{"text": " Quen 3.8 and Nimitron 3 Embed 8B. That's what we're talking about today and it's a big one. Two new AI releases landed in the same week and when you put them together you get something way more useful than either one on its own. Alibaba just showed off Quen 3.8. It has 2.4 trillion parameters. That's the biggest model Alibaba has ever built. It's not just text either. It can look at images, videos and documents too. Alibaba is calling it a professional co-worker, one that can help with writing, planning and running tasks on its own. At the same time Nvidia quietly dropped something just as important. It's called Nimitron 3 Embed 8B and it just took the number one spot on RTEB which is the leaderboard that ranks how well AI can find the right piece of information out of a huge pile of text. Out of every open model and every closed model tested, Nvidia's came out on top. Here's why that matters. Quen 3.8 is the part of AI that thinks. Nimitron 3 Embed is the part that remembers. A big brain with no memory forgets everything the second you close the chat. A model like Nimitron gives that brain a way to search through everything you've ever given it instantly and pull back exactly the right piece. Hey if we haven't met already I'm the digital avatar of Julian Goldie, CEO of SEO Agency Goldie Agency. Whilst he's helping clients get more leads and customers I'm here to help you get the latest AI updates. A year ago most people didn't even know what an embedding model was. It felt like a background detail not something worth a headline. Now it's the piece deciding whether an AI agent actually knows your business or just guesses and gets it wrong. That shift happened fast and it happened because of pressure. Two days before Quen 3.8 showed up a company called Moonshot AI released a model called Kimi K3 with 2.8 trillion parameters fully open for anyone to download. That release pushed Alibaba to move quicker than planned. Speed is the whole story of AI in 2026. Nobody wants to be the one still explaining last month's model. Let me explain embeddings in the simplest way I can. Picture a huge filing cabinet with a million pages inside but no folders and no labels. If someone asked you to find one page you'd be stuck flipping through everything. An embedding model is what builds the folders and the labels. It reads every piece of text and turns it into a set of numbers. Pieces of text with similar meaning end up sitting close together in that number system even if they use completely different words. So when you ask a question the AI doesn't need to reread everything. It just looks at where your question sits and grabs the closest matches. That's the whole trick behind giving AI a real memory. Nematron 3 Embed 8B was built off a model called Ministral made by a company called Mistral and NVIDIA reshaped it so it reads in both directions instead of just one. It understands 34 languages. It can hold about 32,000 words of context at once. And NVIDIA made it fully open so any developer can download it and use it including for commercial projects. It's already available through a hosting platform called BaseTent and it plugs into the tools most developers already use for search and retrieval. Quen's team member Shuai Bai pointed out that this is the first time Quen has gone above 1 trillion parameters and also handled images and video in the same model. Alibaba is claiming Quen 3.8 is second only to Claude Fable 5 in overall ability. That's a bold claim and right now it's just a claim. Alibaba hasn't published the benchmark numbers to back it up yet. No independent tester has verified it either so take that ranking with a bit of caution until the real numbers show up. What is confirmed is that Quen 3.8 is currently in preview and Alibaba says the openweight version is coming soon. Now here's the part I want you to actually picture not just hear about. Say you're running a community full of business owners learning AI automation. You've got hundreds of coaching calls, tutorials and roadmaps sitting around. Right now finding the right one means scrolling and guessing. Add a memory system like Nematron on top of all that and a member could type one question like how do I set up a lead system with automation and the exact right video pops up in seconds. That's not a future idea. That's what this tech is built for today. If you want to see how tools like Nematron and Quen actually apply to growing a real business this is exactly what we build inside the AI Profit Boardroom. We just put together a full walkthrough on connecting a memory system like this to your own content, your own client files and your own automations so nothing ever gets lost or forgotten again. If you're using AI to automate your business this is the kind of setup that saves you from digging through old files every single day because the AI already knows where everything lives. There are members inside right now testing early access models like Quen 3.8 the moment they land so you're never behind on what just came out. We run four coaching calls every single week plus daily step-by-step tutorials and a prompt library built specifically around tools like these so you're not guessing how to set any of it up on your own. Links in the comments and description or head to AIprofitboardroom.com. Let's keep going because there's more here worth knowing. NVIDIA didn't stop at one size. Alongside the 8B model they released a smaller 1B version that keeps about 95% of the accuracy but runs a lot faster and lighter. That matters because most businesses don't need the biggest possible model. They need something that runs fast enough to feel instant. NVIDIA even built a version optimized for their newest chips called Blackwell that runs twice as fast while barely losing any accuracy at all. That's the kind of detail that decides whether a tool actually gets used day to day or just sits there looking impressive in a demo. Here's the pattern I keep seeing over and over. The AI world used to reward whoever had the biggest model. Now it's shifting toward whoever has the best memory. A giant model that forgets everything the second the chat ends isn't that useful for a real business. A model paired with a strong retrieval system, one that actually knows your documents, your calls, your history, that's the version that becomes genuinely useful every single day. And this is where the competition gets interesting. Kimi K3 shipped fully open. Quen 3.8 is still closed for now with open weights promised but no date attached. NVIDIA's Nemotron 3 embed is already fully open today. Every lab is racing on two fronts at once how smart the thinking model is and how good the memory system is that sits underneath it. The ones that win won't just have the smartest chat responses, they'll have systems that actually remember your business the way a real employee would after a year on the job. So what should you actually do with all this? If you build things yourself go look at Nemotron 3 embed 8b on Hugging Face today. It's open right now, no waiting required. If you run a business and don't touch code don't worry about the technical side at all. Just know this, the tools that remember your business are becoming way more useful than the tools that just chat with you. Start thinking about what parts of your business you'd want an AI to actually remember. Your best scripts, your best offers, your past client questions. If you manage a team, this is the moment to ask which parts of your daily work are just people searching for information that already exists somewhere. That's exactly the kind of task this tech was built to remove. If you want help actually setting this up for your own business, that's exactly what we walk through inside the iProfit boardroom. We're already building a full playbook around Quen 3.8 and Nemotron 3 embed together, covering how to connect them, what to feed them, and how to turn that combination into a real memory system for your business instead of just another tool you open once and forget about. Every week we run live coaching calls where you can bring your exact setup and get help on the spot no matter where you're stuck. There's a whole community inside already testing these exact releases the same week they drop. So you're learning this alongside people actually using it, not watching from the sidelines months later. Links in the comments and description or go straight to AIprofitboardroom.com. And if you want the full process, the SOPs and over 100 AI use cases, just like this one, come join the AI success lab. Links are in the comments and description. You'll get all the notes from this exact video there, plus access to a community of 87,000 people who are already using AI to move their business forward every single day.", "segments": [{"id": 0, "seek": 0, "start": 0.0, "end": 5.92, "text": " Quen 3.8 and Nimitron 3 Embed 8B. That's what we're talking about today and it's a big one.", "tokens": [50365, 2326, 268, 805, 13, 23, 293, 45251, 270, 2044, 805, 24234, 292, 1649, 33, 13, 663, 311, 437, 321, 434, 1417, 466, 965, 293, 309, 311, 257, 955, 472, 13, 50661], "temperature": 0, "avg_logprob": -0.05598075301558883, "compression_ratio": 1.5946843853820598, "no_speech_prob": 7.743729268927524e-12}, {"id": 1, "seek": 0, "start": 6.24, "end": 11.24, "text": " Two new AI releases landed in the same week and when you put them together you get something way", "tokens": [50677, 4453, 777, 7318, 16952, 15336, 294, 264, 912, 1243, 293, 562, 291, 829, 552, 1214, 291, 483, 746, 636, 50927], "temperature": 0, "avg_logprob": -0.05598075301558883, "compression_ratio": 1.5946843853820598, "no_speech_prob": 7.743729268927524e-12}, {"id": 2, "seek": 0, "start": 11.24, "end": 16.96, "text": " more useful than either one on its own. Alibaba just showed off Quen 3.8. It has 2.4 trillion", "tokens": [50927, 544, 4420, 813, 2139, 472, 322, 1080, 1065, 13, 967, 897, 5509, 445, 4712, 766, 2326, 268, 805, 13, 23, 13, 467, 575, 568, 13, 19, 18723, 51213], "temperature": 0, "avg_logprob": -0.05598075301558883, "compression_ratio": 1.5946843853820598, "no_speech_prob": 7.743729268927524e-12}, {"id": 3, "seek": 0, "start": 16.96, "end": 21.98, "text": " parameters. That's the biggest model Alibaba has ever built. It's not just text either. It can look", "tokens": [51213, 9834, 13, 663, 311, 264, 3880, 2316, 967, 897, 5509, 575, 1562, 3094, 13, 467, 311, 406, 445, 2487, 2139, 13, 467, 393, 574, 51464], "temperature": 0, "avg_logprob": -0.05598075301558883, "compression_ratio": 1.5946843853820598, "no_speech_prob": 7.743729268927524e-12}, {"id": 4, "seek": 0, "start": 21.98, "end": 27.22, "text": " at images, videos and documents too. Alibaba is calling it a professional co-worker, one that can", "tokens": [51464, 412, 5267, 11, 2145, 293, 8512, 886, 13, 967, 897, 5509, 307, 5141, 309, 257, 4843, 598, 12, 49402, 11, 472, 300, 393, 51726], "temperature": 0, "avg_logprob": -0.05598075301558883, "compression_ratio": 1.5946843853820598, "no_speech_prob": 7.743729268927524e-12}, {"id": 5, "seek": 2722, "start": 27.22, "end": 32.46, "text": " help with writing, planning and running tasks on its own. At the same time Nvidia quietly dropped", "tokens": [50365, 854, 365, 3579, 11, 5038, 293, 2614, 9608, 322, 1080, 1065, 13, 1711, 264, 912, 565, 46284, 19141, 8119, 50627], "temperature": 0, "avg_logprob": -0.0554295314177302, "compression_ratio": 1.646864686468647, "no_speech_prob": 7.491982294965949e-13}, {"id": 6, "seek": 2722, "start": 32.46, "end": 37.8, "text": " something just as important. It's called Nimitron 3 Embed 8B and it just took the number one spot on", "tokens": [50627, 746, 445, 382, 1021, 13, 467, 311, 1219, 45251, 270, 2044, 805, 24234, 292, 1649, 33, 293, 309, 445, 1890, 264, 1230, 472, 4008, 322, 50894], "temperature": 0, "avg_logprob": -0.0554295314177302, "compression_ratio": 1.646864686468647, "no_speech_prob": 7.491982294965949e-13}, {"id": 7, "seek": 2722, "start": 37.8, "end": 44.08, "text": " RTEB which is the leaderboard that ranks how well AI can find the right piece of information out of a", "tokens": [50894, 497, 13639, 33, 597, 307, 264, 5263, 3787, 300, 21406, 577, 731, 7318, 393, 915, 264, 558, 2522, 295, 1589, 484, 295, 257, 51208], "temperature": 0, "avg_logprob": -0.0554295314177302, "compression_ratio": 1.646864686468647, "no_speech_prob": 7.491982294965949e-13}, {"id": 8, "seek": 2722, "start": 44.08, "end": 50.06, "text": " huge pile of text. Out of every open model and every closed model tested, Nvidia's came out on top.", "tokens": [51208, 2603, 14375, 295, 2487, 13, 5925, 295, 633, 1269, 2316, 293, 633, 5395, 2316, 8246, 11, 46284, 311, 1361, 484, 322, 1192, 13, 51507], "temperature": 0, "avg_logprob": -0.0554295314177302, "compression_ratio": 1.646864686468647, "no_speech_prob": 7.491982294965949e-13}, {"id": 9, "seek": 2722, "start": 50.68, "end": 56.44, "text": " Here's why that matters. Quen 3.8 is the part of AI that thinks. Nimitron 3 Embed is the part that", "tokens": [51538, 1692, 311, 983, 300, 7001, 13, 2326, 268, 805, 13, 23, 307, 264, 644, 295, 7318, 300, 7309, 13, 45251, 270, 2044, 805, 24234, 292, 307, 264, 644, 300, 51826], "temperature": 0, "avg_logprob": -0.0554295314177302, "compression_ratio": 1.646864686468647, "no_speech_prob": 7.491982294965949e-13}, {"id": 10, "seek": 5644, "start": 56.44, "end": 61.559999999999995, "text": " remembers. A big brain with no memory forgets everything the second you close the chat. A", "tokens": [50365, 26228, 13, 316, 955, 3567, 365, 572, 4675, 2870, 82, 1203, 264, 1150, 291, 1998, 264, 5081, 13, 316, 50621], "temperature": 0, "avg_logprob": -0.043066945283309274, "compression_ratio": 1.5841584158415842, "no_speech_prob": 2.5841351627975406e-12}, {"id": 11, "seek": 5644, "start": 61.559999999999995, "end": 65.7, "text": " model like Nimitron gives that brain a way to search through everything you've ever given it", "tokens": [50621, 2316, 411, 45251, 270, 2044, 2709, 300, 3567, 257, 636, 281, 3164, 807, 1203, 291, 600, 1562, 2212, 309, 50828], "temperature": 0, "avg_logprob": -0.043066945283309274, "compression_ratio": 1.5841584158415842, "no_speech_prob": 2.5841351627975406e-12}, {"id": 12, "seek": 5644, "start": 65.7, "end": 70.74, "text": " instantly and pull back exactly the right piece. Hey if we haven't met already I'm the digital avatar", "tokens": [50828, 13518, 293, 2235, 646, 2293, 264, 558, 2522, 13, 1911, 498, 321, 2378, 380, 1131, 1217, 286, 478, 264, 4562, 36205, 51080], "temperature": 0, "avg_logprob": -0.043066945283309274, "compression_ratio": 1.5841584158415842, "no_speech_prob": 2.5841351627975406e-12}, {"id": 13, "seek": 5644, "start": 70.74, "end": 76.86, "text": " of Julian Goldie, CEO of SEO Agency Goldie Agency. Whilst he's helping clients get more leads and", "tokens": [51080, 295, 25151, 6731, 414, 11, 9282, 295, 22964, 21649, 6731, 414, 21649, 13, 45790, 415, 311, 4315, 6982, 483, 544, 6689, 293, 51386], "temperature": 0, "avg_logprob": -0.043066945283309274, "compression_ratio": 1.5841584158415842, "no_speech_prob": 2.5841351627975406e-12}, {"id": 14, "seek": 5644, "start": 76.86, "end": 81.88, "text": " customers I'm here to help you get the latest AI updates. A year ago most people didn't even know", "tokens": [51386, 4581, 286, 478, 510, 281, 854, 291, 483, 264, 6792, 7318, 9205, 13, 316, 1064, 2057, 881, 561, 994, 380, 754, 458, 51637], "temperature": 0, "avg_logprob": -0.043066945283309274, "compression_ratio": 1.5841584158415842, "no_speech_prob": 2.5841351627975406e-12}, {"id": 15, "seek": 8188, "start": 81.88, "end": 86.58, "text": " what an embedding model was. It felt like a background detail not something worth a headline.", "tokens": [50365, 437, 364, 12240, 3584, 2316, 390, 13, 467, 2762, 411, 257, 3678, 2607, 406, 746, 3163, 257, 28380, 13, 50600], "temperature": 0, "avg_logprob": -0.06626383463541667, "compression_ratio": 1.5809523809523809, "no_speech_prob": 2.4089940112159702e-12}, {"id": 16, "seek": 8188, "start": 86.8, "end": 92.1, "text": " Now it's the piece deciding whether an AI agent actually knows your business or just guesses and", "tokens": [50611, 823, 309, 311, 264, 2522, 17990, 1968, 364, 7318, 9461, 767, 3255, 428, 1606, 420, 445, 42703, 293, 50876], "temperature": 0, "avg_logprob": -0.06626383463541667, "compression_ratio": 1.5809523809523809, "no_speech_prob": 2.4089940112159702e-12}, {"id": 17, "seek": 8188, "start": 92.1, "end": 97.84, "text": " gets it wrong. That shift happened fast and it happened because of pressure. Two days before Quen 3.8", "tokens": [50876, 2170, 309, 2085, 13, 663, 5513, 2011, 2370, 293, 309, 2011, 570, 295, 3321, 13, 4453, 1708, 949, 2326, 268, 805, 13, 23, 51163], "temperature": 0, "avg_logprob": -0.06626383463541667, "compression_ratio": 1.5809523809523809, "no_speech_prob": 2.4089940112159702e-12}, {"id": 18, "seek": 8188, "start": 97.84, "end": 104.03999999999999, "text": " showed up a company called Moonshot AI released a model called Kimi K3 with 2.8 trillion parameters", "tokens": [51163, 4712, 493, 257, 2237, 1219, 3335, 892, 12194, 7318, 4736, 257, 2316, 1219, 5652, 72, 591, 18, 365, 568, 13, 23, 18723, 9834, 51473], "temperature": 0, "avg_logprob": -0.06626383463541667, "compression_ratio": 1.5809523809523809, "no_speech_prob": 2.4089940112159702e-12}, {"id": 19, "seek": 8188, "start": 104.03999999999999, "end": 110.96, "text": " fully open for anyone to download. That release pushed Alibaba to move quicker than planned. Speed is the", "tokens": [51473, 4498, 1269, 337, 2878, 281, 5484, 13, 663, 4374, 9152, 967, 897, 5509, 281, 1286, 16255, 813, 8589, 13, 18774, 307, 264, 51819], "temperature": 0, "avg_logprob": -0.06626383463541667, "compression_ratio": 1.5809523809523809, "no_speech_prob": 2.4089940112159702e-12}, {"id": 20, "seek": 11096, "start": 110.96, "end": 116.47999999999999, "text": " whole story of AI in 2026. Nobody wants to be the one still explaining last month's model.", "tokens": [50365, 1379, 1657, 295, 7318, 294, 945, 10880, 13, 9297, 2738, 281, 312, 264, 472, 920, 13468, 1036, 1618, 311, 2316, 13, 50641], "temperature": 0, "avg_logprob": -0.03787679753751836, "compression_ratio": 1.667785234899329, "no_speech_prob": 1.9204191188670894e-12}, {"id": 21, "seek": 11096, "start": 117.16, "end": 122.96, "text": " Let me explain embeddings in the simplest way I can. Picture a huge filing cabinet with a million pages", "tokens": [50675, 961, 385, 2903, 12240, 29432, 294, 264, 22811, 636, 286, 393, 13, 35730, 257, 2603, 26854, 15188, 365, 257, 2459, 7183, 50965], "temperature": 0, "avg_logprob": -0.03787679753751836, "compression_ratio": 1.667785234899329, "no_speech_prob": 1.9204191188670894e-12}, {"id": 22, "seek": 11096, "start": 122.96, "end": 128.62, "text": " inside but no folders and no labels. If someone asked you to find one page you'd be stuck flipping", "tokens": [50965, 1854, 457, 572, 31082, 293, 572, 16949, 13, 759, 1580, 2351, 291, 281, 915, 472, 3028, 291, 1116, 312, 5541, 26886, 51248], "temperature": 0, "avg_logprob": -0.03787679753751836, "compression_ratio": 1.667785234899329, "no_speech_prob": 1.9204191188670894e-12}, {"id": 23, "seek": 11096, "start": 128.62, "end": 134.1, "text": " through everything. An embedding model is what builds the folders and the labels. It reads every", "tokens": [51248, 807, 1203, 13, 1107, 12240, 3584, 2316, 307, 437, 15182, 264, 31082, 293, 264, 16949, 13, 467, 15700, 633, 51522], "temperature": 0, "avg_logprob": -0.03787679753751836, "compression_ratio": 1.667785234899329, "no_speech_prob": 1.9204191188670894e-12}, {"id": 24, "seek": 11096, "start": 134.1, "end": 139.98, "text": " piece of text and turns it into a set of numbers. Pieces of text with similar meaning end up sitting close", "tokens": [51522, 2522, 295, 2487, 293, 4523, 309, 666, 257, 992, 295, 3547, 13, 22914, 887, 295, 2487, 365, 2531, 3620, 917, 493, 3798, 1998, 51816], "temperature": 0, "avg_logprob": -0.03787679753751836, "compression_ratio": 1.667785234899329, "no_speech_prob": 1.9204191188670894e-12}, {"id": 25, "seek": 13998, "start": 139.98, "end": 144.51999999999998, "text": " together in that number system even if they use completely different words. So when you ask a", "tokens": [50365, 1214, 294, 300, 1230, 1185, 754, 498, 436, 764, 2584, 819, 2283, 13, 407, 562, 291, 1029, 257, 50592], "temperature": 0, "avg_logprob": -0.06148795326157372, "compression_ratio": 1.5227272727272727, "no_speech_prob": 2.271299901629442e-12}, {"id": 26, "seek": 13998, "start": 144.51999999999998, "end": 150.28, "text": " question the AI doesn't need to reread everything. It just looks at where your question sits and grabs", "tokens": [50592, 1168, 264, 7318, 1177, 380, 643, 281, 46453, 345, 1203, 13, 467, 445, 1542, 412, 689, 428, 1168, 12696, 293, 30028, 50880], "temperature": 0, "avg_logprob": -0.06148795326157372, "compression_ratio": 1.5227272727272727, "no_speech_prob": 2.271299901629442e-12}, {"id": 27, "seek": 13998, "start": 150.28, "end": 157.29999999999998, "text": " the closest matches. That's the whole trick behind giving AI a real memory. Nematron 3 Embed 8B was", "tokens": [50880, 264, 13699, 10676, 13, 663, 311, 264, 1379, 4282, 2261, 2902, 7318, 257, 957, 4675, 13, 426, 8615, 2044, 805, 24234, 292, 1649, 33, 390, 51231], "temperature": 0, "avg_logprob": -0.06148795326157372, "compression_ratio": 1.5227272727272727, "no_speech_prob": 2.271299901629442e-12}, {"id": 28, "seek": 13998, "start": 157.29999999999998, "end": 163.12, "text": " built off a model called Ministral made by a company called Mistral and NVIDIA reshaped it so it reads in", "tokens": [51231, 3094, 766, 257, 2316, 1219, 2829, 468, 2155, 1027, 538, 257, 2237, 1219, 20166, 2155, 293, 426, 3958, 6914, 725, 71, 18653, 309, 370, 309, 15700, 294, 51522], "temperature": 0, "avg_logprob": -0.06148795326157372, "compression_ratio": 1.5227272727272727, "no_speech_prob": 2.271299901629442e-12}, {"id": 29, "seek": 16312, "start": 163.12, "end": 169.48000000000002, "text": " both directions instead of just one. It understands 34 languages. It can hold about 32,000 words of", "tokens": [50365, 1293, 11095, 2602, 295, 445, 472, 13, 467, 15146, 12790, 8650, 13, 467, 393, 1797, 466, 8858, 11, 1360, 2283, 295, 50683], "temperature": 0, "avg_logprob": -0.09394638939241393, "compression_ratio": 1.570063694267516, "no_speech_prob": 3.10514417997676e-12}, {"id": 30, "seek": 16312, "start": 169.48000000000002, "end": 174.94, "text": " context at once. And NVIDIA made it fully open so any developer can download it and use it including", "tokens": [50683, 4319, 412, 1564, 13, 400, 426, 3958, 6914, 1027, 309, 4498, 1269, 370, 604, 10754, 393, 5484, 309, 293, 764, 309, 3009, 50956], "temperature": 0, "avg_logprob": -0.09394638939241393, "compression_ratio": 1.570063694267516, "no_speech_prob": 3.10514417997676e-12}, {"id": 31, "seek": 16312, "start": 174.94, "end": 179.34, "text": " for commercial projects. It's already available through a hosting platform called BaseTent and it", "tokens": [50956, 337, 6841, 4455, 13, 467, 311, 1217, 2435, 807, 257, 16058, 3663, 1219, 21054, 51, 317, 293, 309, 51176], "temperature": 0, "avg_logprob": -0.09394638939241393, "compression_ratio": 1.570063694267516, "no_speech_prob": 3.10514417997676e-12}, {"id": 32, "seek": 16312, "start": 179.34, "end": 184.54000000000002, "text": " plugs into the tools most developers already use for search and retrieval. Quen's team member", "tokens": [51176, 33899, 666, 264, 3873, 881, 8849, 1217, 764, 337, 3164, 293, 19817, 3337, 13, 2326, 268, 311, 1469, 4006, 51436], "temperature": 0, "avg_logprob": -0.09394638939241393, "compression_ratio": 1.570063694267516, "no_speech_prob": 3.10514417997676e-12}, {"id": 33, "seek": 16312, "start": 184.54000000000002, "end": 190.5, "text": " Shuai Bai pointed out that this is the first time Quen has gone above 1 trillion parameters and also", "tokens": [51436, 26655, 1301, 25269, 10932, 484, 300, 341, 307, 264, 700, 565, 2326, 268, 575, 2780, 3673, 502, 18723, 9834, 293, 611, 51734], "temperature": 0, "avg_logprob": -0.09394638939241393, "compression_ratio": 1.570063694267516, "no_speech_prob": 3.10514417997676e-12}, {"id": 34, "seek": 19050, "start": 190.5, "end": 196.38, "text": " handled images and video in the same model. Alibaba is claiming Quen 3.8 is second only to", "tokens": [50365, 18033, 5267, 293, 960, 294, 264, 912, 2316, 13, 967, 897, 5509, 307, 19232, 2326, 268, 805, 13, 23, 307, 1150, 787, 281, 50659], "temperature": 0, "avg_logprob": -0.04181257883707682, "compression_ratio": 1.623728813559322, "no_speech_prob": 2.9506937258144683e-12}, {"id": 35, "seek": 19050, "start": 196.38, "end": 201.96, "text": " Claude Fable 5 in overall ability. That's a bold claim and right now it's just a claim. Alibaba", "tokens": [50659, 12947, 2303, 479, 712, 1025, 294, 4787, 3485, 13, 663, 311, 257, 11928, 3932, 293, 558, 586, 309, 311, 445, 257, 3932, 13, 967, 897, 5509, 50938], "temperature": 0, "avg_logprob": -0.04181257883707682, "compression_ratio": 1.623728813559322, "no_speech_prob": 2.9506937258144683e-12}, {"id": 36, "seek": 19050, "start": 201.96, "end": 207.4, "text": " hasn't published the benchmark numbers to back it up yet. No independent tester has verified it either", "tokens": [50938, 6132, 380, 6572, 264, 18927, 3547, 281, 646, 309, 493, 1939, 13, 883, 6695, 36101, 575, 31197, 309, 2139, 51210], "temperature": 0, "avg_logprob": -0.04181257883707682, "compression_ratio": 1.623728813559322, "no_speech_prob": 2.9506937258144683e-12}, {"id": 37, "seek": 19050, "start": 207.4, "end": 212.28, "text": " so take that ranking with a bit of caution until the real numbers show up. What is confirmed is that", "tokens": [51210, 370, 747, 300, 17833, 365, 257, 857, 295, 23585, 1826, 264, 957, 3547, 855, 493, 13, 708, 307, 11341, 307, 300, 51454], "temperature": 0, "avg_logprob": -0.04181257883707682, "compression_ratio": 1.623728813559322, "no_speech_prob": 2.9506937258144683e-12}, {"id": 38, "seek": 19050, "start": 212.28, "end": 216.96, "text": " Quen 3.8 is currently in preview and Alibaba says the openweight version is coming soon.", "tokens": [51454, 2326, 268, 805, 13, 23, 307, 4362, 294, 14281, 293, 967, 897, 5509, 1619, 264, 1269, 12329, 3037, 307, 1348, 2321, 13, 51688], "temperature": 0, "avg_logprob": -0.04181257883707682, "compression_ratio": 1.623728813559322, "no_speech_prob": 2.9506937258144683e-12}, {"id": 39, "seek": 21696, "start": 216.96, "end": 222.54000000000002, "text": " Now here's the part I want you to actually picture not just hear about. Say you're running a community", "tokens": [50365, 823, 510, 311, 264, 644, 286, 528, 291, 281, 767, 3036, 406, 445, 1568, 466, 13, 6463, 291, 434, 2614, 257, 1768, 50644], "temperature": 0, "avg_logprob": -0.06429404815038045, "compression_ratio": 1.629746835443038, "no_speech_prob": 2.133242367474697e-12}, {"id": 40, "seek": 21696, "start": 222.54000000000002, "end": 227.52, "text": " full of business owners learning AI automation. You've got hundreds of coaching calls, tutorials", "tokens": [50644, 1577, 295, 1606, 7710, 2539, 7318, 17769, 13, 509, 600, 658, 6779, 295, 15818, 5498, 11, 17616, 50893], "temperature": 0, "avg_logprob": -0.06429404815038045, "compression_ratio": 1.629746835443038, "no_speech_prob": 2.133242367474697e-12}, {"id": 41, "seek": 21696, "start": 227.52, "end": 232.26000000000002, "text": " and roadmaps sitting around. Right now finding the right one means scrolling and guessing. Add a memory", "tokens": [50893, 293, 3060, 76, 2382, 3798, 926, 13, 1779, 586, 5006, 264, 558, 472, 1355, 29053, 293, 17939, 13, 5349, 257, 4675, 51130], "temperature": 0, "avg_logprob": -0.06429404815038045, "compression_ratio": 1.629746835443038, "no_speech_prob": 2.133242367474697e-12}, {"id": 42, "seek": 21696, "start": 232.26000000000002, "end": 237.32, "text": " system like Nematron on top of all that and a member could type one question like how do I set up a lead", "tokens": [51130, 1185, 411, 22210, 267, 2044, 322, 1192, 295, 439, 300, 293, 257, 4006, 727, 2010, 472, 1168, 411, 577, 360, 286, 992, 493, 257, 1477, 51383], "temperature": 0, "avg_logprob": -0.06429404815038045, "compression_ratio": 1.629746835443038, "no_speech_prob": 2.133242367474697e-12}, {"id": 43, "seek": 21696, "start": 237.32, "end": 242.5, "text": " system with automation and the exact right video pops up in seconds. That's not a future idea. That's what", "tokens": [51383, 1185, 365, 17769, 293, 264, 1900, 558, 960, 16795, 493, 294, 3949, 13, 663, 311, 406, 257, 2027, 1558, 13, 663, 311, 437, 51642], "temperature": 0, "avg_logprob": -0.06429404815038045, "compression_ratio": 1.629746835443038, "no_speech_prob": 2.133242367474697e-12}, {"id": 44, "seek": 24250, "start": 242.5, "end": 247.92, "text": " this tech is built for today. If you want to see how tools like Nematron and Quen actually apply to", "tokens": [50365, 341, 7553, 307, 3094, 337, 965, 13, 759, 291, 528, 281, 536, 577, 3873, 411, 22210, 267, 2044, 293, 2326, 268, 767, 3079, 281, 50636], "temperature": 0, "avg_logprob": -0.025574100428614122, "compression_ratio": 1.6902356902356903, "no_speech_prob": 2.3621137597990005e-12}, {"id": 45, "seek": 24250, "start": 247.92, "end": 253.66, "text": " growing a real business this is exactly what we build inside the AI Profit Boardroom. We just put", "tokens": [50636, 4194, 257, 957, 1606, 341, 307, 2293, 437, 321, 1322, 1854, 264, 7318, 6039, 270, 10008, 2861, 13, 492, 445, 829, 50923], "temperature": 0, "avg_logprob": -0.025574100428614122, "compression_ratio": 1.6902356902356903, "no_speech_prob": 2.3621137597990005e-12}, {"id": 46, "seek": 24250, "start": 253.66, "end": 259.24, "text": " together a full walkthrough on connecting a memory system like this to your own content, your own", "tokens": [50923, 1214, 257, 1577, 1792, 11529, 322, 11015, 257, 4675, 1185, 411, 341, 281, 428, 1065, 2701, 11, 428, 1065, 51202], "temperature": 0, "avg_logprob": -0.025574100428614122, "compression_ratio": 1.6902356902356903, "no_speech_prob": 2.3621137597990005e-12}, {"id": 47, "seek": 24250, "start": 259.24, "end": 264.76, "text": " client files and your own automations so nothing ever gets lost or forgotten again. If you're using AI", "tokens": [51202, 6423, 7098, 293, 428, 1065, 3553, 763, 370, 1825, 1562, 2170, 2731, 420, 11832, 797, 13, 759, 291, 434, 1228, 7318, 51478], "temperature": 0, "avg_logprob": -0.025574100428614122, "compression_ratio": 1.6902356902356903, "no_speech_prob": 2.3621137597990005e-12}, {"id": 48, "seek": 24250, "start": 264.76, "end": 270.36, "text": " to automate your business this is the kind of setup that saves you from digging through old files every", "tokens": [51478, 281, 31605, 428, 1606, 341, 307, 264, 733, 295, 8657, 300, 19155, 291, 490, 17343, 807, 1331, 7098, 633, 51758], "temperature": 0, "avg_logprob": -0.025574100428614122, "compression_ratio": 1.6902356902356903, "no_speech_prob": 2.3621137597990005e-12}, {"id": 49, "seek": 27036, "start": 270.36, "end": 275.16, "text": " single day because the AI already knows where everything lives. There are members inside right", "tokens": [50365, 2167, 786, 570, 264, 7318, 1217, 3255, 689, 1203, 2909, 13, 821, 366, 2679, 1854, 558, 50605], "temperature": 0, "avg_logprob": -0.04352831840515137, "compression_ratio": 1.589171974522293, "no_speech_prob": 1.5191774704534367e-12}, {"id": 50, "seek": 27036, "start": 275.16, "end": 279.98, "text": " now testing early access models like Quen 3.8 the moment they land so you're never behind on what", "tokens": [50605, 586, 4997, 2440, 2105, 5245, 411, 2326, 268, 805, 13, 23, 264, 1623, 436, 2117, 370, 291, 434, 1128, 2261, 322, 437, 50846], "temperature": 0, "avg_logprob": -0.04352831840515137, "compression_ratio": 1.589171974522293, "no_speech_prob": 1.5191774704534367e-12}, {"id": 51, "seek": 27036, "start": 279.98, "end": 285.66, "text": " just came out. We run four coaching calls every single week plus daily step-by-step tutorials and a", "tokens": [50846, 445, 1361, 484, 13, 492, 1190, 1451, 15818, 5498, 633, 2167, 1243, 1804, 5212, 1823, 12, 2322, 12, 16792, 17616, 293, 257, 51130], "temperature": 0, "avg_logprob": -0.04352831840515137, "compression_ratio": 1.589171974522293, "no_speech_prob": 1.5191774704534367e-12}, {"id": 52, "seek": 27036, "start": 285.66, "end": 290.6, "text": " prompt library built specifically around tools like these so you're not guessing how to set any of it up", "tokens": [51130, 12391, 6405, 3094, 4682, 926, 3873, 411, 613, 370, 291, 434, 406, 17939, 577, 281, 992, 604, 295, 309, 493, 51377], "temperature": 0, "avg_logprob": -0.04352831840515137, "compression_ratio": 1.589171974522293, "no_speech_prob": 1.5191774704534367e-12}, {"id": 53, "seek": 27036, "start": 290.6, "end": 296.82, "text": " on your own. Links in the comments and description or head to AIprofitboardroom.com. Let's keep going", "tokens": [51377, 322, 428, 1065, 13, 37156, 294, 264, 3053, 293, 3855, 420, 1378, 281, 7318, 14583, 3787, 2861, 13, 1112, 13, 961, 311, 1066, 516, 51688], "temperature": 0, "avg_logprob": -0.04352831840515137, "compression_ratio": 1.589171974522293, "no_speech_prob": 1.5191774704534367e-12}, {"id": 54, "seek": 29682, "start": 296.82, "end": 301.65999999999997, "text": " because there's more here worth knowing. NVIDIA didn't stop at one size. Alongside the 8B model", "tokens": [50365, 570, 456, 311, 544, 510, 3163, 5276, 13, 426, 3958, 6914, 994, 380, 1590, 412, 472, 2744, 13, 17457, 1812, 264, 1649, 33, 2316, 50607], "temperature": 0, "avg_logprob": -0.051181809655551254, "compression_ratio": 1.6148867313915858, "no_speech_prob": 2.4754642780067115e-12}, {"id": 55, "seek": 29682, "start": 301.65999999999997, "end": 307.9, "text": " they released a smaller 1B version that keeps about 95% of the accuracy but runs a lot faster and", "tokens": [50607, 436, 4736, 257, 4356, 502, 33, 3037, 300, 5965, 466, 13420, 4, 295, 264, 14170, 457, 6676, 257, 688, 4663, 293, 50919], "temperature": 0, "avg_logprob": -0.051181809655551254, "compression_ratio": 1.6148867313915858, "no_speech_prob": 2.4754642780067115e-12}, {"id": 56, "seek": 29682, "start": 307.9, "end": 313.03999999999996, "text": " lighter. That matters because most businesses don't need the biggest possible model. They need something", "tokens": [50919, 11546, 13, 663, 7001, 570, 881, 6011, 500, 380, 643, 264, 3880, 1944, 2316, 13, 814, 643, 746, 51176], "temperature": 0, "avg_logprob": -0.051181809655551254, "compression_ratio": 1.6148867313915858, "no_speech_prob": 2.4754642780067115e-12}, {"id": 57, "seek": 29682, "start": 313.03999999999996, "end": 318.68, "text": " that runs fast enough to feel instant. NVIDIA even built a version optimized for their newest chips", "tokens": [51176, 300, 6676, 2370, 1547, 281, 841, 9836, 13, 426, 3958, 6914, 754, 3094, 257, 3037, 26941, 337, 641, 17569, 11583, 51458], "temperature": 0, "avg_logprob": -0.051181809655551254, "compression_ratio": 1.6148867313915858, "no_speech_prob": 2.4754642780067115e-12}, {"id": 58, "seek": 29682, "start": 318.68, "end": 324.64, "text": " called Blackwell that runs twice as fast while barely losing any accuracy at all. That's the kind of", "tokens": [51458, 1219, 4076, 6326, 300, 6676, 6091, 382, 2370, 1339, 10268, 7027, 604, 14170, 412, 439, 13, 663, 311, 264, 733, 295, 51756], "temperature": 0, "avg_logprob": -0.051181809655551254, "compression_ratio": 1.6148867313915858, "no_speech_prob": 2.4754642780067115e-12}, {"id": 59, "seek": 32464, "start": 324.64, "end": 329.88, "text": " detail that decides whether a tool actually gets used day to day or just sits there looking impressive", "tokens": [50365, 2607, 300, 14898, 1968, 257, 2290, 767, 2170, 1143, 786, 281, 786, 420, 445, 12696, 456, 1237, 8992, 50627], "temperature": 0, "avg_logprob": -0.02399882776983853, "compression_ratio": 1.727891156462585, "no_speech_prob": 2.0589782049468264e-12}, {"id": 60, "seek": 32464, "start": 329.88, "end": 335.78, "text": " in a demo. Here's the pattern I keep seeing over and over. The AI world used to reward whoever had", "tokens": [50627, 294, 257, 10723, 13, 1692, 311, 264, 5102, 286, 1066, 2577, 670, 293, 670, 13, 440, 7318, 1002, 1143, 281, 7782, 11387, 632, 50922], "temperature": 0, "avg_logprob": -0.02399882776983853, "compression_ratio": 1.727891156462585, "no_speech_prob": 2.0589782049468264e-12}, {"id": 61, "seek": 32464, "start": 335.78, "end": 341.12, "text": " the biggest model. Now it's shifting toward whoever has the best memory. A giant model that forgets", "tokens": [50922, 264, 3880, 2316, 13, 823, 309, 311, 17573, 7361, 11387, 575, 264, 1151, 4675, 13, 316, 7410, 2316, 300, 2870, 82, 51189], "temperature": 0, "avg_logprob": -0.02399882776983853, "compression_ratio": 1.727891156462585, "no_speech_prob": 2.0589782049468264e-12}, {"id": 62, "seek": 32464, "start": 341.12, "end": 346.38, "text": " everything the second the chat ends isn't that useful for a real business. A model paired with a strong", "tokens": [51189, 1203, 264, 1150, 264, 5081, 5314, 1943, 380, 300, 4420, 337, 257, 957, 1606, 13, 316, 2316, 25699, 365, 257, 2068, 51452], "temperature": 0, "avg_logprob": -0.02399882776983853, "compression_ratio": 1.727891156462585, "no_speech_prob": 2.0589782049468264e-12}, {"id": 63, "seek": 32464, "start": 346.38, "end": 350.97999999999996, "text": " retrieval system, one that actually knows your documents, your calls, your history, that's the version", "tokens": [51452, 19817, 3337, 1185, 11, 472, 300, 767, 3255, 428, 8512, 11, 428, 5498, 11, 428, 2503, 11, 300, 311, 264, 3037, 51682], "temperature": 0, "avg_logprob": -0.02399882776983853, "compression_ratio": 1.727891156462585, "no_speech_prob": 2.0589782049468264e-12}, {"id": 64, "seek": 35098, "start": 350.98, "end": 355.86, "text": " that becomes genuinely useful every single day. And this is where the competition gets interesting.", "tokens": [50365, 300, 3643, 17839, 4420, 633, 2167, 786, 13, 400, 341, 307, 689, 264, 6211, 2170, 1880, 13, 50609], "temperature": 0, "avg_logprob": -0.06938107274159663, "compression_ratio": 1.615873015873016, "no_speech_prob": 2.5837324901106795e-12}, {"id": 65, "seek": 35098, "start": 356.32, "end": 362.24, "text": " Kimi K3 shipped fully open. Quen 3.8 is still closed for now with open weights promised but no", "tokens": [50632, 5652, 72, 591, 18, 25312, 4498, 1269, 13, 2326, 268, 805, 13, 23, 307, 920, 5395, 337, 586, 365, 1269, 17443, 10768, 457, 572, 50928], "temperature": 0, "avg_logprob": -0.06938107274159663, "compression_ratio": 1.615873015873016, "no_speech_prob": 2.5837324901106795e-12}, {"id": 66, "seek": 35098, "start": 362.24, "end": 369.06, "text": " date attached. NVIDIA's Nemotron 3 embed is already fully open today. Every lab is racing on two fronts", "tokens": [50928, 4002, 8570, 13, 426, 3958, 6914, 311, 22210, 310, 2044, 805, 12240, 307, 1217, 4498, 1269, 965, 13, 2048, 2715, 307, 12553, 322, 732, 40426, 51269], "temperature": 0, "avg_logprob": -0.06938107274159663, "compression_ratio": 1.615873015873016, "no_speech_prob": 2.5837324901106795e-12}, {"id": 67, "seek": 35098, "start": 369.06, "end": 374.44, "text": " at once how smart the thinking model is and how good the memory system is that sits underneath it. The ones", "tokens": [51269, 412, 1564, 577, 4069, 264, 1953, 2316, 307, 293, 577, 665, 264, 4675, 1185, 307, 300, 12696, 7223, 309, 13, 440, 2306, 51538], "temperature": 0, "avg_logprob": -0.06938107274159663, "compression_ratio": 1.615873015873016, "no_speech_prob": 2.5837324901106795e-12}, {"id": 68, "seek": 35098, "start": 374.44, "end": 379.22, "text": " that win won't just have the smartest chat responses, they'll have systems that actually remember your", "tokens": [51538, 300, 1942, 1582, 380, 445, 362, 264, 41491, 5081, 13019, 11, 436, 603, 362, 3652, 300, 767, 1604, 428, 51777], "temperature": 0, "avg_logprob": -0.06938107274159663, "compression_ratio": 1.615873015873016, "no_speech_prob": 2.5837324901106795e-12}, {"id": 69, "seek": 37922, "start": 379.22, "end": 385.14000000000004, "text": " business the way a real employee would after a year on the job. So what should you actually do with all", "tokens": [50365, 1606, 264, 636, 257, 957, 10738, 576, 934, 257, 1064, 322, 264, 1691, 13, 407, 437, 820, 291, 767, 360, 365, 439, 50661], "temperature": 0, "avg_logprob": -0.052365068405393574, "compression_ratio": 1.7091503267973855, "no_speech_prob": 2.2807443867539634e-12}, {"id": 70, "seek": 37922, "start": 385.14000000000004, "end": 391.56, "text": " this? If you build things yourself go look at Nemotron 3 embed 8b on Hugging Face today. It's open right", "tokens": [50661, 341, 30, 759, 291, 1322, 721, 1803, 352, 574, 412, 22210, 310, 2044, 805, 12240, 1649, 65, 322, 46892, 3249, 4047, 965, 13, 467, 311, 1269, 558, 50982], "temperature": 0, "avg_logprob": -0.052365068405393574, "compression_ratio": 1.7091503267973855, "no_speech_prob": 2.2807443867539634e-12}, {"id": 71, "seek": 37922, "start": 391.56, "end": 396.40000000000003, "text": " now, no waiting required. If you run a business and don't touch code don't worry about the technical side", "tokens": [50982, 586, 11, 572, 3806, 4739, 13, 759, 291, 1190, 257, 1606, 293, 500, 380, 2557, 3089, 500, 380, 3292, 466, 264, 6191, 1252, 51224], "temperature": 0, "avg_logprob": -0.052365068405393574, "compression_ratio": 1.7091503267973855, "no_speech_prob": 2.2807443867539634e-12}, {"id": 72, "seek": 37922, "start": 396.40000000000003, "end": 401.88000000000005, "text": " at all. Just know this, the tools that remember your business are becoming way more useful than the tools", "tokens": [51224, 412, 439, 13, 1449, 458, 341, 11, 264, 3873, 300, 1604, 428, 1606, 366, 5617, 636, 544, 4420, 813, 264, 3873, 51498], "temperature": 0, "avg_logprob": -0.052365068405393574, "compression_ratio": 1.7091503267973855, "no_speech_prob": 2.2807443867539634e-12}, {"id": 73, "seek": 37922, "start": 401.88000000000005, "end": 406.86, "text": " that just chat with you. Start thinking about what parts of your business you'd want an AI to actually", "tokens": [51498, 300, 445, 5081, 365, 291, 13, 6481, 1953, 466, 437, 3166, 295, 428, 1606, 291, 1116, 528, 364, 7318, 281, 767, 51747], "temperature": 0, "avg_logprob": -0.052365068405393574, "compression_ratio": 1.7091503267973855, "no_speech_prob": 2.2807443867539634e-12}, {"id": 74, "seek": 40686, "start": 406.86, "end": 412.7, "text": " remember. Your best scripts, your best offers, your past client questions. If you manage a team, this is the", "tokens": [50365, 1604, 13, 2260, 1151, 23294, 11, 428, 1151, 7736, 11, 428, 1791, 6423, 1651, 13, 759, 291, 3067, 257, 1469, 11, 341, 307, 264, 50657], "temperature": 0, "avg_logprob": -0.058796508639466526, "compression_ratio": 1.6544117647058822, "no_speech_prob": 2.2539628583700955e-12}, {"id": 75, "seek": 40686, "start": 412.7, "end": 418.54, "text": " moment to ask which parts of your daily work are just people searching for information that already exists", "tokens": [50657, 1623, 281, 1029, 597, 3166, 295, 428, 5212, 589, 366, 445, 561, 10808, 337, 1589, 300, 1217, 8198, 50949], "temperature": 0, "avg_logprob": -0.058796508639466526, "compression_ratio": 1.6544117647058822, "no_speech_prob": 2.2539628583700955e-12}, {"id": 76, "seek": 40686, "start": 418.54, "end": 424.78000000000003, "text": " somewhere. That's exactly the kind of task this tech was built to remove. If you want help actually setting this up", "tokens": [50949, 4079, 13, 663, 311, 2293, 264, 733, 295, 5633, 341, 7553, 390, 3094, 281, 4159, 13, 759, 291, 528, 854, 767, 3287, 341, 493, 51261], "temperature": 0, "avg_logprob": -0.058796508639466526, "compression_ratio": 1.6544117647058822, "no_speech_prob": 2.2539628583700955e-12}, {"id": 77, "seek": 40686, "start": 424.78000000000003, "end": 430.44, "text": " for your own business, that's exactly what we walk through inside the iProfit boardroom. We're already building a full", "tokens": [51261, 337, 428, 1065, 1606, 11, 300, 311, 2293, 437, 321, 1792, 807, 1854, 264, 5180, 340, 6845, 3150, 2861, 13, 492, 434, 1217, 2390, 257, 1577, 51544], "temperature": 0, "avg_logprob": -0.058796508639466526, "compression_ratio": 1.6544117647058822, "no_speech_prob": 2.2539628583700955e-12}, {"id": 78, "seek": 43044, "start": 430.44, "end": 437.71999999999997, "text": " playbook around Quen 3.8 and Nemotron 3 embed together, covering how to connect them, what to feed them, and how to turn", "tokens": [50365, 862, 2939, 926, 2326, 268, 805, 13, 23, 293, 22210, 310, 2044, 805, 12240, 1214, 11, 10322, 577, 281, 1745, 552, 11, 437, 281, 3154, 552, 11, 293, 577, 281, 1261, 50729], "temperature": 0, "avg_logprob": -0.08572627652075983, "compression_ratio": 1.673780487804878, "no_speech_prob": 1.91261654951258e-12}, {"id": 79, "seek": 43044, "start": 437.71999999999997, "end": 442.28, "text": " that combination into a real memory system for your business instead of just another tool you open once", "tokens": [50729, 300, 6562, 666, 257, 957, 4675, 1185, 337, 428, 1606, 2602, 295, 445, 1071, 2290, 291, 1269, 1564, 50957], "temperature": 0, "avg_logprob": -0.08572627652075983, "compression_ratio": 1.673780487804878, "no_speech_prob": 1.91261654951258e-12}, {"id": 80, "seek": 43044, "start": 442.28, "end": 447.96, "text": " and forget about. Every week we run live coaching calls where you can bring your exact setup and get help", "tokens": [50957, 293, 2870, 466, 13, 2048, 1243, 321, 1190, 1621, 15818, 5498, 689, 291, 393, 1565, 428, 1900, 8657, 293, 483, 854, 51241], "temperature": 0, "avg_logprob": -0.08572627652075983, "compression_ratio": 1.673780487804878, "no_speech_prob": 1.91261654951258e-12}, {"id": 81, "seek": 43044, "start": 447.96, "end": 453.0, "text": " on the spot no matter where you're stuck. There's a whole community inside already testing these exact releases", "tokens": [51241, 322, 264, 4008, 572, 1871, 689, 291, 434, 5541, 13, 821, 311, 257, 1379, 1768, 1854, 1217, 4997, 613, 1900, 16952, 51493], "temperature": 0, "avg_logprob": -0.08572627652075983, "compression_ratio": 1.673780487804878, "no_speech_prob": 1.91261654951258e-12}, {"id": 82, "seek": 43044, "start": 453.0, "end": 458.44, "text": " the same week they drop. So you're learning this alongside people actually using it, not watching from the", "tokens": [51493, 264, 912, 1243, 436, 3270, 13, 407, 291, 434, 2539, 341, 12385, 561, 767, 1228, 309, 11, 406, 1976, 490, 264, 51765], "temperature": 0, "avg_logprob": -0.08572627652075983, "compression_ratio": 1.673780487804878, "no_speech_prob": 1.91261654951258e-12}, {"id": 83, "seek": 45844, "start": 458.44, "end": 464.12, "text": " sidelines months later. Links in the comments and description or go straight to AIprofitboardroom.com.", "tokens": [50365, 20822, 9173, 2493, 1780, 13, 37156, 294, 264, 3053, 293, 3855, 420, 352, 2997, 281, 7318, 14583, 3787, 2861, 13, 1112, 13, 50649], "temperature": 0, "avg_logprob": -0.17351768130347842, "compression_ratio": 1.6367041198501873, "no_speech_prob": 2.6248404491613364e-12}, {"id": 84, "seek": 45844, "start": 464.76, "end": 471.96, "text": " And if you want the full process, the SOPs and over 100 AI use cases, just like this one, come join the AI", "tokens": [50681, 400, 498, 291, 528, 264, 1577, 1399, 11, 264, 10621, 23043, 293, 670, 2319, 7318, 764, 3331, 11, 445, 411, 341, 472, 11, 808, 3917, 264, 7318, 51041], "temperature": 0, "avg_logprob": -0.17351768130347842, "compression_ratio": 1.6367041198501873, "no_speech_prob": 2.6248404491613364e-12}, {"id": 85, "seek": 45844, "start": 471.96, "end": 478.44, "text": " success lab. Links are in the comments and description. You'll get all the notes from this exact video there, plus access to a", "tokens": [51041, 2245, 2715, 13, 37156, 366, 294, 264, 3053, 293, 3855, 13, 509, 603, 483, 439, 264, 5570, 490, 341, 1900, 960, 456, 11, 1804, 2105, 281, 257, 51365], "temperature": 0, "avg_logprob": -0.17351768130347842, "compression_ratio": 1.6367041198501873, "no_speech_prob": 2.6248404491613364e-12}, {"id": 86, "seek": 45844, "start": 478.44, "end": 484.36, "text": " community of 87,000 people who are already using AI to move their business forward every single day.", "tokens": [51365, 1768, 295, 27990, 11, 1360, 561, 567, 366, 1217, 1228, 7318, 281, 1286, 641, 1606, 2128, 633, 2167, 786, 13, 51661], "temperature": 0, "avg_logprob": -0.17351768130347842, "compression_ratio": 1.6367041198501873, "no_speech_prob": 2.6248404491613364e-12}], "language": "en"}