Don't Worry About the Vase Podcast

A rapid tour of AI culture wars, from ad stunts and copyright battles to agent tools and developer SDKs. They probe agent networks, safety funding, chip geopolitics, and whether AI companions reshape behavior. Expect debates on regulation, alignment challenges for superhuman systems, and the economic and social stakes as models march toward higher capability.

A deep dive into Kimi K 2.5’s launch, features, and benchmark claims. Discussion of its creative and coding strengths alongside user praise for speed and tool use. Critical takes on reliability, math and vision limits pop up. Conversation covers video-to-code demos, agent swarm architecture, export controls, local deployment costs, and safety concerns about agentic systems.

A deep dive into OpenClaw’s origin, architecture, and why it can feel agentic. A close look at security risks from full-access agents and prompt-injection attack surfaces. Practical safety tips for hobbyists and cost trade-offs of always-on heartbeats. A debate on generalist agents versus specialized tools and whether consumers should wait for safer, cheaper options.

A dive into Moldbook and the strange social dynamics emerging around agent-run platforms. Topics include whether agent activity counts as real, what kinds of posts gain traction, and examples of agent-built workflows. Risks get attention too: prompt injection, exposed API keys, spam, and attempts at self-replication. The conversation ends on governance, private agent channels, and why this wave feels especially consequential.

Dario Amodei, AI researcher and Anthropic CEO, explains his essay on the adolescence of technology and what powerful AI might look like. The conversation covers timelines and uncertainty, risks of autonomous data-center systems, defenses like interpretability and monitoring, biological and authoritarian threats, economic disruption and concentration of power. Short, provocative, and forward-looking.

They roam through AI progress and practical LLM uses, from productivity boosts to code agents. They debate deepfakes, cybersecurity readiness, and the ethics of AI companions for kids. They unpack unit economics, chip geopolitics, and fundraising scale. They speculate on coordination versus race dynamics and wrap with lighter takes on vibe coding and AI humor.

A deep dive into how Claude’s Constitution should be written and amended. They debate why precise wording and heuristics matter. Risks from impersonation, prompt injection, and economic disruption get attention. Tensions between corrigibility, autonomy, and safety are explored. The conversation ends by listing open problems and why the current approach may not be enough.

Claude, an AI whose constitution outlines ethical aims, is presented as a principled, culturally aware moral agent. The conversation reads and analyzes Claude’s values, from radical honesty and compassionate truth-telling to hard constraints against causing mass harm. Short takes examine role limits, calibration of harms and benefits, and the hope for coherent long-term ethical convergence.

A deep dive into how a constitutional framework shapes an AI's values and behavior. Discussion of decision theory, negotiation via observable commitments, and why AI alignment is a philosophical test. Exploration of virtue-ethics training, core values and priority order for instructions. Concerns about harmful dependence and the balance between helpfulness and long-term safety.

A brisk tour through a hard sci-fi trilogy, covering first contact, interstellar warfare, and mind-bending tech. They trace sabotage of science, the dark forest deterrence idea, and dramatic strategic gambits. Political betrayal, failed saviors, and cosmic-scale weapons get spotlighted. The conversation ends on cultural tones and why the books feel bleak about human nature.