Thinking
Ideas On Deep Tech, Software Work And Adoption
Ideas is a structured place for analysis, research notes and field reflections. It is not a generic blog; it tracks the themes behind advanced technology adoption.
Editorial Tracks
- AI and software work: how agentic software, AI-assisted development and autonomous systems change engineering practice.
- Quantum computing and security: implications for AI training, cybersecurity, governance and future architectures.
- Organizational adoption: change management, internal capability, technical and executive training, and the shared language required for decisions.
- Deep tech entrepreneurship: how technical depth, market timing, education and execution interact when building in complex domains.
Ideas
AI in School Should Be Assessed, Not Pretended Away
Schools should protect deep reading and handwriting while training teachers to assess how students use AI, so artificial intelligence supports learning instead of replacing it.
5 Principles for Working With AI Support
Artificial intelligence can help people enter unfamiliar domains when they use it to surface risks, options, missing context and better questions, not only to execute tasks.
Conversational Agents Need Service Design, Not Just RAG
Customer service conversational agents need low-latency voice interaction, barge-in, correct escalation, tool calling and omnichannel service design, not only a RAG layer over support documents.
Junior Developers Must Own The Code Agents Write
In agentic coding, code review verifies the code. Organizations also need ways to verify whether junior developers understand and can own the code produced with coding agents.
5 Reasons We Are Moving Coding Agents Toward Self-Hosted AI
Why a software company should treat coding agents as production infrastructure and explore self-hosted AI for availability, cost control, governance and internal capability.
AI Adoption Training Is Not About Tools. It Is About Organizational Trust.
AI and digital transformation training in large organizations becomes useful when people connect tools to their work, trust that change is possible, and build internal capability through practice.
What If LLMs Could Simply Say I Don't Know?
A note on why uncertainty handling matters when AI systems move from isolated tools into software workflows, decision processes and organizational capability.
Discuss research, talks or collaborations around deep tech adoption.
Explore A Conversation