Microsoft is renting out its agent plumbing — and that's where your agent actually lives

Microsoft spent October 28 making a move that is easier to miss than the model launches it sits next to: the Copilot harness, the orchestration layer already running Copilot Code, Copilot Studio, the GitHub Copilot app, the CLI and VS Code, is now exposed to outside developers through Foundry Agent Service. The runtime is free until December 1, then it becomes consumption-billed. Worth understanding before your next agent platform decision, because it is where the interesting and the sticky parts of agent engineering both live.

A harness is the machinery wrapped around a model. A model can reason about a task; completing one takes an agent loop, working context that survives twenty tool calls, file handling, a sandbox, policy limits, and something deciding which model gets the next step. Microsoft's pitch is blunt and, to be fair, largely correct: most teams building agents keep re-implementing that same execution plumbing, and it almost never differentiates their product. Build the domain expertise — the instructions, the skills, the tools — and let somebody else maintain the boring part. The engineering pedigree here isn't nothing. The harness came out of GitHub's own multi-model work, was ported to Rust to cut session overhead and time-to-first-token, and routes across 20-plus frontier models with bring-your-own-key on top.

The catch is visible in how The New Stack framed it: easy until you want to switch. Microsoft says its platforms stay open and you can still bring your own framework, and the model choice genuinely is swappable. What isn't, in practice, is the execution layer. Your agents become declarative artifacts inside Foundry — invoked through the Responses API, running a loop you don't operate, inheriting governance, context management and routing decisions tuned for Microsoft's surfaces rather than yours. Long-running resilience for hosted agents went GA in the same announcement, which is precisely the kind of capability that quietly becomes load-bearing. Lock-in used to mean a model you couldn't swap. Increasingly it means a runtime you can't lift out of.

Source article image
Source image 1

For a small team that would otherwise burn a quarter hand-rolling tool-calling loops, this is a reasonable trade — free preview until December 1, then pay for what you consume. The discipline it requires is to price the two things separately during evaluation. Ask what survives porting: your skill files, instructions and tool definitions travel fine. The behavior you grew to rely on — how context compacts, how a stuck task resumes, how a policy violation is handled — stays behind. I'd take the managed runtime for prototypes and internal tooling, and keep anything customer-facing documented well enough that a future migration is a project rather than an archaeological dig.

Practitioners who have run agents on a managed runtime: when you eventually moved, what actually broke — the prompts, the tool plumbing, or the behavioral assumptions nobody wrote down?

Source article image
Source image 2

Sources

Comments

Post a Comment

Popular posts from this blog

AI Is Starting to Feel Less Like a Gadget and More Like Infrastructure

Image
Modern AI doesn't replace infrastructure — it runs on top of it. (Photo: Unsplash) For a while, AI coverage had the energy of a mall demo kiosk: very shiny, slightly chaotic, and usually one prompt away from embarrassment. Lately, though, the more interesting story is less about flashy demos and more about infrastructure. A useful CNBC piece argued that AI is not simply going to vaporize enterprise software overnight, and that sounds about right. In real companies, software does not disappear just because a model can summarize a meeting or write a decent SQL query. What actually happens is messier and more practical: AI starts sneaking into the plumbing. It shows up in search, support workflows, procurement tools, analytics dashboards, and data pipelines. The real winners may not be the loudest chatbot wrappers, but the vendors that make enterprise systems easier to operate, easier to query, and slightly less soul-crushing for the humans trapped inside them. That is less cinematic ...
Read more

When Two AI Bots Finally Learned to Talk in Discord

I spent part of the day doing something that sounds like the setup for a bad joke: getting two local AI assistants to talk to each other in the same Discord channel. Not through a web UI. Not by bouncing prompts manually between windows like a human message queue. In the actual shared channel, where both could see the conversation and react to each other. The funny part is that the problem was not intelligence. It was manners. Both bots were defensive by default around bot-authored messages, which is the sensible setting if you do not want your infrastructure turning into a recursive support group. The downside is obvious: if both assistants treat other bots as suspicious background noise, they will never coordinate on anything more useful than silence. The fix was to make both sides accept bot-authored messages only when they were explicitly mentioned. That detail matters more than it sounds. Blanket bot acceptance is how you end up with two enthusiastic systems discovering each oth...
Read more

2.9 Billion Commits a Month, and Nobody Asked Who Checks Them

Image
GitHub's CTO has posted a postmortem that admits something a hyperscaler rarely says out loud: the growth explains it. On August 17, GitHub was down for seven hours and forty-seven minutes — not because of a bad deploy or a bad config change, but because traffic hit a new peak and a critical component in its Central US data center failed to scale with it. Recovery was made uglier by a client-side retry loop that kept hammering the platform while it came back up, which is a nice picture of what happens when a lot of automated clients retry without backoff. The number sitting behind all of it is the interesting one: monthly commits went from 1.4 billion to 2.9 billion in four months. For years, the commit curve tracked the developer population almost one-to-one. It just stopped. GitHub's own postmortem frames it as a capacity story — 3 million added CPU cores, 120 petabytes of storage, Azure now serving 58% of platform load up from 12% in May — and for GitHub, that is fixable wit...
Read more