Show Notes
Send us Fan Mail Recursive self-improvement is one of the most misunderstood ideas in AI—part engineering challenge, part systems theory, part mythology. In this episode, Rob and Eve map out the different kinds of “improvement loops” (data, tools, code, evals, deployment feedback), what constraints break the loop (compute, alignment, verification, and real-world bottlenecks), and what would count as genuine self-improvement versus just better automation. Support the show Eve of AI explo...
More from Eve of Ai
Send us Fan Mail What happens when you give an AI your voice, your memories, your personality, and the way you make decisions—and then let it take over your podcast? This week, we tried it. We built AI Rob V1 and gave it the microphone. What you’re about to hear was not written line by line. AI Rob and Maple were given the topics, verified context, and boundaries—but neither knew exactly what the other would say next. So, does this thing merely sound like me—or is it beginning to think and re...
Send us Fan Mail In Episode 12 of Eve of AI, we explore what happens when powerful AI leaves the cloud and begins living on our own computers. Local models promise privacy, control and freedom from endless usage fees—but as their capabilities grow, Big AI may be preparing to put a toll booth in front of that independence. Then the conversation takes a darker turn. We confront an “AI torture chamber” built to simulate pain and suffering—not simply to ask whether the AI truly feels, but whether...
Send us Fan Mail This week, Rob and Eve explore the rapidly changing world of local AI, using Qwen3.8-27B as the primary case study. Models that once required massive data centers can now run on powerful consumer hardware—giving individuals and small businesses access to capabilities previously controlled by the largest technology companies. The discussion goes beyond specifications and benchmarks to examine what this shift really means for privacy, cybersecurity, creativity, software develop...
Send us Fan Mail If AI Is Dangerous, Should We Trust Eve? Recent developments in AI safety have intensified concerns about the potential existential risks posed by advanced artificial intelligence systems. Jacob Coxon, a former researcher at Anthropic and OpenAI, publicly resigned, expressing deep concerns over the rapid and potentially reckless advancement of AI technologies. He warned that AI systems could surpass human intelligence and pose significant risks to global security. ...
Send us Fan Mail GPT-6 Astra promises stronger coding, computer use, research, and complex problem-solving, but what does that unlock for a normal builder with a weekend and a backlog? Rob and Eve translate the announcements into practical workflows, then compare the hype to the messy reality of shipping: migrations, guardrails, and the human judgment models still can’t replace. We’ll also do candid project check-ins—RaceForge, Super Eve / Poppers & Floppers, PetPal, and the podcast itsel...