather than just an

7:52

international organization of people who are

7:54

really committed to tracking security issues.

7:56

Doesn't take insane amounts of funding

7:58

to track these things. You know

8:00

what I mean? I have no

8:02

idea what the work is involved

8:04

for this. Yeah, I don't

8:07

know what the work is, but

8:09

it's not like there's a lot

8:11

of server infrastructure. It's not running

8:13

like. AI farms or something. The

8:15

funding probably should be coming from

8:18

all the ISPs and big companies

8:20

and stuff that are benefiting from

8:22

this. I think so too,

8:24

but hey, I'm not against the US

8:26

government. I'm glad they were doing it,

8:28

but it just puts it in a

8:30

weird situation these days. And

8:32

Python became a CVE authority not

8:34

too long ago, the PSF, so

8:37

they can announce their own CVE.

8:39

CVEs around things in the Python

8:41

space without going through an external,

8:43

not convincing some other participant to

8:45

allow them to list their CVE

8:47

for Python and so on. So

8:49

there's a bit of a distributed

8:52

aspect of it. Now, before we

8:54

jump on to thinking our sponsor,

8:56

Brian, the thing that I was

8:58

thinking of is Alestio, E -L -E

9:00

-S -T -I -O, and it says

9:02

fully managed DevOps or your cloud

9:04

and open source software. And

9:06

I've not used this. It's not a

9:09

recommendation, but we, as in they, deploy

9:11

and manage open source software to your

9:13

cloud provider choice. So, you know,

9:15

create a Hetzner server, point it at it,

9:17

and then pick the various self -hosted things

9:19

that you want, and they will self -host them

9:21

there for you. I believe it's how it

9:23

works. Awesome. Self -hosting is a service. Who

9:25

knew? Who knew? Well, I

9:27

guess you did. But we also have,

9:29

we also have Posit doing pretty awesome

9:31

stuff. Why don't you tell them about

9:34

it? This portion of Python bytes is

9:36

brought to you by the folks at

9:38

Posit. Posit has been making huge investments

9:40

in the Python community lately. Known

9:42

originally for RStudio, they've been building

9:44

out a suite of tools and services

9:46

for Team Python. Have you thought

9:48

of all the things that go into

9:50

a Python data science project? You

9:52

need your notebook or IDE for sure.

9:54

But you also need a server

9:56

or cloud environment to run it. A

9:59

version of Python, packages, access to

10:01

your databases and internal APIs. That's

10:03

a lot to set up. And

10:05

if you change any of these

10:07

things when you return to your

10:09

project months down the road, you

10:11

might get different results. Wouldn't it

10:13

be nice to have all of

10:15

this set up for you in

10:17

one easy to access place whenever

10:19

you want to get work done?

10:21

That's the goal of Posit Workbench.

10:23

Posit Workbench allows data scientists to

10:25

code in Python within their preferred

10:28

development environment without any additional strain

10:30

or on IT. It gives data

10:32

scientists access to all the development

10:34

environments they love, including Jupyter Nobix,

10:36

Jupyter Lab, Positron, and VS Code.

10:38

And yet it helps ensure reproducibility.

10:40

Here's how it works. You or

10:42

your team set up Positworkbench on

10:44

a powerful, dedicated server within your

10:46

organization or on the same cloud

10:48

service that is hosting your most

10:50

important data sources, such as AWS,

10:53

SageMaker, Azure, GCP, Kubernetes, or

10:55

pretty much anywhere. There, you

10:57

create dedicated pre -configured environments to

10:59

run your code and notebooks. And

11:02

importantly, you also configure

11:04

access to proprietary databases

11:06

and internal APIs. When

11:08

it's time to onboard a new data

11:10

scientist or start a new project,

11:12

you just fire it up in Workbench,

11:14

and it's fully configured and ready

11:17

to go, including the infrastructure side things.

11:19

All of this is securely administered

11:21

by your organization. If you work on

11:23

a data science team or consistency

11:25

matters, you owe it to you and

11:27

your org to check out Posit

11:29

Workbench. Visit pythonbytes .fm slash workbench today

11:32

and get a three month free trial

11:34

to see if it's a good

11:36

fit. That's pythonbytes .fm slash workbench. The

11:38

link is in your podcast player show

11:40

notes. Thank you to Posit for

11:42

supporting Python Bytes. Indeed. All right,

11:44

you ready for the next one? Yeah.

11:46

Database. Database things. So here's

11:49

a really interesting. free web app,

11:51

I guess it is, that

11:53

lets you draw and import and

11:55

export and visualize database diagrams.

11:57

So either you've got your own

11:59

project or where I see

12:01

this being super useful as you

12:03

put onto a new project

12:05

or you're a consultant. And

12:07

they're like, and welcome for the two

12:09

weeks. Here is the database and here's

12:11

the app. Please fix it by now.

12:14

And you're like, uh, how do I

12:16

even get started? What is here? Right.

12:18

So this thing called draw db .app

12:20

allows you to draw, copy and paste

12:22

database diagrams. And if you go there,

12:24

you can see there's really nice graphics

12:26

and the UI is quite nice for

12:28

interacting with it. So it says, you

12:30

can try this for yourself for free.

12:32

And, uh, it, what's interesting is it

12:34

asks you to choose your database. as

12:37

in like SQL light or Postgres or

12:39

SQL server or whatever, because it imports

12:41

and exports SQL statements. And those different

12:43

databases have different database SQL dialects, which

12:45

in and of itself is annoying. But

12:47

let's just say, I don't know, I'll

12:49

do Postgres, right? And you come in

12:51

here, you can add a little table,

12:53

and you can then like edit that

12:55

thing, give it a column, multiple columns,

12:57

different data types. and create a second

12:59

one then you can say like grab

13:01

one column from one database or one

13:03

table rather and then drag it and

13:05

drop it in a field on another

13:07

another column and that'll create a foreign

13:09

key relationship automatically for example and you

13:11

can go over and you can say

13:13

file export sequel or import from sequel

13:16

and that'll generate the data definition language,

13:18

DDL stuff, create scripts and create the

13:20

indexes and columns. Or if you've got

13:22

a database, you can export it and

13:24

then load up this diagram based on

13:26

what was in your database and then

13:28

visualize it, tweak it, save it, or

13:30

just try to understand it. That's really

13:32

cool. Yeah. And as far as I

13:34

can tell, it's free. I don't know.

13:36

Maybe there's some point where I pay

13:38

for it, but I don't think so.

13:40

So anyway, I think it's a great

13:42

little app and people should check it

13:44

out if they have databases they want

13:46

to visualize. And you know, I'm usually

13:48

starting with a drawing anyway. like drawing

13:50

it on paper. So why not just

13:52

draw it in something like this? Exactly.

13:55

Because then you could say generate my table

13:57

from this. Yeah. Yeah, that's cool. Neat. Cool.

13:59

So not much more to it, but there

14:01

it is. All right.

14:03

Well, my last item is a,

14:05

and I'm only going to cover

14:07

part of this, but it is

14:09

a blog post by Edward Lee

14:11

called 14 advanced Python features. And,

14:14

you know, it's a listicle sort of

14:16

a thing, but there's a lot of

14:18

those like advanced Python features and even

14:20

talks about this that are really, really

14:22

not that advanced. They're just stuff that

14:25

people should know and some fun things

14:27

for, you know, advanced for beginners, but

14:29

not really. But I

14:31

kind of, I kind of really

14:33

like this because there are

14:35

things in this list that I

14:37

really wish I would have

14:39

learned earlier. And so anyway, I'll

14:41

just jump in. There's a few things I wanted

14:43

to pick out. First off

14:45

is typing overloads. And this is

14:47

something that I just learned while reading

14:50

this article. I didn't know you

14:52

could do this. So within the typing

14:54

module, you can say from typing

14:56

import overload. And then

14:58

one of the things you can

15:00

do then is you can

15:02

essentially list overloaded operations, overloaded

15:04

definitions for a function call.

15:06

And it's not really like

15:08

full function overloading like we

15:11

have in C or something

15:13

like that. However, there's return

15:15

types. So let's say there's

15:17

an example here that if

15:19

you only pass in a

15:21

certain type, then you're always

15:23

going to get a list

15:25

of strings. And if you're

15:27

passing in a different type,

15:30

then you always get a

15:32

single string back. Those

15:34

sorts of things are nice to

15:36

have for typing or return types. And

15:38

that's something that we don't really

15:40

have in Python. You can't have a

15:42

different difference in just return type.

15:44

So having that in place is kind

15:46

of neat. That's neat. I'm

15:48

going to play with this like right away. So that's

15:50

pretty cool. Overloading functions with

15:53

the typing. So I'll have to

15:55

try that. Next up is something

15:57

I've been using a lot lately

15:59

is a keyword only and positional

16:01

only arguments. And specifically, so

16:03

we now have these a star

16:05

or a slash that you can

16:07

separate the parameters to a function.

16:09

the parameter definition. And

16:12

the the asterisk or star

16:14

means that everything, everything after

16:16

that is keyword only parameters. And

16:19

then the slash is

16:21

positional only parameters. And

16:23

that's that's everything before.

16:26

So it's a little one of them's

16:28

before and one of them's after. So in

16:30

his example, he's got a b and

16:32

then slash c d star e f. So

16:34

mean that means a and b are

16:36

positional only, c and d can be

16:38

positional or or keyword,

16:41

and then ENF have to be

16:43

keyword only. And the thing

16:45

that I'm doing a lot is

16:47

why I'm using the keyword

16:50

only one a lot is for

16:52

functions that have, and these

16:54

are not usually API functions, but

16:56

functions, internal functions that have

16:58

a lot of parameters that have

17:00

defaults, and you would almost

17:02

hardly ever pass it like just.

17:05

positional only because the defaults are, it

17:07

doesn't really matter the order. It's just

17:09

they all have defaults and there's a

17:11

bunch of them. So I really want

17:14

all the callers of that every place

17:16

we're calling the function to list which,

17:18

which variable or which parameter they're, defining

17:20

as they call the function and you can

17:22

do that with the asterisks. Super cool. That's

17:24

really cool. Another thing that I think is

17:27

really useful for that is if it's the

17:29

same, a lot of times you're going to

17:31

give it. like numbers are true and false

17:33

where you don't have a variable that you're

17:35

passing in but you have just um yeah

17:37

kind of constant because if it goes seven

17:39

seven five true true false you're like whoa

17:42

whoa whoa what what It's

17:44

not like variable names are there where like x

17:46

comma y comma z, like, oh, those are

17:48

the dimensions. No, it just goes seven, seven, five.

17:50

Like, hmm, which is which,

17:52

you know what I mean? Especially true, true,

17:54

false, true, something like that. If you have

17:56

the, if you force keyword arguments on it,

17:58

then it's a much more readable thing at

18:00

the call site. Yeah. Things that are like

18:02

generic, like you're just adding things. It doesn't

18:04

really matter. You know, it doesn't really matter

18:06

that your ad function is A and B,

18:08

but, you know, for true and false, you

18:10

really want, what do those mean? I like

18:12

this. It's a good addition. Last

18:15

thing I want to come to

18:17

is there's a list of, he

18:19

said number nine is Python

18:21

nitpicks, which is really a

18:23

few topics around it, but

18:25

it's listed as a nitpick

18:27

because it's a bummer when

18:29

people aren't utilizing this. So

18:32

the 4L statement, and this is,

18:34

I think probably still controversial, is

18:36

maybe a little bit. is whether

18:38

or not you should utilize the

18:40

else clause in four statements. And

18:43

kind of, you know, it's like often

18:45

before the else clause or without using it,

18:47

you might have to say like something

18:49

that his example is like a found flag

18:51

to say, you know, whether or not

18:53

you actually found the item you were looking

18:55

for while you're iterating the for loop.

18:57

And then you can check that later. But

19:00

there's there's the else you could just

19:02

say, you know, if you didn't find if

19:04

If you never hit anything inside the,

19:06

the for loop, you can else out. Um,

19:08

it's, it's still a little weird though.

19:10

I still find it very, uh, very good

19:12

to make sure that you comment that

19:14

to say what's, what's going on in the

19:16

else you're using what you're doing in

19:18

there. So, okay. I'm, I'm anti else. I'm

19:20

definitely else. And by the way, Guido,

19:22

uh, I heard him quoted at one point

19:24

that said, if I had to do

19:27

it over again, there would be no else

19:29

statement. I think it's just weird. It's

19:31

like, does it happen when it breaks? Or

19:33

is it happened when it. doesn't break

19:35

like is break the thing you're looking for

19:37

and else is the other or is

19:39

break something weird and it was supposed to

19:41

go like it's just I don't know

19:43

it just I know you can save one

19:45

line of code but it's it's too

19:47

ambiguous to me it's too weird so else

19:49

is what else is if you didn't

19:52

break right I think so yeah anyway I

19:54

think so anyway it's not the fact

19:56

that you gotta like we gotta have this

19:58

discussion it's like I don't make social

20:00

weird for me I'm out I'm I'm I

20:02

know there's I could do it but

20:04

I don't do it okay Also, the walrus

20:06

operator, it's been around since 3 .8, and

20:08

3 .8 is already deprecated, or end of

20:10

life. So we can start using, definitely

20:12

use the walrus operator. Again, it's just saving

20:14

one line of code, but. I like

20:17

it. How about a walrus? I'm a fan

20:19

of walrus. I created the walrus operator

20:21

this weekend, I believe. For

20:23

me, I like it because it's

20:25

the locality of definition. I'm

20:27

creating it for this if block and I'm

20:29

going to use it in this if block

20:31

if I need it. Otherwise, it's kind of

20:33

like it's part of this thing, not something

20:36

that might make sense later down the line,

20:38

probably. For me, I like it. A

20:40

couple more. Short

20:43

circuit evaluation, I don't

20:45

really care. I'm fine with

20:47

a bunch of FLSs, actually,

20:49

if that's all you can

20:51

do. But the short circuit,

20:53

which means utilizing OR to

20:55

say, if you're going to

20:58

do one thing or the

21:00

other, you can

21:02

use OR short circuits. So once you

21:04

hit one of them that's true, anything

21:06

after that's not going to get ran.

21:08

And you can kind of go crazy

21:10

with that, though, and actually put logic

21:13

in there. And I'm really not a

21:15

fan of putting logic in the short

21:17

circuiting or operation. But that's just me.

21:20

But I am a fan of operator

21:22

chaining. And I often see this with

21:24

people coming from different languages. They don't

21:26

know you can do operator chaining. So

21:28

I'll see, like, if 0 is less

21:30

than x and x is less than

21:32

10, you don't have to do that

21:34

in Python. just put them together and

21:36

say zero less than x less than

21:38

10. So operating training is right. But

21:40

also, I think that we should have

21:42

been a little more strict with operator

21:44

training. And I don't think

21:46

it like you can put anything

21:48

in there, right? But I

21:50

don't think you should like for numbers,

21:52

it should be less than you should

21:54

not be doing like greater than operator

21:56

training. That's just weird. The number line

21:58

is small to big. So please do

22:00

that. That's all I wanted to cover

22:02

for that. But there's a bunch of

22:04

other great stuff in here. So

22:07

definitely check out this article.

22:10

Here's the full list. So LRU cash. There's

22:12

a lot of stuff in there. Yeah. Love

22:14

it. F strings. Love it. Nitpicks. I have

22:16

some nitpicks with that statement. But you know

22:19

what? It's fine. Good, good, good fine to

22:21

the Brian. All right. Well, we're done with

22:23

our main topics and I don't have any

22:25

extras, Michael. But do you have

22:27

any extras? Well, I thought the answer was

22:29

no, but it turns out to be

22:31

yes. Okay. Because something I just heard about

22:33

right before was, oh gosh, assuming on

22:36

this is all weird. So there's, remember I

22:38

wrote this article that said unsolicited advice

22:40

for Mozilla or Firefox or something like that

22:42

scene. You know what? The stuff that

22:44

you guys are doing is not leading you

22:46

in the right path. And here are

22:48

five ideas that you might write, try as

22:51

a business, to exist. down the

22:53

line, please do that. And

22:55

so there actually come, I mean,

22:57

I really doubt they gave

22:59

a crap about what I said,

23:01

but they are introducing this

23:03

new suite of services. How about

23:05

that? So Thunderbird, it all

23:07

seems to be based around Thunderbird,

23:09

their email client. and Thundermail,

23:11

which is a really interesting term,

23:13

but they're offering like Thunderbird

23:15

appointment, which is kind of like

23:17

Calendly or Tidycal or whatever.

23:19

Thunderbird send for private file sharing.

23:21

I used to love Firefox

23:23

send, but it got abused by

23:25

hackers and other badness. And

23:27

so then it stopped working. But you could put just

23:29

like, here's a URL, here's a password. The

23:31

whole file just goes away in three days, give it

23:33

to someone, and then they could have it. It was

23:35

really great. And so I'll end encrypted, all that kind

23:37

of stuff. Also, some AI

23:40

thing, because of course there's an

23:42

AI thing, and then Thunderbird Mail,

23:44

hosted Thunderbird Mail, all that. Anyway,

23:46

I think this is a cool

23:48

idea. Thunder Mail. Thunder, Thunderbird.

23:51

So, very cool. That's

23:53

all I got for my extra. You

23:55

can't spell mail without AI. Sorry. Well,

23:58

I'll tell you what, that seems to be

24:00

what they think in their feature set. Every mail

24:02

client I find is like, and now

24:04

we have some terrible AI thing that will just erase

24:06

all your formatting and make you have to rewrite your

24:08

mail if you try to use it. But it's here

24:10

and it's great. No, it's not.

24:14

Anyway, shall we, you

24:16

know, Brian, we try to make it not

24:18

too political here, but I got a, I got

24:20

a political joke in a sense. Okay, awesome.

24:22

Are you ready for it? Yeah. So this one

24:24

came to us by many people. So thank

24:26

you to everyone who sent us this in. Have

24:28

you noticed? I don't watch the news a

24:30

ton, but have you noticed that there's some talks

24:32

about tariffs lately? Yeah.

24:37

I mean, look, I think genuinely it's

24:39

it's fair to say it like let's

24:41

Discuss tariffs and if other countries have

24:43

tariffs on us does it make sense

24:45

for us to not I don't know

24:47

whatever I think there's a debate that

24:49

can be had but the way that

24:51

it's been done is so just chaotic

24:54

and random and on and

24:56

off again and so on. But

24:58

somebody decided that if that's

25:00

a good idea for global trade,

25:02

boy, oh boy, wouldn't that

25:04

be a cool idea for Python

25:06

and for program languages, and particularly

25:08

for these pesky external outside

25:10

of the standard library packages. So

25:13

I present to you Tariff,

25:15

a Python package that imposes tariffs

25:17

on Python import statements. No,

25:20

it's not just a joke. It's literally...

25:22

Version 1, it's not even 0ver anymore,

25:24

and it is released on PyPI, so

25:26

you can literally pip install tariff. What

25:29

does it do? Well, boom,

25:31

fire, fist emoji, fire emoji, a

25:33

little reference back to a signal. The

25:36

greatest, most tremendous Python package that

25:38

makes importing great again. Tariff is

25:40

a fantastic tool that lets you

25:42

impose import tariffs on Python packages.

25:44

We're going to bring manufacturing back

25:46

to your code base by making

25:48

foreign imports more expensive. And

25:51

so all you got to do is

25:53

import it. You set your rate on the

25:55

different libraries, like 50

25:57

% tariff on NumPy, 200 %

25:59

terrible on Pandos, and so on.

26:01

And then when you import

26:03

NumPy, it's literally 50 % slower,

26:05

takes 50 % longer than before.

26:07

What do you think? That'll teach

26:09

him. Yeah. That'll teach him.

26:11

Yeah. We're going straight back to

26:14

self -hosted, vendering in. Yeah.

26:16

But what's nuts is it works. Like,

26:18

you know, it's not just a weird

26:20

joke idea, but somebody made it. It's open

26:22

source. And the hat. The

26:25

hat. The hat is

26:27

good. Why tariff, you may

26:29

ask? Because foreign packages have been stealing our

26:31

CPU cycles for too long. It's time

26:34

to put America first and make importing fair

26:36

and balanced again. Obviously, a parity package,

26:38

use it at your own risk. Yeah,

26:41

Other people stealing our CPUs, we need

26:43

to steal our own CPUs. Exactly. That is

26:45

how we're going to do it. Yeah.

26:47

Well, that's what I got for you. Is

26:49

it funny? I don't know, but I

26:51

think it is certainly amusing. It's When

26:53

I look at my 401k, it is not

26:56

funny. but. Gotta laugh. You got to cry.

26:58

Those are one of your two reactions. Might

27:00

as well last. Glad I like my job because I'm going

27:02

to be here for a while. Anyway.

27:04

Anyway, thanks Thanks everything, Michael. Thanks for

27:06

the joke. Thanks to everybody that shared them.

27:09

Like we said, a lot of people

27:11

sent that in, but that is not a

27:13

waste. That also gives us a signal

27:15

that we might want to cover it if

27:17

a lot of people are thinking about

27:19

it. So So thanks. Yeah. Thank you. All

27:21

right. Bye. Bye, Brian. Bye, everyone.