Algorithms + Data Structures = Programs - Episode 269: 2025 Double Retro

Episode Date: January 16, 2026

In this episode, Conor and Bryce conduct the annual 2025 retro!Link to Episode 269 on WebsiteDiscuss this episode, leave a comment, or ask a question (on GitHub)SocialsADSP: The Podcast: TwitterConor Hoekstra: LinkTree / BioBryce Adelstein Lelbach: TwitterShow NotesDate Recorded: 2026-01-13Date Released: 2026-01-16VOTE FOR YOUR FAVORITE ADSP EPISODES OF 2025ADSP Episode 259: 🇳🇴 NDC TechTown 🇳🇴 Vittorio Romeo & JF BastienNDC TechTown 2025 PlaylistPrefix Sums and Their Applications by Guy BlellochParallel Prefix Sum (Scan) with CUDA by Mark HarrisADSP Episode 263: 🇳🇱 The Sean Parent Origin StoryADSP Episode 232: Algorithms! Live from New York!Intro 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

That brings us to our 2025 retro of what are we talking about? Favorite episodes, favorite moments. You said you came across your favorite. So I want to know, yeah, I want to know what episode from 2025 did best. I'm going to put it close to the camera. Read for the listener what we have here. It says, it says Ayota, hazy IPA. That can't be how they want you to pronounce it.

Non-alcoholic. Oh, you're going to have to send me one. You're going to have to send me one. Welcome to ADSP the podcast, episode 269, recorded on January 13th, 20206. My name is Connor, and today with my co-host, Bryce, we conduct our 2025 annual retro. Austin, Texas. Austin, Texas. What does Austin, Texas mean?

Your shirt. Your shirt. Oh, yeah, yeah, yeah, my bad. I'm wearing one of my many race t-shirts, and I have raced, I think twice. the no longer but formerly 3M Austin Half Marathon. I think now it's called the International Half Marathon sponsored by some other company. It's a great marathon, mostly, not Marathon, Half Marathon. Downhill mostly.

I love Austin. Downhill is my type of running. Ramona and I have started running recently, and it's miserable. It's just miserable. And like she, she doesn't exercise really at all right now because her arm's injured, so she hasn't been able to play tennis. And I exercise a lot, but we go out and she's just like way faster than me as way more energy and endurance. It's unfair. It takes time to get into.

It takes time to get into. But she hasn't, but see, I don't buy that because she hasn't been running for years. She started again, and then she's just instantly like 10 times more athletic than me. I mean, if you used to do it, I imagine it's kind of like. like working out, if folks ever worked out in, like, university or at any point in their life, and then they stop for like a decade, you still have like the muscle memory of it. So it's way easier to get back into it if you've done it at one point in your life. I think she just has a higher tolerance for suffering than me.

But now we take the dog when we're running to it. So we go for, our route is like 4.5 kilometers. She would say it's four kilometers. but her garment watch is just not recording correctly. But the three of us go, and the dog doesn't much likely. The dog just sort of trots. The dog doesn't really run. The dog just sort of trots.

And it's cold, but we've been going out most days. It's good. How is the weather in? Not that we need to bore the, maybe I'll pen X this. How is the weather in New York? Are you running metals behind you? Are they sorted by color? because I'm noticing that all the ones in the right are like,

I see like dark red, then like slightly lighter red, and then some green and then blue, they are organized by color. But the ones on the right side are not like organized. You can imagine why. You can imagine why, right? You start off with it, rainbow color sorted. And then when you gain new ones,

it's not like you're going to do a one rotate and shift everything. You just kind of. I mean, I would. Not every time. You got to wait until you get. like a, what do you call it, a critical mass, so it's worth it. But they, wait, wait, wait, hang on. But all of the, these are all primarily red, white, or blue.

Do they not make other colors? Well, I mean, yeah, that is true. That is true. I got to spin around. There is a ton of blue. That is definitely the most common. I don't actually, I mean, a lot of the American ones, you can kind of tell by this shirt, it's got red, white, and blue.

Marco. Yeah. And I guess the Toronto Waterfront ones are red. There's a lot of holiday. Like there's the San Diego holiday half that I've run twice. And there's the Mesa Marathon that's red. I don't know. I guess red is kind of like a celebratory, you know, jubilant.

You cross the line. I don't really know. There's a... It's a... I just got back from Southeast Asia and it's like everything's red there. Oh, yeah, yeah. Well, red is like literally the... Is it lucky? The color of opulence, wealth, and luck, yes.

Yeah. But wait, here, look what I came across, Bryce. This is, I'm going to put it close to the camera. Read for the listener of what we have here. It says, it says, Iota, hazy IPA. That can't be how they want you to pronounce it. Non-alcoholic. Oh, you're going to have to send me one.

You're going to have to send me one. It is, I couldn't believe it, folks. I've started drinking more non-alcoholic. beverages because, you know, they're better for your health. And I was in a metro, which is a Canadian grocery store. And to my surprise, right there, in between athletic brewing and a couple other brands, is the Phillips Brewing and Maltinko Iyota. Hasey IPA.

I've been waiting. I have not had one of these yet, folks. It's been sitting in my little mini fridge behind me. Here we go. Just like old times, we used to crack the bubblies. now we're cracking non-alcoholic hazy IPAs. And here's the live review. That did not register.

That did not register on my ear pods. Maybe it registers on yours. It definitely registered. I can see the little cracks in the audacity. So I hope this is amazing and it's going to be my favorite thing ever because I really would love to love the IOTA, Hasey IPA. It's everybody, honestly, I'm not even going to explain it. If you don't know what Iota is, go back in the backlog.

You got 268 episodes to listen to. Jeez, we're coming up on 300. It's pretty good. Yeah, we're getting close, man. We're getting close. We're getting there. And, oh, yeah, we got so much to talk about. So I think we're, it's good, folks.

We haven't talked in, like, a while because the holiday is, like, have we recorded since your wedding? No. I mean, the last time we recorded was October 10th or 11th or something like that. That was like a long time ago. Yes. We're going to call this episode the double retro because we're going to do like a retro. of our whatever favorite episodes or highlights, ADSP, whatever.

But also, too, I have had a list of growing things, corrections, references to comments that we've got in the GitHub discussion. And so it's going to be one retro is addressing these things that are outstanding. And the other one is going to be like our top stuff. But yes. And also, Happy New Year. So it's been what? This is almost three months since we have recorded.

It's only been a couple days since we chatted because we chat outside. Believe it or not, we chat outside of these recordings. What should we start with? Should we start with addressing the outstanding feedback items? What do you want? I would say, yeah. All right, let's see if we can burn through these because there's like five or six of them. The reason that we have to do that is because I had a favorite episode of 2025,

but I cannot remember what it is. So if we trust the feedback first, I'll have enough time to remember. All right. So here is my list of stuff. The first one. And I think this is in chronological order. when we were talking to what episode was this, Vittorio and J.F. in episode 259, I did not at the time go to Vittorio's talk, and I said I was going to J.F.'s talk. I did not end up going to J.F.'s talk either. I did, I think,

watch most of it. Oh, I'm going to get text messages from J.F. about this. J.F. apparently does not know how to communicate with Conor, so anytime he needs to communicate with Connor, he sends me a text message, and then I have to text Connor. Well, the last thing that you relayed on his behalf was to add the YouTube link to the show notes of the podcast. But I had already done it. I had seen the NDC talks link in the show notes to the playlist. I don't think all of them are online because you gave a talk at NDC, correct, on top of the two trainings that we did? Yeah, I did.

So that is not online yet. JF's talk is, and I think most or half of the NDC Techtown talks are online. Anyways, Vittorio gave me a hard time. say why are you going to J.F.'s talk? And I didn't end up going to either. So I just, I'm not sure if Vittorio's listening. Don't feel bad. I didn't make it to either. I had to run during both of them. And Vittorio's talk isn't online either. I'm not sure if that's because he didn't want it recorded, which doesn't sound like Vittorio, it might just be in the bucket of tops that talks that are not uploaded yet. Point number two, I mentioned in one of the episodes where

we were in Breda, Netherlands. I believe it was either episode 260 or 261 that the author of prefix sums and their applications was Mark Harris. That is incorrect. Mark Harris has published papers

on prefix scans, but it was not that one. That is authored by Guy Bleleck. So I think I discovered that when I was trying to add it to the show notes and discovered, uh-oh, I misacredited a paper. So I will link both Guy Blalick's paper and one of Mark Harris's papers, because I think that's what I did

back then as well. And there's three left. So we're going to skip this one and come back to it because it will potentially require the most discussion, or we might also just kick it to a whole other episode. Maybe it'll be a bonus episode. Number four, so we'll come back to number three, is that I mentioned in episode 255 that Pagacha was from Denmark. And I was saying I learned all this stuff about Denmark. I think one of them was that Carlsberg was the official beer of Denmark. Pagacha is not from Denmark, folks. He is famously from Poland.

And what's the individual's name? Oh, you're going to get in trouble with Ramona. Yeah, yeah, well, I got in trouble with, we don't have this individual's name, but their username on GitHub is, I'm going to butcher this pronunciation, Vike, S.L. Gartner. And they commented pagachas from Poland. Oh, boy. Hashtaghtag confidently in Khr.

correct. What did you learn on your last trip to Slovenia? Anyways, we do apologize not only to the country of Slovenia, but also to the country of Denmark, although misattributing one of the top cyclists in the world, if not the top cyclist in the world, I guess it's not really a bad thing. Number five, in the conversation with Sean Parent, which was spread across episodes 261 to 263, I think, or maybe it was just 262. 263. I mentioned the word abscond and then asked Sean if I use that word correctly and he said I did

and I looked it up afterward because I was wanting to add the definition of abscond to the show notes. I did not use the word correctly. The definition of abscond is to leave hurriedly and secretly typically to avoid detection of or arrest for an unlawful action such as theft. And the way that I used it was a couple of attendees of the conference were absconding Sean so we weren't able to talk to him right away. And that is definitely... Yeah, they stole Sean. What? No, no, no, no. That's fine. Like, they took Sean and they stole him away from him. No, but that's, that's not. So to abscond is means to like to exit without detection. They did not exit. So anyways, but it's to exit. It's, it's to exit without detection. Like, you could say to abscond with an item. Like, he,

absconded with the gold or something. I guess maybe. I mean, maybe there's some, we've got like several thousand listeners. If someone's a wordsmith out there, we didn't get any comments, but it seemed incorrect when I got the definition. And so I asked a couple LLMs if there was a word that I was trying to reach for but couldn't. But it didn't really give me the top one they suggested was a cost, which is definitely incorrect, because that's when you kind of like aggressively, you know, go up to someone.

talk to them or whatever, which isn't incorrect, but it's definitely not accosted. And then that couple other options were detained, waylaid. Anyways, maybe abscond was correct. Go to the get up discussion. If you are a wordsmith, let us know. And if you know the word that I was reaching for, anyways, enough on abscond. And then number three, so this is last but not least of the five things I had written down. We got multiple comments.

And I think I had mentioned this to you at some point while we were traveling or or whatever. We talked about the Tesla situation and the fact that I don't have a Tesla when we were road tripping from Orhus to Copenhagen in the Polestar. And at one point in that episode, I said, separate the art from the artist. I regret having gotten a Prius. I wish I got a Tesla. And folks, there's a little sidebar.

I regret it even more. Tesla, a Prius in the winter? My goodness. Don't get me started. Don't get me started, folks. And anyways, we had multiple individuals comment. And we'll read in the entirety from, maybe we won't, we'll just, we'll name them Robert. You can go find their username if you really want to because it's publicly posted next to a GitHub account.

This is the response. And there was a second one as well that asked a couple different things, but we'll just read this one. Quote, I'm sorry, I tried. I can't separate the art from the artist when some of his art is about supporting the rise of the far right where I live. I don't have that privilege. I thought I'd check again, seeing as this post-mentioned EVs, I wouldn't even care about any of you owning a Tesla,

but I can't support proposing and or rationalizing, giving Elon even more money to attack the freedom, liberties, and well-being of other people. I've gotten lots of great information from your podcast throughout the years, and I'm still very thankful for that, but for me, it's the end of the road here. I sadly won't be checking in again, but still wish you guys all the best, end quote.

modulo a couple of my speakos. And I mean, we don't want to make this whole episode about Tesla, but do you have an instant response, Bryce? How do we feel about having lost a listener? I'm sorry to see somebody go. And maybe we should, maybe this is a lane we should stay out. The thing is, because I don't want, I don't want people to feel bad when listening to the podcast. and I don't want to bring up stuff that's got lots of baggage for folks. So I'm sorry. Perhaps we should not have gone there. Yeah, I mean, I think in general, we have, I mean, at times we've brought up political stuff

slash adjacent to a political stuff, and it never goes well. So I think over time, we have just tried to either stay on technical topics or non-technical fun topics or just nonsense that people care or do not care about. I mean, one of the comments on one of our Denmark episodes was, you know, I'm 15 minutes in, and they haven't mentioned anything about algorithms. I'm out. So I was like, okay. But people should know by now. Chaos with sprinkles of information.

Sometimes more chaos, sometimes more information. And I mean, but we also, we had some, that road trip produced what I have now found my favorite episode of the year. That, that road trip produced some really gold content, I think. I mean, I don't know, do you have the numbers of what episodes from 2025 did best? I can bring them up. I might need to eyeball it a bit or I could quickly feed it to an LLM. But I mean, whenever I see those comments of like, oh, they're not talking about anything, I completely disagree because a lot of times we're talking about nothing,

but then we instantly drop down into, you know, some discussion about algorithms on the CPU or GPU. So anyways, you know, if you don't like the fluff, that's fine. maybe this podcasting for you. But, you know, it's part of our brand at this point that, you know, sometimes we're talking about random talk. It's that being said, the Tesla slash Priya stuff, it doesn't, it's not intend. I guess when we're talking about can I buy it, it is a political thing. But, you know, it is, it's also like a technology product, if you will.

And I guess that's why we veered into it. I do feel bad about platforming, you know, talking about it and then potentially encouraging folks. That being said, you know, I just, you know, I keep on hearing about these Nvidia, Mercedes, Alpameo, whatever, full self-driving. Can I buy a full self-driving Mercedes? I keep hearing a modem in the keynotes that Jensen gives. But can we buy a self-driving Mercedes?

Anyways, are there other options, you know? I'm happy to support a different company. But yeah, maybe going forward, we won't touch that stuff. I don't know. Maybe it's the people that are upset that we've. comments, maybe other people enjoy the discussion. Do people not care? We need your feedback, folks, on the GitHub discussion. Let us know if you want to veer away from this stuff, because, you know, I do feel bad talking about this content. If it's like, you know, the point is to have

fun and talk about algorithms. And if we're discouraging people from listening because they don't want to hear about this stuff and they're suffering through. And that reminds me, there was another running podcast that I stopped listening to because I loved the technical stuff. But whenever they would talk about personal stuff, it drove me up the mall. And I knew like a few people in my run club that it was the same thing. The technical content was amazing, but they could care less for all the other stuff they talked about. I wonder what percentage of our listeners are like that, where whenever we start talking about coaches, people get irritated. Anyways, if you have feedback. There is, there is a suggestion. I'm talking to the LM. It does suggest that there is a pilot program

for Mercedes self-drive in the U.S., but only in California and Nevada, and only in certain conditions, like when there's, like, if you're in a traffic jam and it's less than, you're going less than 40 miles an hour. But... Yeah, they ain't going to cut it for me. The other two listed is, is, uh,

GM's self-driving product and Ford's self-driving product. But I suspect that those are also not going to do it. But let's ask about Canada. Let's ask about Canada. I mean, I don't think it's going to be different. We did some research into alternatives. I think we mentioned B.D. And we even looked into importing it from Mexico because it's like one of the top electric vehicles

and self-driving options. But it is illegal to import a car from a different country that's not on the sanction list. It would have still been cheaper. Also, even if you could import the car, you probably couldn't use self-driving feature unless it's approved by the local authorities. Oh, yeah, yeah, potentially. Anyways, folks, we don't mean to talk about this more. Will I switch cars in the future? Maybe nobody will know in the future because we're putting our viewership on the line.

And we don't want to do that. Anyways, those are, that's the first part of our retcho was going back. Thanks to everybody left comments and we apologize for the mistakes that were made. That brings us to our 2025 retro of what are we talking about? Favorite episodes? Favorite moments? You said you came across your favorite. So I want to know, yeah, I want to know what episode from 2025 did best.

I mean, I did, they did actually send me in an email, this little, what do you call it if I bring it up? You know how Spotify does the unwrapped and then everyone has started copying them? Buzz Sprout, which is the platform that I use for the podcast, sent me the same kind of thing. And if I try and go find it, it was three bros. Yeah, here it is. And I remember that the number one most viewed episode was still episode zero, overwhelmingly. And at this point, the stats on the episodes are like, I think we have 84,000 listens.

on episode zero, like in aggregate, which accounts for almost like 10%. Like we have like 1.1 or 1.2 million downloads in total. And that is a lesson for future podcasters. Make sure episode zero is a banger and your audio is on point. Because I'm sure our viewership would be like some single digit percentage higher if we had higher quality audio. And if I go to check this playback here, I can share my screen for a second. I'm just asking an LLM while you're doing, while you're doing this manual work, I'm just asking an LOM, what episode of ADSP the Bodd Task in 2025 had the most views slash listen slash downloads.

We published 25 episodes, unsurprising, podcast location. So you're looking at this, right? Maybe I'll post this. I thought it was kind of silly, so I didn't. I am attempting to look at it, but my bandwidth seems to be very bad. So right now it is just showing me a big C.E. for Connor. But I imagine if we wait another five seconds or so. Well, the thing is it's like

an animated video. So it's saying the United States was our most popular country for listenership. And we were listened to in 23,429 different cities. The most city, most popular city was New York. And

Hey, my people. And top countries. So that is United States, Germany, and United Kingdom. Top Cities is New York. New York, Sydney, and Stockholm. Although the animation has gone by. We got over a quarter million downloads in 2025. One episode really stood out.

And that was episode zero, our favorite algorithms with 8,460 downloads. Didn't even come out this year. Or the last year. Or the last year. And we're in the top 5% of podcasts. Whoa! And it's going to show us our other top episodes, I believe, which were episode 247, the philosophy of good software design, and then episode 244, high on AI, part one. The first one.

247, 2025 started with, okay, so that was a 2025 episode. Yeah, both of those were in the summertime. time. One had 3,900 views and the other one had 3,700 views. I'm not surprised the high on AI one is up there. And then there's a summary of ADSP, which shows five years of podcasting, 52 episodes published, top 5%, quarter million down, I guess it's just all the same stats. And then the total number of minutes we published were 1,711, which divided by 52 is probably going to be, what, 32, 33 minutes or something like that? So I'm going to give people a pass because my favorite episode of the year was posted in

December and, you know, nothing happens in Tekken in December. Well, I guess aside from Nvidia, signing deals. But other than that, other than that. Jensen's very busy on Christmas. Other than that, nothing happens in December. So we'll check back in a few weeks and see whether my favorite one is really picked up steam. But do you know which episode I'm thinking of? If it was in December, I'm going to guess it was the Sean Parent origin story, correct?

Yes, of course. It has to be. And I was so sad that I had to leave the hotel because was I flying out? I think I had to go catch my flight. Your flight was the next morning, but it was early in the next morning, and you had to catch the last train to Amsterdam, I believe. Yes, I think I had to leave that night. I had to get on a train. Like, if it was just like I had to, if it was just a question of I would not get any sleep, I would have stayed. But it was like, I needed to be on a train and this was the last train. And so I had to depart because I stayed at a hotel that was actually like inside the airport past security, which was an interesting experience. But yes, I was sad that I had to leave. How much longer did you guys talk after I left? I, well, so I don't even think people noticed in the edit because I cut out when you were leaving. And then I think we recorded for another maybe, I can't remember if it was 10 or 20 minutes.

And so I think everybody thought you were there for the whole time. Yeah, and I only think it was 10 to 20 minutes. We finished that. And then we ended up joining a bunch of folks that were at some restaurant. And we had way too many people at a table, but it was a lot of fun. And then, so we hung out for a few more hours, but we only recorded for 10 or 20 more minutes. And what's interesting is that, so here are the stats. And if you do a reverse max scan on the episodes, you can kind of see the ones that on, like, per week have more.

So actually, those two Sean Parent episodes do have more than the previous two weeks. but that is only for those two weeks and then the NDC Tech Town with J.F. Bastion and Vittorio that one has 45 views more. And then algorithm taxonomy did pretty well. So then that's the ones that kind of stand out, right? But then the algorithm taxonomy

that skips up to 3.200 or 3,200. And then if we keep scrolling back, so when's the next time we have one that's higher? It was the last time we had Sean Parent on before that. So that probably does indicate that the Sean Parent ones will generate more views over time. And then, yeah, this 4001 philosophy of software design.

Wait, wait, wait. Which recording of those Sean Parent ones, those were September? That was when we did it on the Zoom call. So that was late August when we recorded. And we, I think, recorded for, I don't know, a couple hours. And then, yeah, we diced it up. This podcast really is a discussion with Sean Parent. There's really a lot of Sean.

There was a couple different. It's ADSP. A discussion with Sean. unparent, algorithms plus data structures equals programs. There was another one. Do you remember what it was? I can't remember. Anyway, so I will say my favorite.

Well, yeah, I think I didn't spend a ton of time thinking about it, but it was the one in March. And I'm actually sad because we didn't, I didn't update it, because I'm pretty sure I added a, did I add? Oh, no, I didn't. I guess I only add emoji flags when we're at conferences. That looks like the pattern, because we have recorded. live from different countries,

but, and I guess you're based in the state, so that's probably why I didn't do it. But we recorded, if you recall, after that comedy show, and we went back to the hotel. I do. And I just had a ton of fun. Whenever we're walking and talking.

Yeah, that was a good one. That was a good one. And also, too, in hindsight, we commented on, or you commented on the Jewish comedian was your favorite. There is, he's famous. Ari Shafir? Oh, really?

Yeah, he's a famous comedian. Like, he's been on Joe Rogan. He's in, like, the Joe Rogan sphere. And... What, we're not going there, Donna. Anyways, so, yeah, no, I guess our listeners, maybe we lost a couple listeners there because you liked a comedian that's been on the Joe Rogan podcast. Anyways, after the fact, I realized that, like, oh, yeah, I've seen that guy before.

And I'm in most, I most famously know him. from Shane Gillis roasting him. Anyways, we're at the 30-minute mark. Are there other things we want to retro about... Well, okay, so I had asked you about this in December, and maybe the listeners can provide some input. I would love to do some sort of poll to find out what you, the listener, what your favorite episode was from 2025.

And I kind of had in mind, like I sometimes see on Twitter where people do like multiple rounds of polls like March Madness. Like, you know, you make a bracket. You do, you bake off a bunch of different, you know, you take four and you say which of these four is the best and then you go to the next round and you do a bunch of those polls where you do, you know, sets of four and then you move on to the next round and etc., etc., etc., until eventually you advance to a final single poll that determines the winner.

But that's like a lot of polls. and I don't think I don't think I'll be able to retain everyone's attention for that long. So then what is the best option? Do we just go and pick like the 12 episodes that had the highest downloads or 12 episodes that like we like the best and put them into a poll? Or like what's the best mechanism? There must be 52 episodes from the air. So what's the best mechanism for us to let people vote on their 52 favorite episodes? on their favorite episode of those 52.

And also, too, I mean, we already solicited feedback earlier about the Tesla topic. But if you have feedback in general, we've been doing this for over five years now. And if there is, I mean, we've heard different things from different people, you know. Some people like the 10xing, guys. Some people don't like 10xing. Zach Lane. He said that's his like favorite thing about the show. It was whenever I 10X your voice.

I've heard other people, they said, Their favorite story was the back and forth during COVID about you sending me, like me getting COVID, not being able to get a vaccine, or what was it a test? And then you like sending me the test. And then other people, they say that their favorite stuff is the algorithm talk. Anyways, if you let us know not just like what your favorite episode is or how you want us to do that kind of vote, but like what's your favorite part or is it all of the above? Anyways, we're open to feedback.

You know, we want to keep our listeners happy. And, yeah, maybe we'll do, maybe Bryce will do something if we get a good enough suggestion on some kind of poll or something like that. And I guess we could, we could, I know how we could encourage, we could encourage participation is we could give out some prize, right? If you, if you participate in the poll. I don't know what sort of prize we could give out. Do you have a prize to give out, an Nvidia ruler or something like that? I could make ADSP t-shirts or something. That was a quick upgrade.

We finally got stickers, and then a couple months later we got T-shirts. Well, I had to have T-shirts made for another purpose that I cannot disclose on the podcast. But it was pretty straightforward. All right. Well, there you go, folks. Maybe the potential to have a T-shirt. There's no promises here because. I'd be very surprised if these t-shirts materialized, but if they do, you might be the proud owner.

And how do they win it? Is this anybody? Like, no, like, you know, we'll randomly choose five people who, uh, who comment on some platform. Participate in the poll. Yeah, no, no, no, no, who vote on the poll of the, what was their favorite episode of 2025. Okay. Well, have we decided we're doing a poll yet? Uh, if I can find a way to construct the poll, then yeah.

I want to research. There's got to be some clever way of doing this. Okay. Like giving people just 52 options, it cannot be the best way. Like I don't, doing the bracket is like, you know, a good way, but then that requires you to do multiple rounds, right? So people have to be engaged multiple times. I guess, like, asking people to, like, just pick from a list of 52, I feel like that's, that can't be the best way. There has to be theory on this and what's the best.

way to to poll. Well, I mean, technically on this episode, we could just say if you have a favorite episode, go leave a comment within the GitHub discussion. Or maybe I could even create like a separate GitHub discussion for a favorite episode. Go leave the number. And then if people see it in the future, they just upvote it. And then as you're missing such a wonderful opportunity to talk about like ballot design and like algorithms for determining consensus.

You're making it too complicated. You're making it too complicated. Jeffrey, and I'll say this, Jeffrey Askin, who was the Library ofolution Chair when I started on the C++ committee, best library evolution share, I think we've ever had. He did like all this research on polling methods and like, you know, techniques for determining consensus and he had all this theory crafting on it. He would know, he would know what the right mechanism is. He came up with this like, these strategies for how we would prioritize stuff.

Like you're given a list of like a hundred things. How do you ask a group of people to prioritize those hundred things? And one of the things we used to do was we would, you would take an item and you'd ask everybody to raise their hands and you would start at the top of the list and you would ask people to put their hands down as you moved the item further and further down the list you'd ask them to lower their hands when the item reached the place in the list that they thought it belonged and then when roughly half the people in the room had lowered their hands that's where the item would live and it was like a pretty fast way to go through and prioritize a bunch of items because like you just you just go through and then just like once you know some enough hands had gone

down to be like all right and you didn't have to like ask people to uh like say like oh i think it should be 24th or 25th or something there was no like discussion about it it was just like each person individually would put their hand down and i think it was a really a really good strategy All right, well, we're not doing that because we don't have people sitting in a room. What if you... I think we're putting way too much thought into it because in honesty, I can't even recall half of the episodes that we've recorded by like number or title or what we talked about. I have already in my head just like a handful that were the favorites and most memorable for me.

And I think it would be generous. Like, what are the people recording it and producing it? I think it would be very unlikely that our listeners have more of a memory for our episodes than we do. So they probably only have one or two. So you just go ask them to leave a comment and then they upvote it and then we're done. And if we get some overwhelming, you know, 20 different episodes and people disagreeing, then we just look at the upvotes and then we have a final poll or something and that's it. But, you know, it's way too complicated.

But your method produces unclean data. It's just like a series of textual comments and maybe upvotes on a GitHub discussion. My method's more fun. I'm going to do research on this. I'm going to do research on this. I'm going to ask by research, I mean I'm going to ask chat GPT. I tell you what, you build some kind of voting mechanism. People can also just go leave comments on this GitHub discussion.

Or maybe I'll set up a separate one. Just link in the show notes for a link to somewhere where you can leave a comment. And then if Bryce actually puts this together before Friday when this episode drops, then I will link to his as well and we will see which is better. All right, anyways, we're almost at the 40 minute mark now. Is there any in the last 30 seconds here, anything else, retro 2025, things we want to mention before we go forward to our 2026 look ahead, predictions, future episode? No, I think, I think this has gone on as long as it should go on.

All right, folks, that's episode. What episode is that? That is episode 269. We are now on to 270. 2025. It's in the rear view, mere. 2026 is here. It is January 13th as we record this.

Be sure to check these show notes, either in your podcast app or at ADSP the podcast.com for links to anything we mentioned in today's episode, as well as a link to a get-up 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.