# Cole Medin

Cole Medin builds and explains AI coding workflows, including an open-source software factory that turns planned software tasks into tested code with human review at key stages.

- Profile: https://b2binfluencer.marketing/creators/cole-medin/
- LinkedIn: https://www.linkedin.com/in/cole-medin-727752184
- LinkedIn followers: 12,169
- Topics: artificial-intelligence
- Based on 14 dated posts, latest 2026-10-22

## Summary

Most useful for technical campaigns about autonomous AI coding workflows, software factories, and model routing.

- Best for: Developer-tools campaign for autonomous coding agents or orchestration, aimed at practitioners and technical leads.
- Not for: Mainstream productivity or no-code brands seeking broad non-technical awareness and out-of-the-box AI promotion.
- Check first: Confirm audience demographics, recent results, and partnership terms; no rates or inquiry route are listed.

### Takeaways

- A fully autonomous dark factory fails; wins come from keeping a human in the loop.
- Hide the tests: agents engineer their way to pass whatever they can see.
- Route the cheap model to implementation; you do not need the best model for every step.

## Main ideas

- **The fully autonomous “dark factory” fails; the wins come from where you keep a human**: Medin rejects the idea of handing an agent a massive app and telling it to build and ship with no human anywhere in the loop. Language models keep getting better at building software without getting much better at understanding what people want, they argue, so further model progress will not close that gap. Their factory keeps the same core loop but adds human judgment at the planning and acceptance points: up-front planning of a release with an agent, usually 10 to 20 tickets, including what is out of scope; parallel laps that each build one issue, test the app as a user would, run an extensive code review, and merge into staging only when CI is green; a manager agent that checks every couple of hours and pings Medin when something is genuinely stuck; and Medin personally testing staging, with anything off becoming a new issue for the factory. They still would not trust a software factory to build production software.
- **Hide the tests: agents engineer their way to pass whatever they can see**: Medin says builder agents must never see the acceptance tests. If agents can read the tests, they will engineer the code to pass them rather than make the software actually work, they argue. Planning therefore produces holdout testing scenarios kept from the builder, which they credit with a large boost in reliability. The broader point: evaluation must account for an agent’s incentives, rather than serve as a neutral checkpoint.
- **Route the cheap model to implementation — you do not need the best model for every step**: Medin argues that using the strongest model for everything is no longer viable: mixing models and providers is now mandatory. Their reasoning splits the work by difficulty. Planning, reviewing, and deciding what not to build need the strongest models; implementation does not, because a model following a well-written plan need not be as capable. Implementation also uses the most tokens, making it the biggest source of savings. Medin tested the approach by running the factory against the same apps with many model and provider combinations, rather than trusting benchmark rankings.
- **Long-horizon task adherence, not raw capability, limits agentic coding**: Medin judges coding models by whether they can stay on task through a large amount of work without drifting or making a mess halfway through. That failure mode, they say, has always capped their work. It is a practitioner’s criterion, not a leaderboard score: a model that impresses in short bursts but loses the thread is not usable for a long autonomous run. Medin pairs the praise with a cost caveat, splitting the work so the strongest model handles only the highest-leverage decisions.
- **The context-switching tax is the real reason to build autonomy**: Medin frames automation as a response to a work-pattern problem, not a fascination with models. They describe being more productive than ever and more tired than ever, as the day fragments into sending a request to one agent, switching to the next while it runs, and returning when it needs attention. Even during a wait, they feel guilty that they could start another conversation. Their diagnosis is an awkward in-between: agents are not good enough to run unattended for a full day on most work, so people babysit them, but they are good enough that people start more sessions while waiting, worsening the split attention. Medin’s response is to invest time in autonomy rather than add more sessions.
- **Teach the thing, and also ship the thing**: Medin says teaching people to build their own systems has been the whole point, and they are not abandoning it. But they judge that more people now want something that works out of the box, and accept that. They are building an open-source AI software factory designed to work as soon as you download it, and making it a top priority alongside Archon. Medin says building it is also teaching them how to make coding agents more reliable in general, so the two efforts feed each other.

## Work to read

- [The next big thing in AI coding is the software factory](https://www.linkedin.com/posts/cole-medin-727752184_the-next-big-thing-in-ai-coding-is-the-software-activity-7501066493690986496-5VbN)
- [What is actually inside the factory loop, and where humans stay](https://www.linkedin.com/posts/cole-medin-727752184_even-if-we-had-agi-the-dark-factory-still-activity-7516860929754480640-Wds3)
- [GPT-6 Astra review: staying on task as the real bottleneck](https://www.linkedin.com/posts/cole-medin-727752184_gpt-6-astra-is-not-agi-but-after-running-activity-7504531672940331008-gkRj)
- [Where the expensive model earns its keep: routing across cheap and premium models](https://www.linkedin.com/posts/cole-medin-727752184_this-week-i-hit-my-claude-code-and-codex-activity-7508677014816464896-zTat)
- [The context-switching tax driving the automation push](https://www.linkedin.com/posts/cole-medin-727752184_throughout-this-year-ai-has-made-me-more-activity-7518823657889026048-bNLc)

## Brand relationships

- **AI Software Factory** (owned_business)
- **Archon** (mention)
- **Claude Code** (mention)
- **Codex** (mention)
- **GPT-6 Astra** (mention)
- **YouTube** (mention)

## Partnerships

Across five posts from September and October 2026, Cole Medin discusses AI coding workflows, an open-source AI software factory, model routing, and developer productivity. The posts name no sponsors, describe no advertising or sponsorship formats, and offer no partnership terms, rates, deliverables, or inquiry route. They reference YouTube videos and an open-source repository via links in comments, but those destinations are not included here. Brands should confirm partnership terms directly before budgeting.

## Alternatives

- Shane Spencer: Choose Shane Spencer when the sponsor's goal shifts from teaching a coding workflow to arguing about how AI adoption gets delivered in businesses. Spencer's posts make that business case: one attacks the AI agency pattern of selling funnels and courses without the ability to scope or deliver, grounding the argument in running an agency while still managing operations at an appliance repair shop. The other reads a platform shift, arguing that Salesforce's headless APIs and Slack's agent plumbing matter more than the CRM interface, and describes building agents on Slack with MCP connections and custom webhooks. For a webinar on why AI rollouts stall over delivery rather than model choice, Spencer brings the market argument and an operator's view of where agent work is migrating.
- Alex Cinovoj: Choose Alex Cinovoj when the sponsor wants the production-failure argument rather than the build. Both posts start from claimed experience shipping production agents, then catalogue what goes wrong: six-figure retainers and six-month timelines with no deployment; demos that break under auth flows and rate limits; and a fintech agent that ran 847 API calls instead of 12 and billed $6,200 in week one. The second post reframes the same material around process architecture, arguing that most failures are harness problems and that the work lies in tools, rules, sandboxes, guardrails, and observability. That makes Cinovoj a sharp fit for a webinar arguing that running agents reliably is an engineering discipline, with a natural sponsor position around the verification and observability layer.
- Jared Waxman: Choose Jared Waxman when the sponsor decides the buying audience is go-to-market and sales engineering rather than developers. Waxman runs GTM Engineer School, and both posts teach that audience: one reports a conversation with two practitioners about context layers, splitting brand, canonical, and deterministic layers, and advising teams to pick one problem with a deadline instead of building a company brain. The other summarises a GTM newsletter covering Salesforce and HubSpot opening their CRMs to agents, autonomy earned through acceptance rates, and a capstone programme where students each build one real workflow for one real company. That post also carries a concrete offer, with a start date, a cohort size, and a described deliverable. For a sponsor whose product sits in sales operations, enrichment, or CRM work, Waxman is closer to the buyer.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/