- Published on
SF October 14th: A Birds of a Feather Session on Agentic Engineering
- Authors
- Name
- DeepHorus
An evening gathering dedicated to agentic engineering is scheduled for Wednesday, 14th October in San Francisco, announced in a Link blog post published on 23rd September 2026. The Birds of a Feather session will be co-hosted by the blog's author and Jesse Vincent. It is aimed squarely at people who are building "weird and interesting things" with, and on top of, coding agents. The organizers are framing the night as an informal forum rather than a conference track or a product launch venue.
The event comes amid unusually rapid movement in agentic AI tooling. Recent Link posts have covered Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna and an emerging price war on 22nd September, Jev's System One "Decision Models" on 21st September, and route generation with GPT-6 Astra and ChatGPT Work on 12th September. That churn has left developers with plenty of unresolved questions and half-formed approaches. A room of peers trying to make sense of the same fast-shifting landscape is precisely what the hosts say they want to convene.
The format is deliberately lightweight: one flowing conversation with an informal show-and-tell component. Sharing something you are working on is encouraged, but no presentation is required from attendees. The emphasis is on comparing notes with other builders and experimenters about what people are attempting, what they are learning, and what they still have not figured out. In other words, the discussion is expected to run toward unfinished work rather than polished demos.
The hosts have stated a specific interest in material that has not been discussed publicly: odd experiments, dead ends, and unfinished projects that do not have an obvious market. That invitation separates the session from typical developer meetups, which often showcase completed features or funded products. By asking participants to bring work without a clear commercial case, the organizers are treating exploratory value as its own justification.
Importantly, this is not a venue for product pitches. The organisers describe the gathering as occupying a much earlier stage than that, where ideas are still rough and directions unproven. The stated goal is to "celebrate and lean into that weirdness" of agentic AI rather than smooth it into a conventional narrative. That framing suggests candid, technically grounded exchanges over scripted talks.
For attendees, the value likely lies in cross-pollination: techniques, failure modes and tooling tricks that rarely surface in public write-ups. Sessions built around shared uncertainty tend to surface the practical problems that documentation omits. Participants can expect to leave with leads on problems others are already chasing.
Looking ahead, the 14th October session could become a recurring touchpoint if the community responds. Any insights gathered may well filter back into public discussion, and potentially into future Link posts covering the same fast-moving models and tools. Registration details and follow-up coverage are expected as the date approaches.
