Algorithms + Data Structures = Programs - Episode 212: No More AI and Episode 211

Episode Date: December 13, 2024

In this episode, Bryce and I catch up and chat about the AI generated ADSP Episode 211.Link to Episode 212 on WebsiteDiscuss this episode, leave a comment, or ask a question (on GitHub)TwitterADSP: The PodcastConor HoekstraBryce Adelstein LelbachShow NotesDate Generated: 2024-12-10Date Released: 2024-12-13Gantt ChartADSP Episode 211: Power, Politics and Misconduct in C++ ✨ADSP Episode 211 GitHub DiscussionWhisper AIAudacityFinal SolutionBlog Post: On "Safe" C++ - HELL IN A REFCELLReddit CommentsNotebookLMIntro Song InfoMiss You by Sarah Jansen https://soundcloud.com/sarahjansenmusicCreative Commons — Attribution 3.0 Unported — CC BY 3.0Free Download / Stream: http://bit.ly/l-miss-youMusic promoted by Audio Library https://youtu.be/iYYxnasvfx8

I'll be honest, the main reason I still go to C++ conferences is because the majority, 99.9% of the C++ community in terms of speakers and attendees and the listeners of this podcast, I love interacting with that community. And it is sad that I have to miss out on that and not interact with that community. Welcome to ADSP, the podcast, episode 212, recorded on December 10th, 2024. My name is Connor, and today with my co-host Bryce, we chat about our thoughts on AI and LLMs, respond to feedback we got on episode 211, and elaborate on why we posted that episode. I have something to improve your mood. Look what I found at the NVIDIA vending machines.

Wait. Look what I found at the NVIDIA vending machines. Wait, do the NVIDIA vending machines require compensation? They do require compensation, sadly. Why do you, because they have free soda at the cafeterias. Do they not have sparkling water there? It's not currently functioning. Oh, that is a tragedy. The soda is broken. That is the number one

reason that I love going to HQ, is they have free soda, which is sad, because compared to all other tech companies, they've got free food, free snacks, free bunch of stuff. The only thing that NVIDIA gives... NVIDIA, we have free soda. Yeah, we get free soda,

which is... it's great. It's great. We're very cheap, but I'm going to open this up now. Okay. I was not as graceful as you because I got some Buble on my laptop's keyboard, but fortunately I have a keyboard cover. A bubbly. A bubbly.

Bubbly? I thought we were calling it Bublé. No, that's just what Michael Bublé wants you to call it. I've been good. How have you been? Oh, a lot of travel. I'm here at NVIDIA HQ for an internal Python workshop just for a few days. And then we, my girlfriend Ramona and I, we were in Poland, and we went to her hometown, which is called Ostralecka,

which means sharp grass. And that, I just think that's like the most badass name for a hometown. But I love Poland. It was super fun. You know what was not super fun was the sun setting at like 315 every day. Yeah, there's – I understand the Central European like stoicism better now because it gets dark and cold in the winter oh yeah well i just looked it up australica is uh it's north of warsaw and i've only ever been to wracko warsaw and wroclaw and so that is and poland's big so that's you're getting closer to

like poland is big you're closer to to, like... Poland is big. You're getting closer to the Scandinavian countries, which, yes, are famous for having a lack of sunsets in certain parts of the year. I don't, like, we have a co-worker, Jeff Hammond, who lives in Finland, and, like, he's, like, in, like, Helsinki, which is, like, the southern part of Finland, but still, like, I don't know how people, like, survive in the winters there when there's but still like, I don't know how people like survive the winter.

It's there when there's just like no sun. I'm sure, I'm sure Jeff has sorted it out, but it is shocking that the hotel infrastructure there has not discovered like blackout curtains. The only time when we were traveling through Scandinavia that they had, we had any, anything close to like darkness at night for sleeping was when we were on like Scandinavia that we had anything close to darkness at night for sleeping was when we were on the bottom floor of a cruise

that was taking us from Helsinki to Stockholm because it had no windows. Every other time we were in a hotel, and we had an option to upgrade to get one with windows, and we were like, why would we want to do that? We've spent a couple... You were in a cruise ship room with no windows? I didn't even know that that was a thing.

I thought that's where they put the crew. No, no. There's on the bottom floors of a cruise ship. And we were just doing it overnight to avoid having to drive and ferry during the daytime. And then we were like, this is perfect. We'll take the lowest know whatever the the lowest floor to get the maximum darkness it's great my first international trip was to um ulu which is

like halfway oh like oh you halfway l you oh you oh yeah we were there we were there it's in the middle of nowhere what were were you doing there? There was a committee meeting there. Are you serious? It was 2016, yeah. I assume they have an airport with flights in there. I assume so because that must have been how I got there. I remember connecting in Helsinki.

But it's like halfway between – it's like in the middle of Finland like halfway between halfway from the top? Vertically. Okay, because I was going to say it's like it's not halfway between anything my guy. It's just like it's like it's halfway between the bottom of Finland and the top of Finland kind of. And we were there in June and the hotel did

have blackout curtains which was necessary because the sun did not set. It got like a little bit dark at like 2 a.m. What time of year were you there? It was June. It was June. Like we were there for the summer solstice. What did we do?

I'm sure I didn't do it or whatever it was. We happened to be there because we drove through Finland And we actually loved Oulu. It was one of the highlights of the trip. Because they had this sauna on the river. That was like community run. And it was like super cheap. And you had to take like a little kind of raft. It was like you could easily swim out. But you take like a little raft.

That's like I don't know 10 or 20 meters. And then they have this like fire stove. very dark wooden sauna right on the river. And then like you just like you sit in there for however long you can last and then dive into the river. It was unreal. I'm not sure if you did anything like that while you were there. We didn't do that. But when Ramona and I were in Bergen, Norway in September, we went and did like the river sauna thing. And it was very cold.

But it's so good though. Like in between the sauna-ing, you only need to hop in for a second, right? Oh, I wouldn't go in. But she went in. And she's braver than me. But isn't when you get like you sit in a sauna for like 10 or 15 minutes and then you're just like dying. Like you're craving like ice cold. Like that's the benefit of like doing these hot cold circuit things.

It's like. You know, I take it back. I did go in, but like very briefly. She like went in for like a more legitimate period of time. Even a couple seconds does the trick, I think. If you're just hopping in and hopping it was like 45 degree it was 45 degrees in freedom units which uh i forget what the forget how you do the conversion i forget how you do the conversion that's like

a few degrees above celsius so 32 is freezing so but uh but it was a pretty like i we wanted to do one of the ones where it's like a boat, like a sauna boat. We wanted to rent a sauna boat and then just like go out on the fjords. That was a good trip. That was a good trip. And then we're going to Morocco for Christmas. Oh, wow. We have nothing planned.

We have hotels and that's it. I'm a little bit panicked. What's the safety rating? Is it one of those places where you have to cab from place to place because it's not safe to walk? Or is that just a rumor? Is that a different country? So that's the thing. I don't know, but I did hear somebody mentioned to me the other day that it's not the safest place.

So I feel like we probably ought to figure that out before we get there. I mean, a lot of places they it's not and it's just like you know there's parts of every city like i'm sure there's like in every city in america there's a more dangerous neighborhood than morocco actually is as like a country but um i there are certain places where like they actually have advisements and like curfews and stuff um so you just you just have to be smart about how you travel when we were in ecuador for we went on this cruise to the galapagos and we spent uh uh two days in guayaquil uh before and uh the cruise company was like very like they were like we will arrange your cab from the airport to the hotel and like don't leave the hotel because it's like super dangerous um and we just like we kind of like we kind of ignored that we hired a local guide and like you know we went out but then like you know we would walk around to go get coffee or something and like we never felt

unsafe but apparently we we started learning like after we'd spent the two days in guayaquil and like when we were on the cruise we started learning from the the like the the um the locals who working on the cruise, like more and more about what's going on in Ecuador. And we're like, oh, yeah, we probably should not have been walking around in the way we were. And then like the day that we left Ecuador, there was some news story about the ex-vice president of Ecuador who was in was in prison for i don't know for some reason he was in prison but they they broke him out of prison by they loaded up a drone with explosives flew it to the jail and like landed on the roof and then blew up the roof of the prison to try to get he imprisoned ex-president out nice president now and i was like yeah maybe this is maybe this is not the most stable uh country but you know i i i feel like if we just stayed in the hotel we wouldn't have had as good of a time and we never

felt unsafe so speaking of travel we're in the process of booking our stuff for your wedding. Very excited. I've never been to Mexico before. Looks like it's a nice place. I guess the ADSP listeners haven't had any updates. Technically, if you're dual listeners to my running podcast with my other buddy,

we talk about it more over on that podcast because I feel like there's more freedom to talk about non-tech stuff. But yes, we have plans next year, 2024, destination wedding. My parents have booked. They're going to come, which is always good. Was that in question? Were your parents like, maybe we won't? I don't know. I mean, you probably haven't had the experience of booking or planning a destination wedding. But in the back of your head, you do think like, you know, I do expect my parents will no question come. But like you do have like this thought like, you know, maybe they'll just be like, we're going to sit this one out.

Anyways, I think all my family has booked or at least put like deposits down and stuff. So it should be fun. The wedding will be on the first day because even though we planned this like 12 months in advance that weekend, the Friday, Saturday and Sunday were all completely booked with the venue that we wanted because they have multiple venues at the resort. And then we said, oh, well, either we take our third pick venue where we do it on a Thursday. And the advantage of doing it on the Thursday is people are going to get there on the Wednesday. And it's just like, I don't want to say out of the way because it makes it sound like we want to rip the bandaid off. But it's just like there's going to be a certain amount of you know anxiety and boxes

you're trying to check and i think if we had that in the middle of the week we're just going to be thinking about that like leading up to if i can use a tech analogy it's like when you're speaking at a conference exactly it is exactly like your talk first yeah you know and it's not that you're not excited about your talk. It's just, you know, you'll feel less stressed after the talk. Yes. Yes. That's exactly.

It's honestly like a perfect analogy. You are looking forward to giving your talk. It's a positive thing, but it's kind of hard to not have that in the back of your mind while you're attending other talks. And like half the time you're just like, well, maybe I won't go to that talk because I'll go prepare for my talk. And yeah, the sooner are you gonna are you gonna go do a big honeymoon afterwards uh i think that is the plan but we have not actually we spent so much time like doing the planning of the wedding that we kind of took a break from planning stuff although it is we are

we have a gantt chart we'll link in the show notes if you don't know what a gantt chart is folks um this is courtesy of my wonderful fiancee and we are we have met the deadlines of all our gantt chart stuff except we are past the deadline of uh starting to plan and book the uh i really i really want to see the scant chart uh yeah maybe we can arrange that at some point um anyways yes oh that's that's uh that's that's something else we with our catching up out of the way we want to address some of the feedback that we got about what will go down in history as probably our most controversial, at least comment-inducing and probably negative feedback-inducing episode that we've released. We're going to read some of our, just some highlights. We're not going to bore you.

And unfortunately, that means we're going to read some of the shorter feedback that we got. Because we got some longer feedback. We'll link in the description, in the show notes, to the discussion from the previous episode, 2.11, even though this is episode 2.12. We had one individual by the GitHub handle We're Lord, or We're Lord, say, AI-generated content is a hard pass for me. We also had from Goldfish Sound, this was painful to listen to, stripped of any human emotion and reflection. Please do not do this again.

And then we had a few other folks that gave some more lengthy responses saying that they would have preferred commentary from the both of us, not just some AI summary. We then on X slash Twitter got a comment from someone that we've had interactions from with before, saying that I need to do a better job managing your interruptions, saying, wow, this latest episode is so disappointing,

and I think it's the last straw for me. And anyways, to say the least, folks, primarily, I think, hated the fact that it was AI-generated. I personally, I will just speak for myself, I love the LLMs. And on one of my other podcasts, I've released, I think now two and shortly it will be three AI summaries of papers. And don't get me wrong. I understand that it's not as enjoyable as listening to humans. But especially when it comes to content that like people are not covering.

I would rather listen to an AI summary of like a 1971 paper on some topic that I'm interested in, then like have no one talk about it. But the people have spoken, there will never be an AI summary, at least for the next like two or 300 episodes. But then once it gets so good, and we can like feed our own voices in and then trick the audiences, maybe we'll do something like that. And I think you can't, you can't even do that today. So I listen to another podcast, Planet Money. It's like the one main podcast that I listen to when I'm like driving or, you know, bike riding or something. And they did an episode that they produced with AI. And they used – there's some company that has a technology where you can feed in some voice samples. I think it's not like that much content.

I think maybe a couple of minutes or an hour or so. And then it can build a model of your voice like very, very accurately. And so they did an AI episode with the voices of their hosts. And it was really good. It like was really really good and i honestly i thought that the the so you just notebook um llm which i hadn't previously heard of but

i thought it was really good and there were some um there were one or two places where it was summarizing or explaining some concept like through an analogy that I thought were like really good. I don't remember the specific, um, the specific ones off the top of my head, but they were like one or two places where, um, it explained something in a way where I was like, huh, you know, I might actually use that if I was explaining something about the C++ committee to someone. Um, and I I, look, I completely understand why people would dislike AI-generated content. And, I mean, I, you know,

have until recently been not a huge fan of seeing as much AI-generated content as we are seeing. And, like, I can understand why some people would have an aversion to it. But when you first sent me those clips, I just thought it was super cool. I was super impressed that it wasn't like you had gone off and spent weeks

building something bespoke to generate this. You'd just taken an off-the-shelf tool and generated something that I think was pretty good without too much effort. And I thought it was cool. And it was actually, I think, a pretty good summary and it was helpful for me. And I actually think I would probably use something like that

to do something like summarize papers in the future. Because I do use large language models more and more these days for like, you know, if I'm trying to solve some problem or something, I'll go to chat GPT. I find myself doing that more and more frequently. And I feel like I'm slow to adopt things like this. So the fact that I'm starting to do it more and more indicates to me that it really is like a 10x thing because I think for me to make a change to how I work, something has to be like 10x better. But one thing I've learned about myself the past few years

is that it's really difficult for me sometimes to get engaged with something that's purely textual. Like, if I'm gonna, if I'm gonna read a book, which I do enjoy doing, like, I have to be, like, really focused. Like, I have to, I can't, like, do it with, like, similar options. And if I'm gonna read, like, an email, reading an email at work can be really challenging for me, because, like, I have to sort of get into, like, reading mode email at work can be really challenging for me because like I have to sort of get into like reading mode. But like listening to something is substantially easier for me. And I'd never thought before about the possibility of using large language models to talk to me. And so now I'm thinking like, hey, you know, if there's some paper I wanted to read or if there's some, you know, email that's really long that I want to get a summary at, maybe I can just have some large language model talk to me about it.

And I never considered that possibility before. in different forms. And it's very interesting to me to now have the option to be able to generate audio, like to have somebody talk to me from almost any content. And I also, I think it'd be cool if I could, for me to, you know, sometimes people say like, oh, like this, you know,

this meeting could have been an email. i generally prefer to have a uh uh a meeting instead of an email so sometimes i'm like huh i wish this email was a meeting well with with ai maybe i can take an email and then turn it into a meeting because then i can have a meeting with some AI that talks to me about what's in the email. And I think that'd be cool. Yeah. There's a bunch of possibilities in the future that like, I mean, I think we're both biased because we work for a company that is definitely profiting from the success of products like this. So it's a definitely we're both biased, but like, there's a ton of stuff that I do that is, that is just like jumping through hoops. Like for instance, editing these podcasts, a lot of the work that I do, you can currently pay other folks to do, but it's like

more money than I would be willing to pay personally for what I think is a reasonable, a reasonable amount of work for me to do. And it also gives me more creative control of like how the sound bites at the beginning get cut and things like 10Xing voices or I can insert some clip from some TV show. I really enjoy doing that stuff. And I think it adds like entertainment and like there was one time I added a, a audio clip from Parks and Rec, where

was it Ben Wyatt is his name, playing the cones of Dunshire. And that's like one of my favorite things that I was able to do, like in the 200 plus episodes is like, I love that scene. And I'm sure there's only a handful of folks that listen that like actually even know that scene. But probably hearing it for that handful of folks gives them as much joy as know that scene but uh probably hearing it for that handful of folks gives them as much joy as it brought me anyways that kind of stuff i don't want to give up but a part of that is removing ums cleaning up when someone's talking over the other person and all of that is going to be the click of a button within the next couple of years and so my point being is i think that like this is work that like I don't want to have to do. I'm not getting paid to do it.

That is going to like disappear, which is why like any kind of like jumping through hoops work that can disappear, I'm super excited about. And anyway, so the point being is – part's going to disappear. Do you mean that you think that in a few years in your audio editor, there'll be a button that you'll press that'll be like remove ums and interruptions? Or do you mean that you think in a few years there'll be some AI assistant where you'll be able to like, or I guess what I'm asking is, do you believe that the task will be completely automated away where it'll be a click with a button and you will just have to review it or do you think it'll be something more where it'll be like an assistant an aid where you'll just um uh you'll just like be able to like interact with it some way you know through

like like the generative um uh ph fill and in photoshop where like it gives you a prompt and you like type out what you want the photoshop to fill into that space do you think it'll be more like that or more like something where it's just like a button that you press both forms will come but the one that i'm picturing is the button like currently there's a tool called noise reduction and audacity where you give it a sample clip of like some background noise if you've got like a fan or an ac or something that's like a little hum you give it or maybe there's some people a couple meeting rooms over that your mic picked up a little bit of you give it a sample of that it creates a noise profile and then you apply that to the whole track and it gets rid of it like

there's going to be some like ai enhanced versions of this that like you give it some audio clips of certain words or something or like and it'll it'll highlight you know you you give it some audio clips of certain words or something or like, and it'll, it'll highlight, you know, you, you give it one or two ums of an individual and then it'll go and find every single, um, highlight it. And then you can do like a quick, you know, I'm sure there's only 30 seconds of ums that one of us says in a 30 minute conversation. So if it goes and selects everything, then I can just go hit play to make sure it's um, um, um, um, um, um. Although with an episode like this, it would absolutely just fall over

because this is intentional ums that I want kept. Anyways, and then you can just delete those. And it'll do a lot of the stuff, like the passes that I do to normalize volume. It'll probably be some just like, you know, you drop your audio clip and it does a bunch of automatic filtering. Then I just need to go and do the creative parts that I want to do. And that will be 10 minutes compared to 60 minutes for a 30-minute episode. All right.

Is that how long it takes you? It just takes you 60 minutes to to do the if i have decent audio tracks yes but also too like i know that for other folks it takes like hours like i know that cpp cast and other folks they put like a lot more energy into editing their stuff but like my philosophy is uh i'd rather rather have more content and lower quality. And I still think I do a pretty decent job editing stuff. But there are times where some ums get left in and it's a little bit messier. But yeah, if I have clean audio tracks.

But every once in a while, we'll have a guest on and their audio will pick up echo from us. And then it takes like three times as long because I need to go silence their track when any when any when at any time anyone else is talking. And that that's like a nightmare. I would I would never I would never be able to do it just now. It would take me it would take me so much longer just because I'm so much of a perfectionist. You know, it's interesting. Somewhat on topic. I don't think I've ever told you about the first time I used Audacity, which was when, like sometime in the 90s,

my grandmother recorded some conversations with her mother, my great-grandmother, who was the family member who was born in Russia and immigrated to the U.S. through Poland in the 1910s, 1920s, when it became not super safe to be Jewish in Russia. And so my grandmother made these recordings towards the end of my great grandmother's life of her talking about her life in Russia. But these were made on an old cassette tape. And I got interested in family history in the 2010s era. And I got told about these tapes and somebody had remarked like, Hey, we should really digitize these because the tapes are like 20 years old at this point.

And the quality may have degraded. And so I went and did that. And there was one, the record, the initial recording was not particularly good because it was like recording done at a party. And it was like a done at a party. Um, and it was like a mess. Um, but then too, there was also some like degradation to the cassette tape itself.

So there was like this, like every like three or four seconds, there was this like background, like blip noise that was just from some damage caused to the tape. Um, and it was like maybe, maybe like an hour, an hour and a half of content. But I got this device that would like digitize from a cassette tape. And then like I loaded up into Audacity. And I went through and I must have spent like a week like carefully editing out every like three or four seconds this little blip of the bad audio. And I didn't know anything about audio editing. So I'm sure that somebody who was qualified could have done this a lot faster. But I recently had dinner with one of my distant

cousins and I told her about this tape and her great-grandfather was the brother of my great-grandmother and so she was very interested in it. And so I dug up my, you know, digitized version of the recording, and I was going to just send it with her, but then I was like, hey, you know, maybe I can get, like, an audio transcription of it. And I uploaded it to this thing, Otter.ai, and it was super cool because not only did it give me a pretty good transcription of the audio, but it also gave me like a summary.

And it like it caught a few things that I hadn't because like I'd listened to the recording, but like, you know, I hadn't like taken notes or anything. And so like it caught a few things in its AI, like its summary of the whole conversation that I hadn't realized that were like important details. And it was just like, it was a super fascinating tool. And Ramona, when we were in Poland, she got some, she made some recordings of some of her family members talking about their family history with the idea being that we'll also feed it into this, this AI tool. Yeah, it was, it was, it was really cool. Like I couldn't, I wouldn't have been able to get that sort of summary and transcript, you know, 10, 15 years ago. Yeah, it's, and yeah, we should mention Whisper too, which is anyone that's a programmer can go and look it up in Python. You can use it as like a command line tool and just, you feed it any MP4, MP3, and it just does like auto transcribing. Once again, like it's not, it's

not perfect, but that's, that's where like we're headed is I've actually thought about trying to add bad transcripts. Cause I don't have the patience to go through and like touch up the mistakes, but like, that's another thing. Like we don't have transcripts on ADSP. We looked into it. It's exorbitantly expensive we're like from what i think it should cost uh because we don't we don't collect any advertising money from this stuff and we could go spend the money but like it's it's not free at some point like perfect uh transcriptions i mean technically we're recording this in teams right now we could hit the record button and transcribe button and then we could copy and paste that and that does

like a pretty good job i actually do think we it automatically does the recording and transcribing so it we do have the transcriptions from teams yeah but the problem is is that it doesn't i have to go and fix it because we're going to dice this up right whereas what i want to be able to do is like feed it the finished mp3 and that's the other thing is some people have requested a video and for us to upload it to youtube of us chatting but once again like i edit the audio and like having to edit the video would like increase the amount of work by a factor of like four or five however yeah imagine you can just feed it a a uh edited audio clip with the original video and just say like cut my video so that it matches what i did to the and like in the future probably it's just one whole ai thing i drop something in it goes and cuts everything out so like there's going to be steps along the way but the point being is there's

a bunch of stuff that we don't currently do that we could do for free with ai and a lot of the stuff that i do do currently like i won't have to do. And like, so I am just extremely excited because I have like 17 different things on the go at all times. And like, I'm going to free myself up like hours of time a week. Anyways, that doesn't mean that we need to upload AI podcasts though.

The point being, we're not going to do it again. You'll hear from us if we want to talk about something. And anyways, that's that's but I'm not I'm not sorry that we tried it. I'm not sorry that we tried it. I'm not sorry we tried it either. I'm not I'm not sure also to we said we were gonna, or at least I was going to provide a little bit of commentary on what our goals were. I mean, I think we wanted to highlight not just that these AI tools were cool. But you know, I think some of the content that was covered by Izzy, I meant to say this earlier too, we added or I added a audio clip at the beginning of the episode

that a lot of you might not have heard, so it's worth repeating it here, that Izzy, the author of the blog, after having listened to the episode, reached out and said that it didn't capture the essence of the blog correctly and that if you do want to refer to anything from that episode you should really go find the corresponding content from the blog itself so we'll link the blog again don't refer to our source as like the official source because it's not it's it's just a transformed version and and we're sorry is a we you know we should have yeah should have checked with you I. I actually did try to reach out to Izzy beforehand.

Izzy's not on Twitter. I went to Mastodon, found Izzy on Mastodon, but they don't allow, or I'm not sure how Mastodon really works, but I'm on a different server than Izzy is, and I couldn't DM, or there weren't DMs available. Anyways, I gave up after like 10 or 15 minutes, and I should have done that.

Anyways, everything's all good there. But the point being is refer back to the blog, not to the podcast. Anyways, but we did want to shine a light on some of that stuff. And I think it was frustrating for me when I was reading the blog because we both have intentionally decided not to talk on the podcast explicitly about a lot of this stuff because this is now the nth time.

I don't know how many times this stuff has come up. And I think this was kind of the straw that broke the camel's back for me. Not to the point where I wanted to have a whole episode where we talked about this stuff, but enough that I wanted to shed some light on Izzy's blog post. And it was the straw that broke the camel's back for me. I mean, I already, I don't know if I've ever said this, you know, on this podcast, but I made a decision back in 2021 that I was never going to attend a committee meeting again. And there's a couple of conferences that I also chose that I'm not going to attend again, unless, you know, changes were made. And the reason this was the straw that breaks the camel back is that there's a number of individuals who I will not name that I try to

avoid for reasons. And that list of people, I can no longer count on one hand. And it is, it's not even something that like really I am the most affected by, but I am affected by it. I run into these people at conferences and I would prefer not to interact with these folks. And you have no idea who's going to show up at a conference when you go there. And this year, I won't name, I went to many conferences this year, but this year I ran into people on this list that I would prefer to avoid. And it is unfortunate when this stuff comes up and then nothing happens about it.

I go to a conference and in the back of my head, I have to think like, am I going to run into one of these people that I do not want to have interactions with? And this is my personal opinion here. It is a privilege, in my opinion, to be able to be a part of these communities. It is not something that like everyone is owed as a right, you know, just because you exist in society. I think that like if I'm organizing a conference, which I do just because you exist in society. I think that like, if I'm organizing a conference, which I do not, but if I did, I would want to be creating a safe space that folks feel safe and welcome in. And if there are people that are showing up at these

conferences, like that violate that safe space, I'd prefer them not to be there. And I don't think that it is like, like I said, this is just my personal opinion that like, I don't think it is you're doing something wrong by like having a not welcome at my conference because you make other people uncomfortable due to things that you've done in the past. Anyways, and so all that to be said is that like, there's multiple times where we have had private conversations about like, do we want to talk about this on the podcast, we've always come to the decision that a similar to CPP cast, we'd rather talk about technical stuff or fun stuff or catching up than like these topics that are not fun topics to talk about and this time I you know I talked to you about this I talked to other folks about this before we posted it

and I decided that like it wasn't worth staying silent this time at least as far as posting an AI summary of it counts as like not staying silent because like we technically didn't say anything. And even to this extent right now, like a part of me feels guilty that like, you know, I'm being very vague about everything. I'm not actually calling people out. I'm not saying specifics. I'm just saying that like personally,

I think this stuff sucks. I think that like every single time this comes up, nothing ends up changing. And there was like a specific incident that happened at the C++ national body, which I'm not going to get into the details about. But I was not the, I didn't publicly step down from the national body. But there was another individual who I'm not sure if they want to be named. We haven't talked to them. So I'm not going to name them.

But at the time, they quite publicly talked about how they were really disappointed about what was happened. I was just as disappointed, if not more. And I lost a lot of respect for people that I have respect for. And attending these committee meetings, contributing to the evolution of C++ was just something I was no longer interested in, which is just, it's a personal decision, right? Everyone can make their own decisions on this kind of stuff. Anyways, I feel like I've kind of said everything I've wanted to say.

The point being is that in the future, no more AI podcast summaries. If we do want to talk about this stuff, you will hear directly from us. And, you know, if some of this stuff happens in the future, and you're wondering, Oh, how come Bryce and Connor, do not think for like a second that we don't know that this stuff happens. And then we don't privately discuss whether or not we think about whether we should discuss because we always do. Yeah. Con because of disagreements with other people in the CPP Con leadership about certain decisions were made regarding certain people attending the conference. And I think maybe the better way to put it is I resigned as the program chair from CPP Con because there was a disagreement with me and other leaders about values.

And even just my resignation where I didn't air a lot of dirty laundry, I just sort of resigned. Even just that, there was a ton of blowback and nothing really came out of it. And I am always hesitant about, like, when should I speak up? When, you know, like, when is it appropriate and moral to do so? Because, like, yes, on the one hand, I do feel an obligation to speak up when there's something like a safety concern. But like on the other hand, you know, if I'm, if I learn about something in a context that's privileged, you know, I don't necessarily want to betray that trust. And also, it's not always my story to tell. And in a a lot of cases there have been things where you know

i don't want to be speaking up because uh it's not my story and you know i i instead would encourage the person that if they want to they can if they want to because some of the times the people involved in these incidents they they just want it to be forgotten about because there's so much blowback because if people do speak up there's so much blowback and uh and and so in a lot of cases what happens when there's a controversy like this is that people the people who are wrong try to get something to be done behind the scenes because they don't want to have the controversy, the blowback that they get from sharing it publicly, but they want some change. But then, you know, nothing happens. And then eventually they get frustrated and then they decide it's time to go and share something publicly. But not everybody feels comfortable doing that. And the other thing I want to add to

is that a lot of the comments and like negative feedback, not with respect to the AI, but with respect to the content that talks about the unethical behavior and whatnot is this kind of, well, we're only hearing one side. I, how should I say this?

I do not want to go into details of how I know what I know, but of the folks that are on my list of people that I try to avoid, I'm not sure, Bryce, we're not going to get into the specifics of this, but I'm not sure if you remember at one point before the incident with this individual that I'm thinking of, which you might, I'm being so vague, you might not even know what I'm referring to. This individual, they got, we'll say, promoted. That's not actually what happened, but we'll use the word promoted to something. And I was upset. I immediately messaged you. They were a nice enough person, but they had done some things that I was,

I don't know what the words are, like not okay with, but like wasn't comfortable with. And not things that were necessarily code of conduct violations, but things that like made me kind of like raise my eyebrows. Like, really, is this the kind of thing that you should be doing at a conference? And anyways, then this person got quote unquote promoted. I immediately messaged you and said, should I talk to the conference organizer about my discomfort around this? And then we both chatted about it and we're conflicted and like, well, we're not really in like I'm not on the board of this conference. I'm not super involved. Like it's not my what is my place to say that?

Like, I think that this is kind of an odd pick. And and also to like, what would I say? Like a couple things were done that made me feel weird or like uncomfortable. Like there was no actual code of conduct violations. I never spoke up at the time that the things happened because I didn't think they were worthy. It was just stuff that kind of made my, you know, the hairs on the back of my neck stand up. Which doesn't seem like enough to go and say like, oh, I have a hunch. Fast forward like, I don't know what it was,

a year or two years, and now they're on my list of folks. And I remember talking to you after the fact and I felt like, I felt bad. I was like, we talked about this. It's like, I had my suspicions and then something ended up happening and I never said anything before.

And it's like, well, we talked about it. It's like, what are you going to say? It's like when there's something that's not a code of conduct violation, but that is just kind of like, it doesn't feel great. It doesn't, like, we talked about it. It's like, what are you going to say? It's like when there's something that's not a code of conduct violation, but that is just kind of like it doesn't feel great. It doesn't like it makes you uncomfortable. I think we even chatted about like, should there be something like less than a code of conduct violation, but more of like a let your anonymously like your friendly conference goer know that you did something that kind of made someone uncomfortable. And it's like, I would always prefer to know if I've done something that made someone uncomfortable, especially if it's like a zero cost thing to me, maybe I use some word or I use some phrase. It was similar to that, like a YouTube comment that

one time where I had used the words final solution as the name of a function in a YouTube video. And then I got a couple of different comments being like, ah, great video, but like, don't use final solution. And I was like, what? Like, I guess it's vague. I could have been more descriptive. And I had no idea. I did not pay enough attention in history class that like it's a reference to the Holocaust and link in the show notes. The point being is that like that kind of stuff, I would prefer to be brought to my attention because it's an easy enough thing for me to avoid.

And if it makes other people feel more comfortable, perfect. It's not like it's not like, you know, censorship or like, ah, I can't say the thing. I don't care. Like, it's not important to me to be able to say that. So the point being is that like, I don't want like people to be scared about, you know, what they're saying or the words they're using at a conference. But like, I think a lot of folks, if they were told that they said something or did something that just kind of like a little bit made someone uncomfortable and that there was a healthy way that they didn't feel like they were being accused everyone would want to know i think everyone would want to know that information without thinking that like we're headed into like big brother territory or whatever my point being of this long-winded ramble is that is that like i i saw this happening like in slow

motion and was not surprised at all that this happened so it's like it's like some people are like oh well we haven't heard the other side of the story. I don't need to hear the other side of the story. Like I was already uncomfortable with some of the stuff this this individual had done in the past before they made my list of like short list of people that I want to avoid. Like I do me personally, I always want to hear if I've said something that might offend or upset someone. And, and to me, it's very low cost to change our language. And that's, you know, I used to, a few years ago, I worked on the Insights Inclusive Terminology standard,

which is still work that I'm pretty proud of. Because I feel like, like, I, why would I want to use language that has ambiguity? Like, like setting aside an immoral argument. If I've said something in a way to you that you've taken some second meaning from it that I didn't intend, like, yes, I may have unintentionally hurt you. But also, like, I've just been whenever we're talking about naming stuff, I'm a big believer in avoiding what I call, like, semantic overloading. Like, you know, using, like, in concurrency, like, the word task. Like, you should almost never call something task because task just, like, it means so many different things to so many different people. Like there was a proposal for a thing, a sender's facility that was going to be

called stood task. And I was like, no, we can't call that stood task. Like, like that's just, that's going to have so many different meanings to so many different people. Like just, just pick a different word. You know, like you just, you, why would you want to pick a term that's going to have so much ambiguity and when you could instead, like, there's so many words out there. Like, search for a different word that won't have this overload. And that's sort of how I feel about, like, you know, inclusive terminology in the language that we pick. It's like, why not just be as clear and unambiguous as possible? And I am sympathetic to the people who get that feeling that they are, that they're being watched or that, like, they have to, that they're being policed, that what they're saying is being policed.

Like, I get that. I understand that. And I understand that some people feel a fear that they're going to be attacked for what they've said. And I think it's important that when somebody does say something that might be offensive to you, that we assume good faith unless there's evidence otherwise. Assume that the person didn't mean to do that. And if you want to tell them about it, tell them about it in a nice way, not a way that's going to feel like they're being attacked. Because then that will make people less afraid feel less policed um you know it's more like constructive criticism it's like if there's a if there's a type the same in the same way if there's a typo in my slides of course i want to hear about that you know yeah yeah it's

like like i said it's the reminds me of the youtube comments because they weren't harsh about it they said great video like really enjoyed this. However, like only piece of feedback is would prefer if you didn't use, you know, that name of a function and they didn't even like explain. And then that was like they didn't go. They weren't even they were too nice. They they could have said that and then drop the link so that I because yeah, like I said, it's as long as you do it in a nice way. I think a lot of this stuff are like super low cost to like it's not the end of the world.

And every once in a while still I will find myself, especially when I'm iterating, even if I'm not naming a function, the final solution. I might refer to some solution that I'm working up to as the final solution. And like literally like in the last couple weeks, I've been recording a bunch of Advent of Code YouTube videos. And a couple times I've said that, and you know, I just pause while I'm recording. And I think one time I referred to it as solution two, or the other time I referred to it as the ultimate solution. Like it's the final solution to me. Sure. And do I mean anything bad by it? Absolutely not. But like, should I be getting upset about the fact that I need to avoid it? Or it takes me a couple extra seconds to like change my verbiage? I don't think so.

Like there's a bunch of other things that I can refer to it and have the basically identical semantics. And nothing, there's zero cost to me in terms of like the educational or the entertainment value. Anyways, we're kind of off on a tangent here. But the point is, no more AI. The people have spoken and we have listened. But also, too, I don't want in the midst of the AI it to get lost that there's a problem and it's not being fixed.

We're not even really advocating for any solutions. I think in general, I've kind of my involvement in C++ has just, my, my decision has just been to kind of step back from that. Primarily, like I, I write in Python for my research work these days. And it's funny because I'm doing more and more Python stuff too. And even over the last couple of years, like the main, I'll be honest, the main reason I still go to C++ conferences is because the majority, 99.9% of the C++ community in terms of speakers and attendees and the listeners of this podcast, I love interacting with that community.

And it is sad that I have to miss out on that and not interact with that community. So that's why I go to conferences like CPP North, where I know that they're really prioritizing welcomeness and safety. And there's a couple other conferences that are really trying to do that. And that is the beautiful thing about the C++ community. There are a ton of conferences, way more than any one person, although I'm sure some people have tried, if you're a dev advocate, to hit all in one year. And so you can choose the one or two that you want to go to.

And then I've just been prioritizing other conference circuits, to be honest, which is sad. Like there are some folks that I really enjoy seeing and we'll see less of because of this. But yeah, until that change or if that changes, you know, that's kind of my personal decision. It's sad to say it is what it is. You know, I hope that the people in charge

do the right thing. I have zero trust that that's going to happen though because this has not been going, this is not like the first time. This has been happening for years now. And I've seen that the right thing typically doesn't happen, at least at a certain level.

There are certain conferences that have really, really tried to prioritize doing the right thing. And those are the conferences that I'm going to spend my time, if I'm going to one or two C++ conferences a year, those are the ones that I'm going to prioritize. I'm not sure if you have any final thoughts. This is probably going to be a super long episode because it's going to be hard to chop this up into two things. So we'll probably just be one and done.

Any final thoughts? No, I've been desperate to talk to you about Python and algorithm stuff. All right. So end of episode 212. That was 58 minutes of recording, folks. I'm probably going to trim it down to hopefully 40 or 45. Be sure to check these show notes either in your podcast app or at ADSPthepodcast.com for links to anything we mentioned in today's episode, as well as a link to a GitHub discussion where you can leave thoughts, comments, and questions. Thanks for listening. We hope you enjoyed and have a great day. Low quality, high quantity. That is the tagline of our podcast. It's not the tagline. Our tagline is chaos with sprinkles of information.