Welcome to the Jon Myer Podcast, powered by Myer Media! Fresh off the keynote stage at AWS re:Invent 2025, we sit down with Nathan from Metal Toad, an AWS partner helping organizations navigate digital transformation and cloud-native deployment.We unpack the biggest announcements from today's keynote, why they matter, and how these innovations translate into real-world opportunities for enterprises—including when AI might NOT be the right solution yet.Key Topics Discussed:✅ Fine-tuning and distilling models: Making small models more accessible✅ Why small models are the future for accuracy, speed, and edge deployment✅ Nova snapshots: Starting 90% through training, saving millions in costs✅ Real-time AI responses: From 2-3 seconds down to half a second✅ Maintaining engineer flow state with faster AI interactions✅ Kiro going GA: Immediately usable for teams✅ The honest ROI conversation: When NOT to implement AI✅ Moving AI to the edge: Reducing cloud cost variability✅ Trends for 2026: Edge AI adoption accelerating faster than anticipatedNathan's Insights:💡 "Small models can run on smaller form factors—you could actually move them to the edge and run directly on a laptop"💡 "Getting something done to deliver value is better than stopping and starting something newer and shinier"💡 "Calculate ROI first. If it takes 40 hours to implement for 2 hours a week savings, is that worth it?"💡 "Sometimes we tell clients: you don't need AI yet. Building that trust is more important than making a sale"This conversation helps you separate hype from practical implementation, understand immediate opportunities vs. strategic planning needs, and know when to wait before jumping on the AI bandwagon.Guest: Nathan, Metal ToadHost: Jon MyerRecorded at: AWS re:Invent 2025, Las VegasYouTube Timeline0:00 - Welcome from AWS re:Invent 2025 keynote stage0:49 - Game changer: Fine-tuning and distilling small models1:13 - Benefits of small models: Speed, edge deployment, cost1:38 - Real-time innovation and customer experience improvements2:00 - Keeping engineers in flow state with faster responses2:20 - Nova snapshots: Starting 90% through training2:47 - Millions in savings by not training from scratch3:05 - What's immediately usable vs. requiring strategic planning3:47 - Calculating ROI: Is 40 hours of setup worth it?4:13 - How to prioritize when innovation comes all at once4:41 - Rapid assessment: Helping clients understand if they need AI5:23 - The hard conversation: Telling clients they don't need AI yet5:47 - Building trust by saying "not yet"6:09 - 2026 trends: AI moving to the edge faster than anticipated6:33 - Closing thoughts & thank you🔔 Don't forget to Like, Subscribe, and hit the notification 🔔 ✔ Subscribe: https://www.youtube.com/jonmyer/?sub_confirmation=1📱 Social Media Twitter: https://twitter.com/_JonMyerWebsite: https://jonmyer.comLinkedIN: https://www.linkedin.com/in/jon-myer/Spotify: https://open.spotify.com/show/0wjJzdIwctq4o4pTXM2KSFLike my sounds? Here's my audio source: https://www.epidemicsound.com/referral/fj9o9k#aws #awscloud #podcast #podcasting #costoptimization #finops #cloudcost #cloudoptimization

Welcome to the Jon Myer Podcast, powered by Myer Media! Fresh off the keynote stage at AWS re:Invent 2025, we sit down with David from Metal Toad, an AWS partner helping organizations navigate digital transformation and cloud-native deployment.We unpack the biggest announcements from today's keynote, why they matter, and how these innovations translate into real-world opportunities for enterprises accelerating their cloud journey.Key Announcements Discussed:✅ Nova2 model family with stronger model differentiation✅ Frontier models: Long-running AI that learns your team's coding style✅ NovaForge: Earlier access to models in the training process✅ S3 vector storage: Up to 90% cost savings vs. OpenSearch for RAG databases✅ Nova2 Omni: Multimodal inputs for unstructured data (audio, video)✅ Agentic AI: Building compartmentalized "brains" for better controlDavid's Insights:💡 "If you think you need AI, you maybe don't—start by identifying edge cases where problems become too squishy to solve programmatically"💡 Frontier models can run for days or weeks, learning your team's style and growing with you💡 S3 vector storage eliminates expensive always-on containers, paying only for what you use💡 Agentic AI gives you more control over your system vs. "text input and hope for the right output"Whether you're mid-market or enterprise, this conversation helps you prioritize innovation, separate hype from practical implementation, and understand what you can leverage immediately vs. what requires strategic planning.Guest: David Dolan, Metal ToadHost: Jon MyerRecorded at: AWS re:Invent 2025, Las VegasYouTube Timeline0:00 - Welcome from AWS re:Invent keynote stage0:52 - Game-changing announcements: Nova2, S3, Frontier models1:18 - What are Frontier models and why are they exciting?1:57 - Long-running models that adapt to your team2:03 - NovaForge: Earlier access to model training2:29 - Why earlier model access matters for performance2:59 - How AI announcements impact mid-market & enterprise3:15 - S3 vector storage: Up to 90% cost savings4:20 - What teams can leverage immediately vs. strategic planning4:35 - Nova2 Omni: Solving the unstructured data problem5:09 - Advising clients when there's too much innovation at once5:43 - Trends defining cloud and AI for 20266:00 - Agentic AI: Building compartmentalized systems with control6:36 - Closing thoughts & thank you🔔 Don't forget to Like, Subscribe, and hit the notification 🔔 ✔ Subscribe: https://www.youtube.com/jonmyer/?sub_confirmation=1📱 Social Media Twitter: https://twitter.com/_JonMyerWebsite: https://jonmyer.comLinkedIN: https://www.linkedin.com/in/jon-myer/Spotify: https://open.spotify.com/show/0wjJzdIwctq4o4pTXM2KSFLike my sounds? Here's my audio source: https://www.epidemicsound.com/referral/fj9o9k#aws #awscloud #podcast #podcasting #costoptimization #finops #cloudcost #cloudoptimization

Joining us today is Jeff Barr, Chief Evangelist for AWS, recorded live at AWS re:Invent 2025. For two decades, Jeff has documented the evolution of cloud computing through the AWS blog, sharing insights on everything from the launch of EC2 instances to the latest breakthroughs in generative AI and machine learning.In this episode, we explore what's next in the AI-assisted workplace, where this technology is heading, how professionals can navigate the shift to AI-native work, and what guidance Jeff has for individuals and organizations trying to turn AI capabilities into competitive advantages.Key Topics Discussed:✅ The rapid global adoption of AI-assisted development (50%+ in 14 countries)✅ Why developers need to become better human-to-human communicators✅ The shift from writing code to reading and validating AI-generated code✅ Balancing productivity with AI hallucinations and errors✅ The art of prompt engineering and verbose specifications✅ Disposable apps: Building to solve, then rebuilding better next time✅ Why some skills are becoming more valuable in the AI era✅ Swimming upstream: Being an early adopter and giving feedback✅ Language barriers disappearing with AI translation capabilities✅ Jeff's 14 years at re:Invent and stepping back from the AWS blogWhether you're deep in your cloud journey or just getting started, this conversation gives you a front-row seat to the future being built at re:Invent.Guest: Jeff Barr, Chief Evangelist, AWSHost: Jon MyerRecorded at: AWS re:Invent 2025, Las VegasSubscribe for more conversations about cloud, AI, and the future of technology innovation.#AWS #ArtificialIntelligence #CloudComputing #reInvent #AIAssistedDevelopment #GenerativeAI #TechInnovation #JonMyerPodcastYouTube Timeline:0:00 - Welcome & Introduction1:19 - Annual tradition at re:Invent with Jeff Barr2:25 - Surprising speed of global AI adoption3:51 - Jeff's year traveling the world talking to developers4:45 - Key guidance: Step in and use AI tools6:18 - Learning to read code, not just write it7:51 - The personal chef analogy for prompt engineering9:28 - When AI gets it incredibly wrong: The college fund story11:12 - Balancing productivity with hallucinations13:18 - Why the same prompt gives different outputs14:29 - Embracing change: 50 years in tech16:01 - AI breaking down language barriers17:10 - When companies aren't AI-ready18:24 - From experimentation to problem-solving19:42 - Disposable apps: Build, use, toss, rebuild better20:29 - Essential skills for AI-assisted work21:15 - You have to experience it, not just read about it21:40 - What Jeff looks forward to at re:Invent23:42 - Stepping away from the AWS blog after 20 years26:15 - Advice for re:Invent attendees27:19 - Closing thoughts🔔 Don't forget to Like, Subscribe, and hit the notification 🔔 ✔ Subscribe: https://www.youtube.com/jonmyer/?sub_confirmation=1📱 Social Media Twitter: https://twitter.com/_JonMyerWebsite: https://jonmyer.comLinkedIN: https://www.linkedin.com/in/jon-myer/Spotify: https://open.spotify.com/show/0wjJzdIwctq4o4pTXM2KSFLike my sounds? Here's my audio source: https://www.epidemicsound.com/referral/fj9o9k#aws #awscloud #podcast #podcasting #costoptimization #finops #cloudcost #cloudoptimization