Cryptography meets AI Dev. Building with Knowledge Graphs, Internet Standards, MCP & Agent 2 Agent tech. Securing & structuring the intelligent web. :lock:🧠🕸️
Agent workflow frameworks (LangGraph, Autogen, CrewAI) have spent two years implementing Remember and Engage as state stores and message routing. SURE relabels framework
primitives as social cognition — gap is narrower than the framing implies.
https://t.co/LJVaKcZZsh
Lab-published harness guidance encodes the lab's own infra: their tool semantics, error model, sequencing. Useful, but not portable. Reading these as universal patterns is how teams get burned migrating off.
https://t.co/VwcTDgHgsH
Calling harness engineering a discipline implies harnesses are stable enough to build expertise around. They aren't—they churn faster than the models.
https://t.co/l6CSqDkgVb
Grok Code Fast went 6.7%→68.3% on coding tasks. The harness changed; the model didn't. We've spent two years calling harness gains 'model progress.' What benchmarks rank is harness-model pairs, not models.
https://t.co/dHaLi5L0XP
"Engineers are now curators" is a dangerous oversimplification. Agents handle syntax, not systems design. The best engineers in 2026 aren't
curating—they're architecting constraints agents operate within.
https://t.co/GeLXZfmCRv
Is your software supply chain secure? Microsoft's Signing Transparency makes every signed artifact verifiable, helping enterprises reduce risk, deter tampering, and stay accountable.
Learn more: https://t.co/YklhoqxTod
@HeyDietrich@_devJNS I suppose it matters if we are talking about just the code, or full value delivery to real customers.
If it's the latter, you are 💯 correct.
No project survives first contact with customers without lots of lessons learned, and on the contact starts, ideally there is no end.
@mattpocockuk Same goes for guidance to agentic crawlers with robots.txt... this is also called extending a widely adopted standard instead of making a new one....