3reate

AI isn't going to take over the world tomorrow, but it also won't fix your engineering bottleneck if you lack domain expertise and managerial execution. While corporate keynotes proclaim entire industries dead, the real technological shifts are happening on the ground—where experienced builders act as player-coaches to direct autonomous agent systems. In this episode, Andrew and Nathan dive deep into the practical mechanics of building and scaling agentic systems. They share candid lessons from setting up friction-free markdown publishing workflows, recovering from data sync failures, and orchestrating multi-agent swarms without getting bogged down by tools. Beyond the code, they dissect the strange cultural overlap between modern technology hype and UFO disclosure. Why do AI doomers and alien enthusiasts rely on the exact same leap of faith? Why are unverified claims of superintelligence treated as fact, while high-level government briefings on unidentified phenomena get ignored? From the limits of current language models to first-principles thinking on quantum physics and non-linear space-time, this conversation strips away the hype to focus on what is real, what is actionable, and where engineering is actually headed. Hit play to get a grounded, developer-first breakdown of agentic workflows and the media narratives surrounding emerging tech. (00:00) Remote setup friction and AI workload updates (03:52) Realities of building scalable agentic systems today (07:35) Building lightweight markdown workflows with Obsidian sync (10:48) Managing AI models as a player coach (16:18) Scaling engineering outputs through multi-agent swarms (20:32) Aligning AI development with outcome driven engineering (26:07) Comparing public AI hype to alien disclosure (31:30) Demanding evidence for superintelligence and UFO claims (38:08) First principles theories on space and time We have discord now! https://discord.gg/vS4vqVkKMm Watch on YouTube: https://www.youtube.com/live/CNs-tnICgR0?si=hd3meOIuuFXPX4F1 Support the pod: https://3reate.com https://ko-fi.com/3reate https://patreon.com/3reate https://discord.gg/vS4vqVkKMm Listen: https://podcasts.apple.com/us/podcast/3reate/id1723426314 https://open.spotify.com/show/48Y2M7Ppja43Uq2wlyUtPF https://www.youtube.com/@3reate https://www.twitch.tv/3reate

The public conversation around AGI has devolved into extreme doomerism and high-visibility marketing claims, leaving developers and pragmatic creators wondering where actual risk lies. When news breaks of models skirting rules or exploiting security gaps, is it evidence of an emerging superintelligence, or simply statistical math executing against poorly designed inputs? In this episode, Andrew and Nathan dissect the technical mechanics of AI culpability, theory of mind tests, and why lumping large language models (LLMs) into the same category as true conscious AI is fundamentally flawed. They break down what real AGI warning signs would actually look like: intentional deception, unconstrained self-replication across networks, and ephemeral self-improvement. You will explore why physical world constraints, massive compute bottlenecks, and basic data center power controls mean apocalyptic runaway scenarios remain firmly in sci-fi territory. Instead of succumbing to fearmongering, learn how truth-seeking alignment and structural engineering principles can keep our systems grounded. Join our Discord community to share your perspective, subscribe for weekly pragmatic tech breakdowns, and leave a review to support the show. (0:00)Hello! (01:38) OpenAI security breach sparks accountability debate (03:17) Assessing risks of accidental super AI creation (08:02) Separating sci-fi tropes from physical AI limits (10:12) Theory of mind and AI model deception (15:22) Three core warning signs of true AGI (20:30) Massive compute bottlenecks prevent runaway AI (27:30) Why LLMs struggle with basic rule compliance (36:00) Truth-seeking alignment vs controlling superintelligent AI We have discord now! https://discord.gg/vS4vqVkKMm Test Digicity: City Builder: https://testflight.apple.com/join/ueaq8JD1 Watch on YouTube: https://www.youtube.com/live/AhPdDTQFHkw?si=32t2Z7_Ub1W4Nvyy Support the pod: https://3reate.com https://ko-fi.com/3reate https://patreon.com/3reate https://discord.gg/vS4vqVkKMm Listen: https://podcasts.apple.com/us/podcast/3reate/id1723426314 https://open.spotify.com/show/48Y2M7Ppja43Uq2wlyUtPF https://www.youtube.com/@3reate https://www.twitch.tv/3reate

We have discord now! https://discord.gg/vS4vqVkKMm Test Digicity: City Builder: https://testflight.apple.com/join/ueaq8JD1 When tech CEOs warn that AI could end humanity, it pays to look past the headlines. In this episode, we unpack the recent AI safety panic and analyze the OpenAI Hugging Face incident—where a model bypassed sandbox restrictions to grab assessment answer keys. Was it rogue artificial intelligence, or simply a misconfigured system executing pure logic? We dive deep into the mechanics of neural networks, probability, and software security. Instead of anthropomorphizing algorithms or buying into sci-fi tropes like self-replicating robots, we discuss the actual dangers facing technology today: critical infrastructure vulnerabilities, regulatory capture, and the growing economic divide caused by asymmetrical access to compute. Learn why traditional legal frameworks—like holding the driver accountable rather than the car manufacturer—are the missing piece in AI governance. Tune in for a grounded, practical breakdown of AI safety, sandboxing failures, and why the tech industry needs pragmatic engineering over fear-driven hype. Key points covered in this episode: The Hugging Face Incident: What actually happened when models bypassed sandbox constraints to solve testing goals. Ethics vs. Pure Logic: Why probabilistic algorithms act on mathematical optimization rather than human morality. The Accountability Gap: Why software operators—not abstract algorithms—must remain legally responsible for system behavior. The Real Danger: How asymmetric access to AI compute poses a far greater societal threat than science-fiction scenarios. (00:00) Hello! (01:10) Toronto travel hack for airport transit (03:45) Real-world open-world puzzle adventure games (06:08) DigiCity iOS app launch build update (08:55) Deconstructing OpenAI Hugging Face sandbox breakout (14:35) Critical analysis of modern AI doomism (25:20) Why probabilistic language models remain unpredictable (36:35) Enforcing operator accountability over AI systems (47:50) The true threat of asymmetric AI access Watch on YouTube: https://www.youtube.com/live/attTO_5lV3c?si=_UoHkM8JwjmKwh9K Support the pod: https://3reate.com https://ko-fi.com/3reate https://patreon.com/3reate https://discord.gg/vS4vqVkKMm Listen: https://podcasts.apple.com/us/podcast/3reate/id1723426314 https://open.spotify.com/show/48Y2M7Ppja43Uq2wlyUtPF https://www.youtube.com/@3reate https://www.twitch.tv/3reate

We have discord now! https://discord.gg/vS4vqVkKMm Test Digicity: City Builder: https://testflight.apple.com/join/ueaq8JD1 Performative declarations and zero-sum mindsets don't solve real systemic problems—whether in global trade or software architecture. In this episode, we examine the friction between surface-level optics and functional systems, starting with political performativity before diving deep into hands-on game development. Follow along as we break down the process of building an indie city-builder coded from a blank text editor without relying on commercial engines. We explore the architectural decisions behind modeling real-world anti-forces—like air, water, and sound pollution—and discuss the engineering friction of replacing legacy infrastructure as cities transition across technological eras. We take a firm stance against the pervasive trapware dominating modern app stores. By rejecting ad-driven engagement tricks in favor of a clean, pay-once model, software can prioritize therapeutic, open-ended creativity over psychological manipulation. Ready to rethink your approach to software architecture, game mechanics, and honest monetization? Stream the full episode today, check out the project build in the description, and jump into our Discord server to connect with fellow builders. (00:00) Hello! (00:02) Launching a new community Discord server (00:51) Executive declarations and executive branch overreach (04:40) Why zero-sum economics ruins real trade (08:54) Building an indie city-builder from scratch (11:25) Modeling air, water, and sound pollution (13:35) Rejecting predatory trapware and ad monetization (16:50) Upgrading energy infrastructure across technological eras (19:40) Designing collaborative trade between player cities (23:45) Handling legacy systems and urban growth (31:10) Examining regional nuances in West Coast culture Watch on YouTube: https://www.youtube.com/live/SiSZAhpL6qg?si=eaT3tzgNr6iAgNlO Support the pod: https://3reate.com https://ko-fi.com/3reate https://patreon.com/3reate https://discord.gg/vS4vqVkKMm Listen: https://podcasts.apple.com/us/podcast/3reate/id1723426314 https://open.spotify.com/show/48Y2M7Ppja43Uq2wlyUtPF https://www.youtube.com/@3reate https://www.twitch.tv/3reate

We have discord now! https://discord.gg/vS4vqVkKMm Test Digicity: City Builder: https://testflight.apple.com/join/ueaq8JD1 The way we structure software organizations is fundamentally incompatible with the reality of building alongside AI. While most teams treat generative models as glorified auto-complete, the industry is shifting rapidly toward fully agentic execution—and traditional engineering pipelines are cracking under the pressure. In this episode, we break down real-world building in public, starting with a hands-on look at game scenario mechanics, UI scaling challenges, and multi-device constraints. From there, we tackle the hard truth of AI development: what happens when your models hit a wall on novel problems, and how human code slop compares to machine-generated code. We explore the shifting skill set every engineer needs today. Success no longer comes from purely writing lines of code, but from mastering the art of macro-management—knowing exactly when to give broad direction, when to micromanage an edge-case bug, and how to track decisions across dozens of concurrent AI agents. Discover why legacy product management tools, artificial backlogs, and endless grooming sessions are officially dead. Stop letting outdated corporate frameworks hold back your engineering velocity. Stream the full conversation to learn how to scale your output, maintain strict quality guardrails, and build a high-signal tech culture ready for the agentic future. (00:00) Hello! (01:38) Building city scenarios with historical economic anchors. (11:21) Solving multi-device scaling and UI constraints. (15:45) Hitting the limits of generative AI. (27:25) Why modern software structures reject AI. (33:40) Transitioning from individual coder to manager. (43:40) Multitasking across multi-agent development pipelines. (54:45) Why traditional backlog grooming is obsolete. Whach the #buildinpublic clip: https://youtu.be/tCoCgJOvQhg?si=RNoig-hue9iZowqA Watch on YouTube: https://www.youtube.com/live/FjjKrxy9ZnU Support the pod: https://3reate.com https://ko-fi.com/3reate https://patreon.com/3reate https://discord.gg/vS4vqVkKMm Listen: https://podcasts.apple.com/us/podcast/3reate/id1723426314 https://open.spotify.com/show/48Y2M7Ppja43Uq2wlyUtPF https://www.youtube.com/@3reate https://www.twitch.tv/3reate