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-lubken-software-of-one/references/

Source outline

Read transcript.md for evidence. Promotion cues 230–240 are omitted from the bundled transcript.

TimeCuesTopic
00:00–01:131–18A small program instead of excessive agent prompting
01:18–03:2619–50Product context and software-of-one definition
03:31–05:0551–75Personal app examples
05:11–07:1076–107Inputs and the reported 2,500-app organization
07:16–09:46108–150Desired stack qualities and exploratory scope
09:49–13:30151–205JavaScript, Durable Objects, celld and compute claims
13:33–15:10206–229Nearby SQLite and durable state
15:56–19:02241–287Gates, throughput and napkin cost estimate
19:06–23:39288–349Access-control architecture
23:43–25:46350–379Usage, value, quality and overlap insights
25:49–27:58380–413Rebuild/promotion and stable backends
28:01–32:13414–471Agent-built interfaces and closing principles

Short quotations

  • “Software built for one task.” — excerpt, cue 37, 02:34.
  • “cheap to operate.” — excerpt, cue 139, 09:06.
  • “Make sure that these apps are cheap to build, to run and retire.” — cue 449, 30:48.

Limits

The stack reflects the speaker's preferences. The 2,500-app report is secondhand. Costs are a napkin estimate with partial assumptions, not total operating-cost evidence. Throughput and cold-start figures are not workload-independent guarantees. Acquisition, open-source commitments and product pricing are event-time accounts. Async approval is conceptual; it does not authorize access before a decision.

skills

talk-lubken-software-of-one

tile.json