CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

ainativedev/aidevcon-2026-nyc

AI Native DevCon 2026 New York — practical agent skills grounded in conference talks

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

outline.mdskills/talk-sharma-mohapatra-evolving-intent/references/

Source outline

Source: Akhil Sharma & Sudipta S. Mohapatra, The "Evolution of the Intent", AI Native DevCon New York 2026. Recording. All anchors refer to original numbered cues in the transcript. Short quotations below preserve the supplied transcription, including spoken wording.

TopicSource anchorExact short quoteApplication and scope
Intent definition00:00:57–00:01:55, cues 15–25“CreateTaxiReservation”The taxi example maps language to a specific capability.
Legacy constraints00:03:57–00:06:17, cues 51–85“the schema of the entities was fixed.”The described legacy flow uses fixed grammar and entities.
Legacy safeguards00:06:24–00:07:41, cues 86–103“without authentication.”The source already recognizes protection before spending.
Tool descriptions00:07:50–00:09:15, cues 105–126“you can describe what, the function is doing in English itself,”Descriptions help the agent choose a tool.
Structured arguments00:10:00–00:11:49, cues 138–160“pickup, drop-off,”Tools retain structured input requirements.
Changing destination00:11:52–00:12:30, cues 161–170“the actual cab ride is supposed to go to Bronx,”The example illustrates following a correction.
Ambiguous flights00:12:35–00:14:39, cues 171–203“I found two trips.”The illustrative flow asks the user to disambiguate.
Dependent recalculation00:14:42–00:15:57, cues 204–222“mid flow correction”Changed intent should update the ongoing plan.
Consequential pause00:16:01–00:17:39, cues 223–243“asking, hey, do you want to confirm”The example pauses before a money-committing action.
Trade-offs00:17:43–00:19:06, cues 244–261“the main thing of giving up is determinism,”Interpretation and latency become less deterministic.
Broader application00:19:14–00:20:36, cues 262–279“an MCP layer”HR and MCP are discussed as extensions, not fully specified integrations.

Source limitations

  • The taxi workflow is illustrative; the source is not evidence that the named providers expose these exact tools or support this workflow today.
  • The transcript's interrupt terminology is imprecise and does not establish a runnable framework API.
  • Conversation-as-state does not prove durability, recovery, or external transaction correctness.
  • No quantified latency, reliability, or implementation-effort comparison is substantiated by a reproducible evaluation in the transcript.
  • Promotional interlude cues 127–137 are omitted from the bundled transcript; the remaining cue numbers retain their source positions.
  • This is a spoken transcript; unseen slides, linked repositories, and third-party research were not independently inspected. No outcome beyond the described talk is established.

skills

talk-sharma-mohapatra-evolving-intent

tile.json