Agent demos depth EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Cameron Rohn emphasizes that there’s a wealth of information in the agent demos, not just a single showcase. Uncertain AI future EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Tom Spencer notes a general sense of uncertainty regarding the future of certain AI developments, with speculation continuing without clear answers. Show more OpenAI Sora de-prioritization EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Cameron Rohn suggests OpenAI is not prioritizing Sora likely due to challenges accessing training data, though the exact reasons remain uncertain. Show more Benchmark Variability Concerns EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Cameron Rohn points out that image model generation showed promise on certain benchmarks, but there remain unclear variables affecting those benchmarks. Show more Post-Diffusion AI Community EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Cameron Rohn observes that after the initial release of diffusion models, a significant AI community emerged, fostering increased professionalism in the field. Show more Iterative Experimentation Process EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Tom Spencer notes that his experimentation with video and image generation models requires many iterations, wondering if this stems from his perfectionist tendencies. Show more Micro Evaluations Wasted Effort EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions Performance - 6/27/2025 Cameron argues that micro evaluations are often not helpful when trying to assess performance. Tool Avoidance to Adoption EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Tom admits he intentionally tried to avoid these tools but now sees legitimate uses for them after the discussion. Tech Should Drive Compliance EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Both speakers agree compliance innovation should originate from and be led by the tech industry rather than accounting. SOC 2’s Accounting Roots EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Cameron Rohn was surprised to discover that SOC 2 and ISO standards originate from accounting practices, with most participants having financial backgrounds. Show more Regulators Lag Behind EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Tom Spencer notes that regulators often lag behind and lack technical expertise, leading to extensive time answering compliance questions. Show more Teams Beyond Product EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Cameron Rohn emphasizes that a team includes processes and how people behave, not just the product, and suggests agents could simplify tasks. Show more Agentic AI Replacement EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Cameron Rohn awaits a solution providing an agentic AI replacement for existing tools and is unclear why the software industry has not addressed this need. Show more Automated Security Perspective EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions Authentication - 6/27/2025 Tom emphasizes the value of using security tools that scan repositories for vulnerabilities and automatically apply fixes. Show more GitHub Integration Concern EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions Testing - 6/27/2025 Cameron notes that tight, effective GitHub integrations in coding workflows expose a competitive landscape. Distinct Tool-Calling Skills EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Cameron notes that while eightway contributes to MCP, it also brings its own tool-calling and skills framework. Protocols Are Complementary EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Tom clarifies that eightway and MCP are not separate but complementary technologies in multi-agent systems. Latency and Error Focus EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions Performance - 6/27/2025 Cameron highlights that the key difference of the eightway protocol is its design for low latency and minimal error in agent-to-agent communication. Show more Similarity to Claude Code EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions - - 6/27/2025 Tom reports that his initial install of the eightway implementation looks very similar to Anthropic’s Claude code. Waiting for Agent Space EP 5 - The Build - Agent Architectures: The Next Frontier in AI Opinions AI Development - 6/27/2025 Cameron is holding off on deeper experimentation until the team releases their promised agent space. 