← The AI Daily Brief: Artificial Intelligence News and Analysis
The AI Daily Brief: Artificial Intelligence News and Analysis
The Surprising Way AI Expands Markets Instead of Capturing Them
00:30:20
Loading summary
Transcript2 lines
- [00:01]
A
Today on the AI Daily Brief how SUNO uniquely might show the future of AI While in the headlines, SoftBank greenlights a big OpenAI investment contingent upon their for profit conversion. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Alright friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG Robots and Pencils, Super Intelligent and Assembly AI. To get an ad free version of the show, go to patreon.com aidaily brief or subscribe on Apple Podcasts and to learn about sponsoring the show, send us a Note@ SponsorsIDailyBrief AI I also want to announce something that I'm really excited about. I am convinced that in 2026 a huge amount of attention is going to shift to demonstrating the performance and ROI of AI deployments. Unfortunately, right now I think we're all flying a little blind and with that in mind, I'm excited to announce the 2025 AI ROI benchmarking study. Basically, we want to know which use cases are driving the most value for you and what type of impact you're seeing. Whether it's time saved, cost saved, new revenue generated, new capabilities unlocked, or something else. The survey can be found at roisurvey AI or on a link from my aidaily Brief AI website and you can contribute by adding as little as a single use case. Just share the name of your use case and you can choose from a list of primary benefits. Each use case is going to take you less than a minute to add. Anyone who shares between one and three use cases will be entered to win an Amazon gift card or get AI credits of an equivalent value. Anyone who shares three or more use cases gets a big report at the end of this, not only summarizing the results but also providing an ROI framework. And then the person who shares the most use cases gets a one on one with me. The survey is live now again at roisurvey AI and I'm really excited to discover what use cases are driving the biggest value. Welcome Back to the AI Daily Brief Headlines Edition. All the Daily AI news you need in around five minutes. Earlier this year, SoftBank committed to make a $30 billion investment in OpenAI split across two stages. The second stage, which is now approved, was contingent on OpenAI completing their for profit restructuring. According to sources speaking with the information that proviso is still in place, but the board is otherwise ready to go ahead. Now it's unclear whether we should draw any conclusions about the likelihood that OpenAI's for profit conversion gets approved by the California attorney general. Former OpenAI safety researcher Miles Brundage suggested that SoftBank might be getting a little ahead of themselves. He argued that public information about the restructuring has been, in his words, all bad so far. On the other hand, Microsoft has reached an in principle agreement on the conversion, so so there are no remaining roadblocks from investors. Meanwhile, for those observing Softbank's moves in capital markets, it's far from clear how the funding for the deal is going to come together. Earlier this month it was reported that Softbank were trying to borrow 5 billion from global banks pledging their ARM stock as collateral. More recently, they were said to be tapping global bond markets for 2.9 billion, funded in both euros and US dollars. This weekend, the Japan Times reported that softbank had issued 2 billion in hybrid dollar bonds across an ultra long 40 year term. The instruments would be subordinate to senior debt and allow Softbank to defer interest payments. Initial pricing has the bonds priced at around an 8.5% interest rate, well above the 6.8% for SoftBank's existing long duration bonds. The Japan Times writes, The fact that SoftBank CEO Masayoshi Son is resorting to issuing expensive dollar and euro denominated hybrids smacks of desperation. Indeed, SoftBank has been very busy looking for money lately. They noted the margin loans, sale of T Mobile shares and record breaking issuance of yen denominated debt. Adding, what has largely been missing amid all this action, however, is good old fashioned bank loans. As an Investment holding company, SoftBank doesn't have reliable operating cash flows to boast about when asking for mega financing deals. As a result, sun has come to public markets. Ratings agency S and P Global, meanwhile, has said that they would consider downgrading SoftBank's bonds if their loan to value ratio gets above 25%. Whether you're bullish or bearish, this is certainly another sign that we are in uncharted territory when it comes to the AI buildout. Next up, Mistral has launched a new enterprise control center called AI Studio. The new platform will provide agent building, orchestration, observability and governance tools designed to help enterprises deploy AI at scale. In an announcement blog post they wrote, mistral AI Studio brings the same infrastructure, observability and operational discipline that powers Mistral's own large scale systems. Now packaged for enterprise teams that need to build, evaluate and run AI in production, the new platform highlights just how extensive Mistral's model range has become. The company now offers 19 different models, including both proprietary and open sourced as well as multimodal coding specific and speech enabled options. The platform represents an evolution in the sophistication of AI tools for business. One interesting feature offered by Mistral is called AI Registry, which serves as a system of record for all AI assets across the company. In other words, enterprises can track every agent, dataset, tool and workflow, registering their ownership and versioning throughout the production lifecycle. The system can manage access controls, moderation policies and a promotion pathway to full deployment. It also integrates directly into observability and orchestration tools, Mistral writes. This unified view enables true governance and reuse. Every asset is discoverable, auditable and portable across environments. Mistral is basically articulating a view that raw model performance is increasingly going to give way to governance as the most important aspect of enterprise deployments, they wrote. Enterprises are entering a new phase of AI adoption. The challenge is no longer access to capable enough models, it's the ability to operate them reliably, safely and at scale. That shift demands production infrastructure built for observability, durability and governance from day one. Certainly this hearkens to all the things that we've been sharing about what we've been seeing at Superintelligent with these recent episodes. And as this Mistral AI studio rolls out further, we'll see how it is received by the market. Next up, an interesting Partnership Stability AI has signed a partnership with EA to provide AI tools for the game making process. EA said that the two companies will quote, co develop transformative AI models, tools and workflows that empower our artists, designers and developers to reimagine how content is built. Now at this point, using visual AI tools for everything from initial design to final asset generation is increasingly commonplace in the games industry, so the partnership doesn't come as much of a shock. In one example, also from last week, PUBG developer Krafton said they were becoming an AI first studio and building their own GPU cluster to support the effort. The interesting part is that EA is rapidly executing on the AI strategy that underpinned the decision to take the company private. The Financial Times reported last month that the investor group were quote, betting that AI based cost cuts will significantly boost EA's profits in the coming years. EA could have retooled using AI while remaining a public company, but going private affords them the ability to move quickly while frankly ignoring what might be an inevitable backlash. Business Insider reported last week that EA was facing morale issues due to a broad AI mandate premised on using faulty tools. Gaming Reddit is of course, no fan of AI cost cutting by a developer already criticized for the quality of their releases. A complete AI overhaul then, is much safer without the risk of the stock plummeting on negative headlines. As EA figures out how to get this right, it's also intriguing to see Stability AI granted a second life as a bespoke AI partner. In 2023, stability was one of the hottest AI startups with the success of their stable diffusion image model. Since then, there's been acquisition talk, the resignation of their CEO, and a debt restructuring to keep the company afloat. Given how much enterprise demand there is for serious talent to rebuild from inside, this might be an interesting pathway as we get to something more of a consolidation period in the AI industry. Lastly, an interesting one that comes from a talk in San Francisco Thinking Machines Lab believes that learning, rather than scaling, will be the next big unlock for AI models Late last year, an entire narrative cycle played out around model scaling hitting a wall. The major labs were seeing disappointing results from scaling up training data sets and using more compute for training runs, leading to widespread concern that model performance had plateaued. OpenAI then released 01 and demonstrated that reasoning and test time compute were another avenue for model improvement. With improvements to reasoning now slowing down, there has been a large focus on context engineering tools like advanced memory. Many believe that continuous learning will need to be developed to unlock the next big jump in model performance. Speaking at the TED AI Conference in San Francisco, Rafael Rifalov, a reinforcement learning researcher at Thinking Machines Lab, said, I believe that the first superintelligence will be a superhuman learner. It will be able to very efficiently figure out and adapt, propose its own theories, propose experiments, use the environment to verify that, get information and iterate that process. He doesn't expect that adding training data will be a viable path to superintelligence. Commenting Learning is something an intelligent being does. Training is something that's being done to it. Regarding reinforcement learning, though, he still thinks there's a lot of space to explore, arguing, I don't believe we're hitting any sort of saturation points. I think we're just at the beginning of the next paradigm, the scale of reinforcement learning, in which we move from teaching our models how to think, how to explore thinking space, into endowing them with the capability of general agents. Ultimately, Raphael is looking to apply the same techniques that allowed models to learn to code and to search the Internet, to learning itself. He commented, learning in and of itself is an algorithm. It has inputs the current state of the model. It has data and compute. You process it through some sort of structure. Choose your favorite optimization algorithm and you produce hopefully a stronger model. I believe that under enough computational resources and with broad enough coverage, general purpose learning algorithms can emerge from large scale training. The way we train our models to reason in general over just math and code and potentially act in general domains, we might be able to teach them how to learn efficiently across many different applications. Interesting stuff to kick off this Monday, but that's gonna do it for the headlines. Next up, the main episode what if AI wasn't just a buzzword, but a business imperative? On youn can with AI, we take you inside the boardrooms and strategy sessions of the world's most forward thinking enterprises. Hosted by me, Nathaniel Whittemore and powered by kpmg, this seven part series delivers real world insights from leaders who are scaling AI with purposes from aligning culture and leadership to building trust, data readiness and deploying AI agents. Whether you're a C suite, executive strategist or innovator, this podcast is your front row seat to the Future of Enterprise AI. So go check it out at www.kpmg.us aipodcasts or search you can with AI on Spotify, Apple Podcasts or wherever you get your podcasts. Small, nimble teams beat bloated consulting every time. Robots and Pencils partners with organizations on intelligent cloud native systems powered by AI. They cover human needs, design AI solutions, and cut through complexity to deliver meaningful impact without the layers of bureaucracy. As an AWS Certified Partner, Robots and Pencils combines the reach of a large firm with the focus of a trusted partner. With teams across the us, Canada, Europe and Latin America, clients gain local expertise and global scale as AI evolves. They ensure you keep peace with change and and that means faster results, measurable outcomes, and a partnership built to last. The right partner makes progress inevitable. Partner with Robots and pencils at robotsandpencils.com aidaily Brief Today's episode is brought to you by Superintelligent. Now for those of you who don't know who are new here, maybe superintelligent is actually my company. We started it because every single company we talk to, all the enterprises out there are trying to figure out what AI can do for them. But most of the advice is super generic, not specific to your company. So what we do is we map your AI and agent opportunities by deploying voice agents to interview your teams about how work works now and how your people would like it to work in the future. The result is an AI action map with high potential ROI use cases and specific change management needs. Basically everything you need to go actually deliver AI value. Go to BeSuper AI to learn more if you're building anything with Voice AI, you need to know about Assembly AI. They've built the best speech to text and speech understanding models in the industry. The quiet infrastructure behind products like Granola, Dovetail, Ashby and Clulee. Now, as I've said before, voice is one of the most important modalities of AI. It's the most natural human interface and I think it's a key part of where the next wave of innovation is going to happen. Assembly AI's models lead the field in accuracy and quality, so you can actually trust the data your product is built on. And their speech understanding models help you go beyond transcription, uncovering insights, identifying speakers and surfacing key moments automatically. It's developer first. No contracts, pay only for what you use and scales effortlessly. Go to semblyai.com brief, grab $50 in free credits and start building your voice AI product today. A couple times in the past week or so, AI music has been in the news and it's generated a lot of conversation around is using these tools and who is paying for these tools in ways that I think are perhaps non obvious. So let's talk about the news and then get into the broader discussion. The most recent story is that OpenAI is potentially getting into this space. The information reports that the company has begun working with students from the Juilliard School to annotate music scores for use in training. Data sources said that the design brief includes the ability to use AI to add elements to existing tracks, for example the ability to add a guitar to existing vocals. Product designers have apparently also discussed integrating a music model into the rest of the OpenAI ecosystem. It could be combined with Sora 2 to allow for greater control over the music that accompanies videos. Or OpenAI could also bring everything together into an advertising platform that could generate video assets with appropriate music. Now, even just in this little tiny article about all these different rumors, you have a bunch of wildly different use cases. You have the idea of AI generated music as a creative accompaniment for video and Sora 2. You have the idea, it sounds like, of something that would be in the digital audio workstation space if they really are interested in the idea of being able to take existing tracks and add elements to them. And then of course, you have the advertising dimensions where ad agencies could potentially use these tools to integrate music into their work in a different way. Now, so far there are very few details. OpenAI declined to comment and the sources didn't have any sense of timing. It just seems like something that they are exploring. Part of the reason that they might be exploring it though, is the success of Suno. In the middle of this month, Bloomberg reported that Suno was in talks to raise over $100 million at a valuation of more than 2 billion. The last time they raised money was in May of 2024 at a $500 million valuation. One overhang for them had been a big copyright case. In June of 24, Universal Music Group, Warner and others sued both Suno and its competitor Udio, claiming that those companies had trained on their copyrighted music, with the suit seeking damages of as much as $150,000 per work infringed, which would be obviously the end of those companies entirely. In June, however, it was reported that those labels were looking to settle the case. Bloomberg once again reported that the companies wanted to collect license fees for their work as well as receive a small amount of equity in the companies. Now this to me seemed like inevitably the end game that they were always careening towards. In the Internet era, the record labels were first to feel the pinch with the rise of Napster. However, they were also the first to develop their strategy of co optation and in many ways have been more successful in evolving with the medium of the moment than other industries have. Frankly, one of the reasons that Spotify is always looking for new business models outside of music is how much the record labels are still integrated into their business. Meaning, in other words, that the overhang around legal threats seems to be potentially on its way out. Meanwhile, just before news of the OpenAI story, the information also reported that Sunos revenue had quadrupled over the past year to reach 150 million in ARR. That puts SUNO in rarefied air. The information noted that only 20 other companies in their Gen AI startup database have reached that level. What's more, one of the sources said that Suno's margins are over 60% even if you include the cost of serving free customers. That's closer to a SaaS company than a gen AI startup, many of which are suffering significant margin compression as the sophistication of AI models advances and as paid users have to subsidize free users which create the onboarding for the paid users in the first place. We've talked about this in the context of AI coding startups in the past. The information again back in August reported that Replit had seen their gross profit margin drop to negative 14% earlier this year with a high point of 36%. Now Replit disputed the characterization a little bit, claiming that enterprise contracts carried 80 to 90% margin and that it was free users and power users who were driving the margin down. They've also stated that as they've moved to a more usage based model that has also helped margins. All of which is to say if Suno is really carrying around a 60% margin with 150 million in ARR, that makes them potentially one of the financially healthiest gen AI companies out there. Now the information noted that audio models tend to be smaller and cheaper than LLMs, which could go aways to explaining that healthy profit margin. But Suno isn't a particularly expensive Service. They charge 10 bucks a month for their mid tier subscription. That allows 500 song generations a month or $30 for a tier that quadruples that usage allowance and also opens up new features like Suno Studio. Current revenue then would imply around 5 million paying customers, which also would put them in the top tier of AI startups. CEO Mikey Shulman attributed the growth to increasingly advanced features. A year ago, Suno could only really work with basic prompts. You could ask for a country song about a down on his luck robot and get a reasonably interesting song out of the model. However, melodies and lyrics would tend to be a little repetitive and simplistic. Also, the fidelity to the genre that you were looking for was a little mixed. Users now have much more fine grained controls. You can upload your own lyrics. You can even hum a few bars of a song and get something much closer to a first draft of a professionally produced song. A little bit later in the show. I'm going to actually show you generations from the same prompt two full years apart. Just to get a sense of how things have evolved now in these stories, about 150 million in revenue came out. Some people were frankly gobsmacked, decipher co founder Michael Rosenfield wrote. Can someone explain where this revenue comes from? Who's paying? That post went viral, getting over 1.3 million views and hundreds and hundreds of responses. So let's talk about where this revenue is coming from. The first thing to note is that while there were some negative responses in the comments, Lovespurts on Twitter writes, I tried the new Suno and it's absolutely ruthless in its blandness, capable of turning any genre, no matter how eccentric, into generic drivel for the most undiscerning consumer hog. While there were one or two like this, by and large, this thread is about people who absolutely love this platform and some of those people are pretty well known. Harun, the founder and CEO of Rocket Money, as well as Andrew founder Palmer Lucky were among those who raised their hand and said they were users. Now I think that the obvious assumption for many is that much of this had to be a business use case, right? Michael Cove writes, it's mostly from content creators. It's cheaper than buying licensed soundtracks for just about anything. Shreya Nevatiya says, I use it to make my podcast intro music can see it being useful for anyone who needs non copyrighted music. Ilya Platinov writes, music licensing is a nightmare. It takes forever to find a good track, then you pay a bunch of money, then you cannot use it on YouTube without a strike. Now as someone who has used a lot of different music licensing services, the pain that he's describing is real. In the past, for example, when I've done a special series where I wanted a distinct music to differentiate it, I would spend literally hours browsing tracks to find exactly the right one. And so for people who have needs for commercial usage of music in whatever products they're creating, there are potentially some big upsides here. Vibe code founder and AI educator Riley Brown writes background music for YouTubers and short form creators ads with music with funny lyrics, and he also pointed out that some production studios are using it and that for a time people were mass uploading to Spotify to try to game the payout system. Although that use case is going away as they crack down, ads and digital content are certainly going to be a primary use case for apps like Suno. Every time you're creating a video ad, every time you're creating some specialized content marketing video, it potentially represents just such a better, cleaner, faster experience that I think that there's going to be a ton of that usage there. However, I don't think that that's primarily where this revenue is coming from right now. So who else is using it? It does seem like there are some traditional musicians who are starting to use the more advanced features. Adobe AI evangelist Chris Cashtanova writes, most of my friends subscribe to it. They are traditional musicians singers who found it useful to make their own music singing themselves but track with this tool. The last update lets using a lot of tracks and took it to another level. She's referring to Suno Studio, which they advertise as a complete creative workspace. This goes way beyond just a simple prompt. You can generate stems which are the specific tracks that come together to make a music. Think isolated vocals or isolated drums or isolated guitars and edit IT in the way that you could using another software like Ableton or Fruity Loops. You can also export everything and bring them into those digital audio workstations of choice. Gabbersoter writes, Suno is eating up the DAW market. Ableton FL Studio and Logic Pro and all the VST market. It's a thousand x faster and cheaper to go from an idea to a 90% version. Convexity on X writes I'm paying because it's damn good. I'm a highly experienced music producer. I'm fluent in DAWS plugins, MIDI controllers, effects, synth sample libraries, you name it. This AI stuff has a tremendous value proposition for even deeply experienced musicians. If the task is to write say four or five songs for a reality show, I can get this done in an hour or two of work as opposed to a couple of weeks minimum by traditional methods. The sound quality has gotten pretty darn good and now you can actually guide the result a lot more. You can whip up a demo and let the AI finish the song. That means it's doing performances in final arrangement, but it's your lyrics, your chord changes, your melody, and that's the strength of AI. It can do final polish complete arrangements in seconds. That would take a producer long days of work, if not weeks of work. You just can't afford to work the old way exclusively anymore. Snobs beware. So you're starting to see the glimmers of how professional musicians working musicians, especially those who supplement their income or primarily get their income with commercial production, are using this to improve how they work. And yet again, I don't think that's the big thing here. Justine Moore from A16Z writes what people don't realize about AI creative tools. For many users, making things with AI has become a hobby and or a form of entertainment. This is like asking why people may pay money to stream tv, go to a show, or join a soccer league. It's just fun. And I think more than anything else, that is what this 150 million in ARR represents. If you go to that thread from Rosenfield and look at what people are responding, it's all about personal use cases. For some people, it's about some very niche thing that they want to hear that they can't find otherwise or can't get enough of. Janik Meisner writes, I love Suno. There is nobody who makes songs about things I want to hear songs about. But now I can. New Pytorch version. There's a song now with the new features in the lyrics. Bryson Ran out of music from his favorite artist and so created a supplement until he could get more. He writes, I made an entire album of music that sounds like FKJ just because I ran out of new music. From his discography. George from Prodmanagement World says, making catchy songs for myself to remember Key ideas and affirmations Arkane's Valor writes, I'm paying Rolled up to a dinner party last night with a Suno song to some lyrics I wrote. It was a huge hit. Shocked at how good it's become. Also put a poem I wrote my wife on there and it was awesome. Justin Schroeder who doesn't pay It's a banger product. I use it just to make songs with my kids about our family. Connor Dempsey writes, suno is secretly my favorite AI product. My favorite use cases turning dumb inside jokes into a song for the group Chat birthday songs lyrics generated in ChatGPT with a few details about the person letting my 6 year old niece suggest a theme, have ChatGPT generate lyrics then having it generate a children's song. Manaj Nacom writes, I don't know anything about music but I purchased a yearly subscription after trying one month premium to create lullaby and songs for my toddler. It's damn good and over and over you see these types of personal experiences. Obi writes, My 74 year old dad writes original lyrics and pays for a SUNO subscription to make songs for my 97 year old grandma and also friends in his congregation. Jeanine Johnson writes that it's got classroom usage too. Teachers love Suno for creating custom learning songs. Honestly, the tone of most of this is incredulous that someone would even question why SUNO is valuable enough to pay for Jamji. Jamaramji writes, who's not paying for suno? In which age are you living? It's one of the most beautiful AI products out there. It's fun, it works and it produces amazing quality output. You can use it for tons of stuff for fun and business. It's super fun with family, friends and kids. Investor Amy Wu Martin put it in the context of social media more broadly. She writes in Social creation started democratized. What's my friend up to today? Turn Prosumer Pro creators make much better content and now becoming democratized again with AI, hobbyists can approach the quality of pro creators. The rule of thumb used to be that passive consumers of content outnumbered creators 10x to 1 on social networks including Roblox. That's changing. On some AI assisted UGC platforms, ratios approach 50% and for music as the first form factor to pass the Turing test 5050 on whether people can tell a song's been created on SUNA or not, it's no wonder an explosion of democratized creation has happened. Gregory Kennedy writes what everyone underestimates about AI creative tools and people just making stuff with them is that it's irrelevant if the output is objectively good. What matters is that the people making it believe they had a hand in creating it. AI creative tools lower the barriers to creation and increase the self satisfaction people get from making things. This is what many missed, even myself, about social media. I thought for sure it would never scale and people would tire of it. What I missed is that social media is always and forever about everyone's favorite topic themselves. All of this gets to the very heart of what it means to be human and what precisely is consciousness. We all experience life from our single perspective and have a running internal narrative that only we are aware of. I find it endlessly fascinating and so poorly understood. It drives so much of our lives. It strikes me that one of the things that we've been looking for is what native social media AI is going to open up. In other words, where AI creates some new type of creative experience that justifies a whole new network. When I talked about Sora too, basically my question was whether the idea of putting yourself and your friends and objects around you in AI generated videos was enough of a difference to justify a fork away from TikTok or YouTube or Instagram. It strikes me that there's a chance that music might be that. Up until now, the barriers to entry were way too high for music creation to actually be the core of a social experience. And for a while, even with AI generated music, the output quality wasn't high enough for it to be more than novelty. But now this entire world of creation is open where you can perfectly tune the lyrics or the style to whatever you want it to be. Music gets to be a part of any experience you want. Now that's everything from BSing with friends in the group chat to creating songs with your kids. Music registers super high in emotional impact relative to other media, and so being able to tune it more closely to yourself feels to me like it could be the type of thing that is not just a change in scale, but a change in kind. And here's what I think is interesting. While yes, there will be some segments of what is already commercial music that are likely to be disrupted by new forms of creation, we already heard about some of that. When it comes to the concern that AI generated music is going to overwhelm existing, human created music. It strikes me that what we're seeing is actually a fundamentally different use case for music opening up. AI generated music, at least at this stage, appears to me to not be competing, at least not over much with the music that people are putting on Spotify. And yes, of course, hold aside all the gamesmanship to try to get revenue from the Spotify algorithm. I'm talking about real, actual trying to be your next favorite musician. Instead, what we're seeing with AI music so far, and this 150 million in revenue that it represents, is an expansion of the total addressable market of music. It is a fundamentally different use case. It is not competing. And maybe this is why when you hear many musicians and producers talk about AI, they're not that scared. They understand that it is the quirks and weirdness and lived experience of musicians that differentiate great music from boring music. In other words, things that don't enter the training set for AI music. And maybe they intuitively get that having more people, being able to create more music is going to open up different ways to experience and interact with music. It's super interesting to me and I think something that's really worth keeping an eye on for now. I want to leave you with the difference that two years makes. Back in 2023, I entered the prompt nostalgic pop punk anthem about a 5 year old girl and a 2 year old boy at Christmas, full of traditional callbacks and glee. This was on Suno V2 and I'll play you a short clip of what came out. He's got his fingers on she's got a pingo running down the hallway Putting on a show because two little rebels Christmas time They're all about love It's Christmas time this is something that you could barely call novelty. At best, it showed the future where we might be headed. Compare that to a generation from Just this Morning with the same prompt, except their ages updated to 7 and 4. I'm going to let this play out. Appreciate you listening or watching as always and until next time. Peace.
- [29:35]
B
She tells him Santa knows their names he gasps like magic's in the flames Cookies crumble, Milk mustache pride Little feet in a sledding slide the world's so big but the tree feels small Their wonder grows it's the best of all Candy canes and paper snowflakes Little hands making big mistakes Tinsel tangled Lights too bright but it's perfect, perfect tonight Cake canes and paper snowflakes.