commit

audio/52.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e52db3d1baae13cb60d3b4e75636edccf4dc4bac3e309c1b7a95e49092a0431
+size 843884

audio/53.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61fec2fa4c42c0de16de1b43b36c3ae289d0beade9694c8808c6811aa4056956
+size 2620844

audio/54.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8068fe124987d739fa55b48277cab1125a911737c471048b0fbcf37c6e10dd0f
+size 2431438

audio/55.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:878736f5276281480be2fee8ea026c1c70ccacab4be9dcf4a3ec6653709da3c3
+size 930284

audio/56.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6c71a8e66f3edfa8422d29f011514aa0446f52583d18d5b793331547684a77f
+size 2431916

audio/57.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a7def0a97459befe1cd55cbc452d8e0695a030ed817f24722516ce8bea0ea7c
+size 1655915

audio/58.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fb3937ecc7b74acfb6467ef0025fecd22ab047d100b82821d6223ec8bbe51fa
+size 940076

audio/59.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2127d470dc3d151dbbd356895d9e53333dcab6a26561dd9b552f0d4e0a26fd6a
+size 1150316

audio/60.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:491df709dd55b5d7420a06d960920a1221007d180b4a22de147baffc7ee7f518
+size 490436

audio/61.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df80ae5b349edc19406f35b00bf5bab5ce7f6338d4aa58f321ec2d87d144f6ce
+size 672236

audio/62.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13217f7c5b152ea405d4ac8dce65725d8f35a8ff3ce781bd810b30d4434a2301
+size 755756

audio/63.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be7f0521dd20bb1a4fbe5fae2412edf74a99880eea0fb6a77d371aa7e711b3ee
+size 2551724

audio/64.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a79a5f58f68686dd5e210877eae1b71a9b8a1ca4df8c7b2d3be244475f6f0b5
+size 2002612

audio/65.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cf8d2c1e254dcfffc4ed357458a074487d8dbf9e8fcc561c2809e409767fe55
+size 1270124

audio/66.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d438450a327cd29a1f0605124f909fdea7c03042e59ac275979ede93ac0da00
+size 1170664

audio/67.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45ee1956e721aadcba20ed615e56404f2dea0151583ee77b1588e137bbeababd
+size 1188846

audio/68.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db9bedc5f245d7dd5c8d5b898569aa35977b33e5cee4dee761cf0ada4e90af39
+size 2375641

audio/69.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fecce6c836b3ac7680556524183a4515f4820c403bce2b3752bcdb27327a77c4
+size 1755116

audio/70.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c1bfab7018f4772aadd2b01974f6ae87fc587e1d9c4cdaa7181d169c6264636
+size 915884

audio/71.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0db9d33707a61c2db8d870810919f4dbfe502747dc36e0e527b79ef7a1083607
+size 911276

audio/72.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:789515ffb288610200283d99ee8065e6b542f8230e78d85b08587e63b3392f35
+size 3753652

audio/73.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c8fbeae9c6e2995d5f79b519678f0e5e3f96c4ac67f700ed79568e0b1eab2a4
+size 1511092

audio/74.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fafd63e7481ad9c465e1807c30d541181409eeca8055adb3209895d617477bc4
+size 748844

audio/75.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76c6129f011ea391a5e15cf8546b9332f1845061cbe7ebd60e42ca6735268b7a
+size 2875436

audio/76.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55b24b69b61156e384b81ec5ac4fc3b09ec57d7bfe69bfcf74d67a6e3a7f9f14
+size 980396

audio/77.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83a57cf2058106a894108649ce73bffd6d562a0d73e66eb11e0009e4e922a080
+size 1134764

audio/78.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b61df3605409a546cea32f231e525a0904b834d9e9a058d49eb0b1e74ed26062
+size 1196996

transcripts/uncorrected/52.txt

@@ -0,0 +1,3 @@
+For Cloudinary, I need some repetitive URLs benchmarked or saved, I should say. And I should have also an account for DSR Holdings. I'm not certain that I do.  
+  
+Maybe it's a spreadsheet, maybe it connects to, I can integrate it with the backend. But every time I need to use the logo, I don't want to be looking up the path of that every single time for email banners, etc.

transcripts/uncorrected/53.txt

@@ -0,0 +1,11 @@
+I was asking for a recommendation yesterday. Recommended a very good product, but I forgot the name of it on AliExpress. This is the stuff.  
+  
+So for our bedroom, I'm looking for a speaker that would be good for playing sound. We have a projector and a smart TV. So the connection can either be one of Bluetooth or it can be a 3.5 millimeter connection. But there isn't much space to work with. So it has to be a very small speaker and it has to be one speaker.  
+  
+One speaker that gives a, packs a lot of punch for its size and delivers a wide sound. It's going to be positioned at approximately the base of our beds, like on the footboard. We're gonna be in front of it.  
+  
+So what I'd like is, what I'm trying to find is one compact speaker that has a very good, you know, 180-degree throw of sound. It doesn't have to be 360 because it's, you know, the wall's on the other side, but something compact as well.  
+  
+So for example, a circular Bluetooth speaker is very good, but it's not angled correctly and it generally doesn't really throw the sound too far across.  
+  
+So I'm looking for something like a, obviously it's not gonna be as good as the sound bar or a proper surround system, but something that's as good as you can get in a compact size and a single speaker.

transcripts/uncorrected/54.txt

@@ -0,0 +1,9 @@
+Is there any speaker? My wife and I are living in a, watching movies a lot in the bedroom with a projector. It's a very small room. We rented it in order to put something permanent.  
+  
+So I'm thinking like, even if this wasn't, let's say, the greatest speaker ever, I use, for example, a Sony XPS 13, and it's pretty decent Bluetooth.  
+  
+So, but I want to keep that for use on the go. You see what I bought it for? Let's put something of that nature, like a little compact circular speaker, that we can maybe creatively affix to the bottom of the bed, on like a telescoping thing, and just kind of pivot it forward a bit, or just leave it there.  
+  
+It tilts it forward a bit so the sound comes towards us. I'm just trying to think of something that would, like a small circular speaker, that really delivered impressive sound for its size.  
+  
+I've definitely heard a couple of them, and ideally would kind of give a good, let's say we position the circular speaker in the middle of the bed, that it would give quite a well-developed array of sound covering both sides of the bed.

transcripts/uncorrected/55.txt

@@ -0,0 +1,9 @@
+I'd be very interested to know how Sonnet 4 and Gemini Pro 2.5 fare off one against the other in terms of agentic coding abilities.  
+  
+Sonnet 4 is kind of the benchmark so far, but I feel it really struggles with context retention.  
+  
+Although it's just maybe massively challenging for an LLM to attempt to do that, I find it very hit and miss using it in Windsurf.  
+  
+It tends to fix one issue and break another.  
+  
+So I'm just wondering if in that respect, what benchmark might evaluate for that and any meaningful differences that have been observed?

transcripts/uncorrected/56.txt

@@ -0,0 +1,7 @@
+So, this folder consists of its mp4s from an IP camera, an overnight shot, and each mp4 captures a 2 minute period. The timestamps are noted in the file names.  
+  
+What I'd like to do in this repository is for the script to do as follows: for each hour, that should be concatenated into one recording. So, for example, between 2 and 3 am, that's going to be determined in the timestamps, or the nearest clip to the beginning of each hour, until the nearest clip to the end of each hour. That should be concatenated into a single mp4.  
+  
+Once that is done, as a second step, the clip should be compressed as far as possible while preserving a reasonable amount of quality. So, stage 1 is concatenating it into one clip per hour, and then stage 2 is running a compression script over those concatenated files in order to compress them a little bit.  
+  
+So, the end result should be in a separate folder called processed clips. We'll have however many hours there are, one clip per hour, and in a compressed format, preserving the audio and the video.

transcripts/uncorrected/57.txt

@@ -0,0 +1,7 @@
+Alright, so we have a... I added a Zigbee lightbulb into our living room. And I would like to configure a lighting setup.  
+  
+So here's the thing. I have aggregated all the lights for each room into one single group entity. And in this way I can turn off the light. But then I realized if I do that as a toggle, when I toggle on of course it will turn on all of the lights. Which I don't want.  
+  
+One in prayer to think of is kind of a gentle light in one area and then two switches to turn off. So do you think the best way to do that is to create a scene for these? Because we will almost never want all the lights in a zone turned on.  
+  
+So is it really just I want a scene as a single one and then I want a double action to turn off or do you have another suggestion?

transcripts/uncorrected/58.txt

@@ -0,0 +1 @@
+So if I create an assistant in VAPI and I want to be able to call it at a number. So I have my phone number from Twilio here. What do I need to do in Twilio to configure the routing that it will go to the assistant?

transcripts/uncorrected/59.txt

@@ -0,0 +1,5 @@
+I'll try to see for any demo. I'll connect both for object storage. I'm going through as many locally that I need to connect. If that's in any way useful, I think it's very less useful as the CLI is all that's available now. So probably anyway, more it makes a lot more sense of external storage.  
+  
+The ones I created for B2, I want to have that account as a credential and for personal storage as well. I have B2 actually listed there. I just want to make sure that that key has access to all the buckets.  
+  
+And for Wasabi, I will add as well just a couple of corresponding buckets should I choose to use that for storage, but I think B2 is a better fit.

transcripts/uncorrected/60.txt

@@ -0,0 +1 @@
+What makes more sense to connect to locally? Is it RTSP? On my Android to get the cameras Or is it the restringed HLS? And if so which app? I see the Tinycam Pro Seems to only support RTSP

transcripts/uncorrected/61.txt

@@ -0,0 +1,3 @@
+Okay, so my question is this. If I, as I plan, connect the UI to Cloudflare so I can look at the things externally, will the streams pass through the tunnel, proxy through the tunnel, thanks to the configuration that we have set up here?  
+  
+Or do I need to create dedicated routes to each of the streams and then have an external NVR in some other way?

transcripts/uncorrected/62.txt

@@ -0,0 +1,5 @@
+See if it's possible to take a similar approach to the Hugging Face repositories, and they should probably be consolidated in the menu at least.  
+  
+Projects can have repos, and maybe we can have the Hugging Face in GitHub just as kind of tabs on a navigation.  
+  
+They can stay as separate pages, but if they could be kind of grouped navigation, that'd be excellent.

transcripts/uncorrected/63.txt

@@ -0,0 +1,5 @@
+Dr. Popplebury, so there's, it strikes me that there are contradictory, there's contradictory information out there about white noise in children. We discussed before that the general position is that so long as the volume is not damaging to hearing, the white noise is pretty innocuous. I've been using white noise for years and I'm a believer in it. I think it's actually therapeutic personally, but that's colored by my sound sensitivity.  
+  
+On the one hand, you said that it's used in NICUs, which is a pretty strong endorsement that it is indeed safe. On the other hand, we read these reports of it being potentially damaging. So it seems to me they're very opposing positions. How do you think that they can be squared in your opinion?  
+  
+The second question is on binaural beats. These are often marketed in wellness communities. I'm not sure what how robust the science is behind them in general but is there any evidence for their use in with infants? I think that more than going to sleep getting them to sleep the objective for the period we're in of four to six weeks is trying to get them to calm and remain asleep for a sufficient period. And so on.

transcripts/uncorrected/64.txt

@@ -0,0 +1,5 @@
+I would love to get your thoughts on a suitable hardware, something that could be used as a control surface for a kid in life. Could be a macro pad of sorts. The things I would like to have when I'm editing video are usually sliders for, well knobs probably for subtle controls, brightness and contrast. Colour correction typically benefits from that kind of stuff too. Volume control.  
+  
+I'm thinking really about, I think I've seen a couple of projects over the years whereby of course there aren't any official control surfaces as there are for the bigger editing tools like DaVinci and the Adobe products. But I believe I've seen people repurpose and create their own control surfaces from a whole bunch of different products, maybe tended for slightly different issues.  
+  
+Given the constraints and limitations of Ubuntu, Linux, what would be my best bet for a control surface and is there anything kind of classic that people use for this purpose?

transcripts/uncorrected/65.txt

@@ -0,0 +1,5 @@
+So I've been in the process of taking my main what were previously assistants and converting them into agents. Right now, predominantly fed by voice notes, but they, of course, can be more versatile than that. And I'm just seeing at the moment how they work.  
+  
+I also want to add today, I think, a balance into OpenRouter. While I'm doing all this experimentation, I might be racking up some significant costs with the APIs. And I think it would be good to have those, just to know that there's a good balance in the account, maybe like even a hundred bucks or something, and give myself plenty of room to experiment with the different models. It certainly gives you the best value for money in terms of being able to use different models and iterate basically.  
+  
+Likewise for Hetzner, the critical infrastructure of this process, maybe figuring out if I can automate the payments as a high priority. We checked with their support today for that, or creating a balancing account.

transcripts/uncorrected/66.txt

@@ -0,0 +1,9 @@
+Generate a lyrics to a song about a sloth called Cornelius who on a very hot day in Jerusalem decided to go do groceries in the market.  
+  
+Cornelius bought a lot of wine because he loved to drink wine but his little paws couldn't carry it and they ended up bursting.  
+  
+His bags burst and the wine spilt all over the market.  
+  
+But there's a happy ending to the story because people doing their shopping were holding up, filling up their cups from the ground and filling their glasses with the wine that broke.  
+  
+Everybody joined in in a celebratory toast to Cornelius the Sloth.

transcripts/uncorrected/67.txt

@@ -0,0 +1,3 @@
+I recently installed some IP cameras and learned the basics of PTZ control, different protocols for it, etc. Out of curiosity, when you see a building that has, let's say, really really well wired up with security cameras, let's take a government building for example, because I used to live down the street from one exactly like this. They have like maybe 6 very conspicuous big cameras. How much would those cost?  
+  
+And at that level of professionalism, moving away from my amateur home use case, what kind of NVR would be used? And how would remote access be made available if I needed to remotely monitor the cameras?

transcripts/uncorrected/68.txt

@@ -0,0 +1,5 @@
+I'd like you to run a couple of calculations for me. Firstly, try to find a speech-to-text algorithm or API that is very cost-effective. Ideally, one that is both cost-effective and affordable, however it can be implemented.  
+  
+Next, take someone who's recording speech almost continuously during the working hours. So let's say about 16 hours per day and then remove some time for showering, eating, toilets. Let's just say, just to work with this example, that the person was involved in some occupation or doing something that they were essentially always recording speech, sending it off maybe in loops, but doing that as much as one human could.  
+  
+And then combining that number with the cost of the most cost-effective API that you can find and one that is then a couple of cheaper, more or less cheap ones, the ones that are more expensive but ones that are all in the category of let's say performant ASR and then calculate the daily cost, the daily API costs for those three models under the same assumption of whatever we calculate this continuous recording to be.

transcripts/uncorrected/69.txt

@@ -0,0 +1,11 @@
+Now that we've finished development of the first version of this app, could you do the following?  
+  
+I'd like to create two versions of it in the build script.  
+  
+The first would be for my computer which is Ubuntu Linux so a Debian file would be preferred.  
+  
+The second one would be a Windows MSI file.  
+  
+Could you create a build script that will create a V1 in the repository with both of those targets and the build script should be excluded from the git ignore.  
+  
+And finally...

transcripts/uncorrected/70.txt

@@ -0,0 +1,9 @@
+Okay, the second job is for Homebox, the Homebox volume.  
+  
+Could you create a script? This is just a one-time script, so it can be at the root of the repository or wherever you want.  
+  
+And what I want to do is just create a nice compressed archive of the Homebox data volume today.  
+  
+So give the file name the timestamp 040825 underscore whatever Homebox data.  
+  
+And try to create the most space efficient archive possible of the data volume as of today.

transcripts/uncorrected/71.txt

@@ -0,0 +1 @@
+Could we now create a separate file within each a CSV within each collection currently just for for AI operations and AI outputs, listing the collections within that workspace, returning it as a CSV collection name, collection ID.

transcripts/uncorrected/72.txt

@@ -0,0 +1,11 @@
+So I have an AI agent that is doing quite well and has proven quite important. It's for answering parenting questions which my wife and I have.  
+  
+So the current workflow is that we ask, I created two ways, two ingestion routes. The first one is, I'll use an app called Voice Notes, which is a voice note processor. It will transcribe my voice notes and then send a webhook which will pipe it into the N8N workflow and that will then create a note in Nuclino which is a note manager that I use.  
+  
+Email us the answer. The agent is instructed to provide HTML and Markdown so that I can put the Markdown version to Nuclino and the HTML to the email. It would, of course, have been easier to just use one output format. But if I did that, it wouldn't have necessarily adhered to the specific HTML instruction which is that it needed to be HTML for internal injection into an email template without outer elements.  
+  
+So that's why I in the end ended up getting the agent to just do the two output formats: one is Nuclino, as mentioned, the other is email.  
+  
+The one that I've been struggling with, I would love to get it into a Google Drive folder. I've been having a really hard time figuring out how to do that, which seems very odd. It's very challenging to just create a Google Doc from either of these.  
+  
+Which one is the best, is a better starting point, the HTML or the Markdown? And is there a node that can just reliably create a doc from content in a shared drive? Because that's what we need.

transcripts/uncorrected/73.txt

@@ -0,0 +1,7 @@
+I want to see if there's a way in Home Assistant to create a binary entity whether anyone's at home. I want to have a condition that if we come home, air conditioning will turn on automatically, but then I would need to know that we were not at home previously.  
+  
+The logic would be something like if no switches were changed for an hour perhaps, then the home empty state will turn on or home vacant or something like that. What do you recommend the logic would be? I don't want to create false positives, obviously.  
+  
+Maybe it'll be during certain times of the day. Or is there any pre-built way to do that? I don't like the idea of using human presence detectors around the house. I find it a little bit invasive.  
+  
+So I was thinking that you could just kind of figure it out based on whether anything was turning on or off.

transcripts/uncorrected/74.txt

@@ -0,0 +1,13 @@
+I'll create an agent for creating the JSON schema and the JSON examples.  
+  
+User will provide the system prompts and say what, and then describe in natural language which elements they wish to standardize.  
+  
+It's going to be streamless.  
+  
+And then it will generate both of those.  
+  
+It can be a form input, form output actually.  
+  
+Returning to them with the items.  
+  
+Thanks for watching!