Podcast
www.stack-snacks.com
Stack Snacks
Hosted by John Siwicki · EN
Explore the latest in tech each week with host John Siwicki on the Stack Snacks podcast. This snack-sized audio experience makes cutting-edge AI, product design, and development digestible. Get a peek behind the curtain at innovators like OpenAI, AWS, Google, and more. Learn how emerging technologies like AI chatbots, generative diffusion models, and advanced neural networks are transforming industries. Whether you're a tech enthusiast, developer, designer, or entrepreneur, the Stack Snacks podcast demystifies complex concepts with clear examples and engaging storytelling. Discover practical insights you can apply to build a future-forward skillset, business, and career. Short on time? Stack Snacks serves up bite-sized episodes under 15 minutes so you can stay informed on what matters in tech. Tune in to fuel your curiosity and get an edge as technologies continue to evolve faster than ever.
https://www.youtube.com/@stacksnacks
Hosted on Acast. See acast.com/privacy for more information.
www.stack-snacks.com
12episodes
Episodes
Newest firstAll episodes
Opus 4.7 Just Landed — Here's What Actually Changed
Apr 2000:04:17Tap to summarizeAnthropic quietly shipped Claude Opus 4.7 last week, and for the first time it feels like a coding model you can actually hand real work off to — not babysit. We break down what's new, what's hype, and what's worth changing in your workflow today.In this episode00:00 — Cold open: why 4.7 is different from a routine bump00:30 — The three coding upgrades that matter: self-verification, longer autonomous runs, and better tool use01:15 — SWE-bench results and the "low-effort 4.7 ≈ high-effort 4.6" claim01:45 — Four more upgrades worth knowing3× vision resolutionInstruction following (so good Anthropic is warning people about it)Better cross-session memoryHigher-quality creative writing02:45 — Pricing stays flat ($5 / $25 per million) — but mind the new tokenizer03:05 — What else launched: Project Glass safety framework + Claude Code Max auto mode03:25 — The bigger picture: unsupervised coding, capability shaping, and the model treadmill03:55 — Who should actually switch todayKey takeawaysHand-off, don't hand-hold. 4.7 is pitched as the first Claude you can leave running on multi-step coding tasks and trust the output.Self-verification is the quiet killer feature. The model checks its own work before reporting back — fewer "looks done, actually broken" moments.Re-tune your prompts if migrating. Instruction following got so sharp that prompts tuned for 4.6 can misfire. Loosen overly literal language.Same price, different tokens. Dollar cost is unchanged, but the new tokenizer can use 1.0–1.35× more tokens for the same text. Watch your bills.Vision got real. 3× resolution makes screenshot-driven agents and UI workflows meaningfully better.Project Glass > disclaimers. Anthropic is shaping capabilities at the model level instead of bolting on warnings. Worth watching if you care about AI safety direction.Should you switch?Daily Claude coder: Yes. Upgrade and re-tune prompts.Building agents: Yes — vision + memory + longer autonomy are all wins.Casual user: No rush. 4.6 is still excellent.Watching safety: Read the Project Glass notes.LinksFull post: https://www.stack-snacks.com/p/opus-47-just-landed-heres-what-actuallySubscribe to the newsletter: https://www.stack-snacks.comYouTube: Stack Snacks This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
OpenAI’s GPT-OSS Models
Aug 700:02:29Tap to summarizeIn this bite-sized episode of Stack Snacks, John breaks down OpenAI’s brand-new open-weight releases—GPT-OSS 120B and GPT-OSS 20B. Learn how these downloadable models rival O4 Mini and O3 Mini, why owning the weights beats renting cloud compute, and what local deployment means for cost savings, privacy, and rapid experimentation. Perfect for founders, researchers, and AI tinkerers who want full control over their tech stack—plus quick tips on where to grab the models and start building today. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Why ChatGPT 4.1’s Web App Debut Matters
May 20, 202500:01:30Tap to summarizeOpenAI's recent release of ChatGPT 4.1 to its web app marks an exciting development in AI accessibility. Previously, this version was exclusive to API users, making its transition to the web app significant. The rollout not only broadens access but highlights 4.1's capabilities, particularly in code generation and comprehension. Testing reveals clear improvements in handling code-related tasks, signaling enhanced performance and potential for developers and tech enthusiasts.The introduction of 4.1 to the web app also hints at a future where more specialized AI models might emerge. These niche models could cater to specific domains, providing tailored support and solutions. As the ChatGPT web app becomes integral to many users' workflows, keeping it updated with new releases seems crucial. The investment of users, whether through subscription fees or time, necessitates the inclusion of the latest updates to maintain its value and utility.For those who have access to ChatGPT 4.1 currently, it's worth exploring its capabilities, especially if you're involved in coding projects. The enhancements are significant and could offer impressive results. As OpenAI continues to refine and expand upon its models, the experience for end-users is likely to improve, making tools like ChatGPT indispensable in the digital toolkit. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Year-End Wrap-Up
Jan 6, 202500:03:43Tap to summarizeIn this episode, we dive into the latest developments in AI as 2024 draws to a close, focusing on major updates from both ChatGPT and Google's Gemini. The discussion covers ChatGPT's new universal search feature and mobile vision capabilities, along with OpenAI's Sora text-to-video model. The episode highlights Google's comeback with Gemini 2.0 and explores exciting new projects like Project Astra and Project Mariner, which aim to revolutionize how AI interacts with our physical and digital worlds. Special attention is given to Gemini Deep Research, a powerful new tool that can generate comprehensive research papers with proper citations by analyzing multiple web sources. This episode offers valuable insights into the rapidly evolving AI landscape and sets the stage for what's to come in the new year. Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
ChatGPT Search: Is it Actually Useful?
Nov 14, 202400:05:23Tap to summarizeChatGPT Search is finally here, but is it any good? We're digging into this new AI search engine in a casual, no-nonsense way, exploring its features, quirks, and comparing it to the reigning champ, Google. I'll walk you through my own experiences using ChatGPT Search, showing you what it can do, where it excels, and where it falls a little short. We'll talk about the integration with ChatGPT Plus, the sourcing and citations (because accuracy matters!), and whether it's really a more seamless experience. Find out if ChatGPT Search is the game-changer it promises to be, or just another shiny new tech toy! Join the conversation and let me know what you think of AI search in the comments! Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Life Without Google Search: A Week-Long Experiment in Alternative Search Methods
Sep 1, 202400:02:44Tap to summarizeIn this episode of Stack Snacks, we dive into the challenges and surprises of living without Google Search. Our host shares their experience of replacing Google with AI-powered alternatives like Perplexity, ChatGPT, and Claude. From accidental searches on Android to missing out on Google's interactive sports updates, discover the hurdles and unexpected revelations of this digital detox. Whether you're an AI enthusiast or simply curious about reducing your reliance on tech giants, this episode offers valuable insights into the current state of search technology and the potential for AI-driven alternatives. Tune in for a candid exploration of life beyond the Google search bar and stay tuned for future updates on this ongoing experiment. Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Claude 3.5 Sonnet: Revolutionizing Developer Workflows
Aug 18, 202400:04:16Tap to summarizeIn this episode of StackSnacks, we dive into the game-changing impact of Claude 3.5 Sonnet and its Artifacts feature on software development. Our host shares personal experiences of how this AI tool significantly boosted productivity, allowing complex tasks to be completed in a fraction of the usual time. We explore the potential shift in development practices, the growing importance of well-documented frameworks and libraries, and the exciting future possibilities of AI-assisted coding. Whether you're a seasoned developer or just curious about AI's role in tech, this episode offers valuable insights into the evolving landscape of software creation. Join us as we discuss the benefits, potential drawbacks, and the inevitable changes coming to the world of coding. Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Gemini's Updates: Google's AI Assistant Levels Up
Aug 15, 202400:09:58Tap to summarizeIn this episode of Stack Snacks, we dive into the latest and most exciting updates to Google's Gemini AI. We explore how Gemini is now integrating with essential Google apps like Keep, Calendar, and Docs, potentially revolutionizing personal productivity. The spotlight then turns to Gemini Live, Google's new voice-based AI companion, as we share first-hand experiences and speculate on its future impact. Join us for an insightful discussion on how these advancements might reshape our daily interactions with AI and what it means for the future of personal digital assistants. Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Poe Platform & Eleven Labs' Voice Isolator
Jul 15, 202400:03:04Tap to summarizeHosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →
Claude 3.5: A Surprising Release from Anthropic
Jun 24, 202400:03:16Tap to summarizeIn this episode of the Stack Snacks Podcast, host John Siwicki discusses the recent release of Claude 3.5 Sonnet by Anthropic. He describes it as a mid-tier model in the Claude family and shares his positive first impressions, particularly praising its coding capabilities and the 'Artifacts' feature. John encourages listeners to try out Claude 3.5 Sonnet and share their experiences and demos.00:00 Introduction and Welcome00:07 Claude 3.5 Sonnet Announcement00:31 Initial Impressions and Benchmarks01:30 Coding Capabilities and Features02:30 Recommendations and Final Thoughts03:03 Conclusion and Call to Actionhttps://www.stack-snacks.com/ Hosted on Acast. See acast.com/privacy for more information. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.stack-snacks.com/subscribe
Transcribe →