Coding-agent talks from AI Engineer: a guided collection

Choose a talk for the decision in front of you: aligning a team, understanding a repository, checking generated code, or operating agents across an organization. This editorial guide selects nine talks from the Coding Agents playlist curated by AI Engineer and organizes them by question rather than viewing order. Each speaker offers a distinct workflow, demonstration, or proposal.

Where to start

1. How should a team align before faster implementation creates more review work? Maggie Appleton(opens in a new tab), a GitHub staff research engineer on GitHub Next, argues that product, business, design, and organizational context must become shareable before implementation rather than being recovered in a pull request 01:03–06:49(opens in a new tab). Her ACE demonstration is a pre-preview research prototype, not evidence that shared agent workspaces have solved coordination 06:49–12:47(opens in a new tab).
2. How can an agent help you understand an unfamiliar repository? Priscila Andre de Oliveira(opens in a new tab), a senior software engineer at Sentry, turns recurring architecture, convention, feature, syntax, test, and history investigations into a local comprehension skill 09:36–13:23(opens in a new tab). Her own session sample and prepared demonstration support a workflow example, not a general accuracy result; developers still need to understand and steer the research 13:23–15:26(opens in a new tab).
3. Which deterministic feedback belongs inside the agent loop? Daniel Szoke(opens in a new tab), the Rust SDK maintainer at Sentry, shows how compiler-enforced type, memory, null, and concurrency constraints can reject specific invalid programs and return actionable errors to an agent 09:26–15:13(opens in a new tab). He does not treat compilation as proof of intended behavior; tests and review remain separate checks 04:59–05:50(opens in a new tab).
4. How much should a coding-agent harness control? Mario Zechner(opens in a new tab), creator of Pi—the source supplies no employer affiliation—argues for a small, observable, user-extensible harness and for bounding generated work by task scope, evaluation quality, and human review capacity 05:16–10:32(opens in a new tab), 15:53–17:41(opens in a new tab). His comparisons reflect personal experience rather than controlled harness evaluation, and Pi's malleability leaves baseline security policy to the user 00:54–04:27(opens in a new tab), 06:19–08:49(opens in a new tab).
5. What does an end-to-end human-led delivery loop look like? Matt Pocock(opens in a new tab), who introduces himself as a teacher and provides no employer affiliation, moves from alignment and a PRD through dependency-aware vertical slices, visible implementation, fresh-context review, manual QA, and branch integration 28:53–35:09(opens in a new tab), 39:38–57:48(opens in a new tab), 1:04:23–1:22:13(opens in a new tab), 1:28:02–1:32:23(opens in a new tab). The workshop's generated increment passes automated checks before manual QA finds a missing table, making the limits of any single proof layer concrete 1:08:57–1:12:21(opens in a new tab).
6. What changes when agents operate across an organization? Ian Ballantyne and KP Sawhney(opens in a new tab) of Google DeepMind describe quotas, model fallback, curated skills, shared workspaces, trajectory observability, task-specific evaluation, and automated review as operating requirements rather than optional polish 04:03–08:55(opens in a new tab), 12:37–24:00(opens in a new tab). Several systems remain internal or exploratory, and the talk does not provide a comparative reliability result.
7. How should a company staff adoption and inspect its claimed gains? Brian Scanlan(opens in a new tab), a senior principal engineer in Intercom's platform group, presents adoption as leadership-backed change management with enablement staff, shared guidance, testable skills, telemetry, and review controls 02:56–07:41(opens in a new tab), 08:45–13:28(opens in a new tab). His reported throughput increase coincides with stronger models and uses activity-heavy measures, so review capacity, defects, and customer outcomes still need separate measurement 15:18–18:58(opens in a new tab).
8. How do you distinguish attention from durable product use? Quinn Slack(opens in a new tab), a Sourcegraph co-founder, argues that subscriptions, survey interest, suggestion exposure, and sustained active use are different signals; he favors frequent workflows with low verification cost and explicit correction paths 01:03–04:12(opens in a new tab), 13:43–16:54(opens in a new tab). His market estimates mix public figures, private information, and intuition, and Sourcegraph has a commercial stake in the adoption claims 08:28–12:40(opens in a new tab).
9. How do you turn an agent session into a durable workstream? Jason Liu(opens in a new tab), who says he works at OpenAI, combines named project threads and persistent context with scheduled heartbeat messages, observable goal checks, and explicit stopping criteria 03:02–07:11(opens in a new tab), 34:49–41:08(opens in a new tab), 1:11:47–1:13:46(opens in a new tab). These are evolving personal workflows rather than controlled reliability evidence: Liu notes unusually generous token access, difficult-to-reproduce connector workflows, and unresolved cross-project memory scoping 23:42–26:53(opens in a new tab), 1:01:36–1:04:29(opens in a new tab).

Related summaries

Appleton, Zechner, and Liu also have dedicated talk summaries in this resource hub. Beyang Liu's related Sourcegraph/Amp talk(opens in a new tab) provides another view of detailed prompts, explicit context, review, subagents, and selective parallelism 11:20–33:38(opens in a new tab).

Watch the full Coding Agents playlist(opens in a new tab).