The gap in Vilius Vystartas’s essay is not a gap in intelligence. It is a gap in the work each person is currently immersed in. A SharePoint specialist can celebrate Markdown support as real progress while the author, after a long weekend of building an autonomous agent environment, sees that same feature as a routine improvement. The harder lesson in his account is that an agent ecosystem has to survive its own operating environment before it can do anything autonomously worth trusting.
The essay was published on May 4, 2026, and is available through a syndicated listing of the DEV Community article. The author’s own summary of the argument also appears in a LinkedIn post. The original DEV page was not retrieved directly, so the details below come from the indexed text and that post.
As an Amazon Associate I earn from qualifying purchases.
What the essay actually argues
Vystartas recalls watching a SharePoint MVP celebrate Markdown support. He does not dismiss it. For SharePoint users it is meaningful progress. What interests him is how quickly his own frame of reference moved. After spending time building autonomous agents, a feature that would have looked like a frontier only a short while ago now reads as ordinary platform housekeeping. His question is how quickly yesterday’s frontier becomes today’s routine, and what that shift hides.
The essay is reflective commentary. It is not a controlled study, and it does not measure productivity across teams. Its value is in describing what operational agent work feels like from the inside.
#1 Best Overall
Why the same feature looks routine from the other side
The gap is a difference in focus. A platform specialist spends time on the features a platform ships and how users adopt them. A builder of agent infrastructure spends time on whether the pieces connect, keep running, and behave predictably when nobody is watching. Both views are accurate. They are simply pointed at different problems, which is why one person’s breakthrough is another’s Tuesday.
That framing is the essay’s strongest point, and it matters for anyone choosing where to spend learning time. The features that feel mature to one group are often the parts that still need the most engineering to rely on in production.
The grind: what broke
The practical center of the essay is a list of failures the author reports from his own build. He presents them as experience, not as a survey of how often each problem occurs across projects. They are useful because they show how many layers an agent touches before it does any real work.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteEnvironment and permissions
- macOS permissions blocking expected actions
- Gateway restarts needed to bring components back into a working state
- Configuration problems spread across the setup
Build and tooling
- A mismatched SCSS build toolchain
- CLI flags that were silently ignored
- Naming conventions that caused components to fail to connect
- C++ modules that would not work under Node 22
Agent behavior
- Agents that looped instead of finishing a task
- Sub-agents that timed out
- Repeated patches to the memory layer
Performance
- Slow inference, which the author associates with a thinking-mode setting
Taken together, these are not exotic bugs. Most of them are the ordinary friction of connecting tools, permissions, build systems, and model calls. The essay’s argument is that this friction is the engineering work, and it is easy to underestimate from the outside.
What the author reports the system eventually did
Vystartas says the system eventually handled text and images without constant supervision, produced usable first drafts, enforced standards, and kept audit trails. He also reports that agents scaffolded 111 web parts and five backend services during a long weekend.
These are his figures and his observations about his own build. Nothing in the material establishes independent measurements of output quality, autonomy, or audit completeness. The counts do not show that a few days of agent work replaces months of human work, and they should not be read that way for typical development teams.
What Microsoft’s documentation says about SharePoint agents
The essay does not evaluate Microsoft’s platform, but the documentation gives useful context for the reliability question. According to Microsoft’s guidance on managing access to agents in SharePoint, an agent answers based on each user’s permissions to its data sources. If a user cannot access a referenced site or library, the response does not include that restricted content for that user. Microsoft also documents controls over who can use an agent, what information it can reach, and where it is available.
Microsoft’s Agent 365 integration with SharePoint Online and OneDrive describes agent access insights, permissions reports, and controls for restricting external sharing and site access. These are documented capabilities. They do not show that a given organization has turned them on, and they do not prevent every runtime failure. Check availability and licensing for your tenant before relying on any of them, because these details change.
Best Value
A governance checklist for agent reliability
Microsoft’s governance guidance supports a practical point: agent reliability depends on identity, data, action boundaries, oversight, and observability, not only on prompting. Microsoft’s guidance on governing agents by risk recommends calibrating oversight to risk and distinguishing agents that assist a person from agents that execute changes in a system. Its guidance on how enterprises control what agents can do frames tool permissions and actions as central to what an agent is able to do.
| Dimension | Question to answer before automation | Documented Microsoft capability |
|---|---|---|
| Agent identity and access | Who can use the agent, and which users can reach it? | Access controls for agent use, per the SharePoint access guidance |
| Data permissions | Does the agent respect each user’s rights to sites and libraries? | Permission-sensitive responses, per the SharePoint access guidance |
| Tool and action permissions | Which tools can the agent call, and what actions can it take? | Tool-governance guidance on controlling what agents can do |
| Assistive versus executing | Does the agent only suggest, or does it change systems? | Risk-based guidance that separates the two, per the risk guidance |
| Human oversight | Where does a person approve or stop the work? | Oversight calibrated to risk, per the risk guidance |
| Audit and observability | Can you see what the agent accessed and did? | Access insights and permissions reports, per the Agent 365 integration documentation |
The table lists what Microsoft documents as capabilities or guidance. It does not establish that any particular tenant has configured these controls effectively.
What this does and does not establish
The essay supports a narrow but useful conclusion: operational resilience is a large part of agent engineering, and people who spend their time there can reasonably find platform announcements unremarkable. It does not establish how common the author’s failures are, how much time they cost across projects, or whether his autonomy and audit results would hold in another environment. No independent study was identified that supports the essay’s broader claims about productivity.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11The Markdown announcement and the agent build are both real. The distance between them is the work of making autonomous systems dependable, and the essay is most convincing when it describes that work honestly and specifically.
The Bottom Line
The “agentic gap” is a gap in focus, not in capability. Autonomous agents depend on permissions, build tooling, configuration, and oversight working together, and the author’s account shows how much of that work sits underneath any visible feature. Treat his counts and results as his own report, and check Microsoft’s current licensing and availability details before applying any of the governance controls.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
