Skydance, as it begins integrating Paramount and Warner Bros. Discovery, is offering voluntary buyouts to some employees and imposing a five-day-a-week office requirement. Layoffs are also expected as the company targets more than $6 billion in annualized cost savings within three years. Variety reports: The new program was outlined by chairman and CEO David Ellison and co-CEO Ynon Kreiz in a memo to staffers Monday. "As we bring our teams together, we recognize that not everyone will see their future here, and we respect that. To give eligible employees the option to move on in a way that works for them, we are introducing a Voluntary Buyout (VBO) Program," the execs wrote. "Those who qualify will receive an individual communication with their personalized details and terms, and a window to decide whether to accept this one-time offer."
They added, "Please know we don't take any of this lightly. We are clear about what we're building: a company that tells the stories people around the world love, embraces new ways to connect with audiences, and leads the next era of media and entertainment. We have the talent, the creative firepower and the ambition to do it, and we will get there together."
[...] In the Oct. 26 memo, Ellison and Kreiz also said the new Skydance is moving to a mandatory five-days-per-week in-office policy across the company. "We believe in-person collaboration is vital to our culture and to the success of our business. Especially in a creative company like ours, the best ideas come from people in the same room, sharing perspectives, challenging one another and solving problems," the execs said. "We truly believe this is in the best interest of our company and our people. We've seen the benefits firsthand at Paramount, and we want every part of Skydance to share in them."
As you may have heard, loops are the hottest thing in AI. The buzz is that rather than you writing prompts, you let your agent do so. The problem is that while all the videos I’ve seen are great at touting the advantages and even explaining the moving parts (though each one has slightly different architecture), I’ve yet to see one that gives concrete C# examples—let alone using the Microsoft Agent Framework.
Soooo, this is the very first in a series. I’m finding my way, stumbling in the dark, so some of this may be a bit simplistic—but I believe it is correct as far as it goes. Remember, however, that the first draft of this post was written by AI. I’ve edited it, but your mileage may vary.
Here’s a thought experiment. Imagine having an agent that keeps trying fixes until a web service stays healthy—retrying configuration changes, re-checking status, and stopping only when a health check returns OK or a safety limit is reached. That behavior—automatic, bounded repetition with an evaluation step between attempts—is exactly what loop programming enables in modern agentic AI. This post explains the pattern, why it matters, and how to implement it in C# using the Microsoft Agent Framework (MAF) concepts.
One-sentence takeaway Loop programming = agent acts → observes → evaluates → repeats (until accepted or bound). Use MAF’s loop primitives (loop wrapper or middleware + evaluators) to implement this pattern safely in .NET.
Why loop programming matters
Iterative refinement: reduce hallucinations by asking the model to rework outputs until they meet quality criteria.
Automation & robustness: automatically retry flaky tools with idempotency and backoff, or orchestrate multi-tool workflows until success.
Cost/latency tradeoffs: each iteration consumes time and tokens—a deliberate loop strategy balances accuracy and cost.
Core MAF loop terms
Base agent / client: This is the model wrapped as an agent (an abstraction that receives requests and produces responses).
Loop wrapper or middleware: A piece that re-invokes the base agent automatically, applying evaluator logic between iterations.
Evaluator: A component that inspects the previous response and returns Continue or Stop. Examples:
Marker-based evaluator: stop when the output contains a specific marker (e.g., “DONE”).
Judge evaluator: ask an auxiliary judge model to say whether the output is acceptable.
Custom evaluator: your own logic that checks outputs, tool results, or external state.
Bounds: MaxIterations, timeouts, and cancellation tokens—mandatory to avoid runaway loops.
AgentRequest: the input to an agent (prompt, tools, options).
AgentResponse: the output (text, structured data, tool results).
Minimal, runnable-style C# example Below is a full Program.cs and a simple csproj. The example uses a completion-marker evaluator (“DONE”) and a MaxIterations bound. It is a bit brain-dead, but it will give you a sense of how to string one of these together.
using System; using System.Threading; using System.Threading.Tasks;
using Microsoft.Agents.AI; using Microsoft.Agents.AI.Evaluators; using Microsoft.Agents.AI.OpenAI.Clients;
class Program { static async Task Main(string[] args) { // Read API key (set as env var) string? apiKey = Environment.GetEnvironmentVariable("OPENAI_API_KEY"); if (string.IsNullOrEmpty(apiKey)) { Console.WriteLine("Set OPENAI_API_KEY environment variable and re-run."); return; } // 1) Create provider client (provider-specific) var openAiClient = new OpenAiProviderClient(new OpenAiOptions { ApiKey = apiKey });
// 2) Convert provider client to an agent abstraction AIAgent baseAgent = openAiClient.AsAIAgent();
// 3) Create a simple marker evaluator that stops when "DONE" appears var evaluator = new CompletionMarkerLoopEvaluator("DONE");
// 4) Wrap the base agent with loop behavior and bounds var loopOptions = new LoopAgentOptions { MaxIterations = 5 }; AIAgent loopAgent = new LoopAgent(baseAgent, evaluator, loopOptions);
// 5) Prepare request var request = new AgentRequest { Prompt = "Summarize this paragraph and keep refining until you append DONE:\n\n" + "The quick brown fox jumps over the lazy dog. This sentence includes all the letters of the English alphabet." };
// 6) Use a timeout for safety using var cts = new CancellationTokenSource(TimeSpan.FromSeconds(30)); try { AgentResponse response = await loopAgent.InvokeAsync(request, cts.Token); Console.WriteLine("Final response:\n" + response.Text); } catch (OperationCanceledException) { Console.WriteLine("Operation timed out or was canceled."); } } }
How to adapt this to the real MAF packages
Install packages
dotnet add package Microsoft.Agents.AI
dotnet add package Microsoft.Agents.AI.OpenAI
The logical steps are always pretty much the same: build a base agent, choose an evaluator, apply loop wrapper or middleware, and call it.
Designing better evaluators (AI judge and custom evaluator)
Marker evaluator (simple)
Pros: deterministic and cheap.
Cons: brittle.
AIJudge evaluator (higher quality)
Pattern: after each iteration, call a small judge prompt that returns YES/NO and a short reason. Use a very explicit evaluation prompt to reduce ambiguity.
Example judge prompt (send as a separate small completion request): “You are a strict grader. Given the task: [task description]. Given the model response: [response]. Is the response acceptable? Answer EXACTLY YES or NO followed by a short reason.”
Custom evaluator pattern (checks content + side effects)
Inspect response text for task markers (like DONE) and for required structured outputs.
Check external state (database, API) to determine if the side effect succeeded.
Optionally consult a judge model only when heuristic checks are inconclusive.
Sample custom evaluator
public class KeywordOrJudgeEvaluator : ILoopEvaluator
{
private readonly string _stopKeyword;
private readonly AIAgent? _judgeAgent;
public KeywordOrJudgeEvaluator(string stopKeyword, AIAgent? judgeAgent = null)
{
_stopKeyword = stopKeyword;
_judgeAgent = judgeAgent;
}
public async Task<LoopDecision> EvaluateAsync(AgentResponse lastResponse, CancellationToken ct)
{
if (lastResponse.Text?.Contains(_stopKeyword, StringComparison.OrdinalIgnoreCase) == true)
return LoopDecision.Stop();
if (_judgeAgent != null)
{
var judgeRequest = new AgentRequest
{
Prompt = $"Is the following response acceptable? Answer YES or NO.\n\nResponse:\n{lastResponse.Text}\n\nCriteria: [list your criteria]"
};
var judgeResponse = await _judgeAgent.InvokeAsync(judgeRequest, ct);
if (judgeResponse.Text.Trim().StartsWith("YES", StringComparison.OrdinalIgnoreCase))
return LoopDecision.Stop();
}
return LoopDecision.Continue();
}
}
Practical safety and production best practices
Always set MaxIterations. Decide MaxIterations based on expected task complexity (2–5 for refinement tasks, more for longer repair flows).
Use a wall-clock timeout (CancellationTokenSource) to cap total runtime.
Log each iteration: input prompt, model output, evaluator decision, iteration duration, and any tool calls.
Be idempotent for side-effecting tool calls: include an idempotency key or track completed steps to avoid duplicate side effects.
Use backoff and limited retries for transient tool failures.
Consider human-in-the-loop for high-risk operations.
Capture metrics: iterations-per-task, average cost per task, success rate after N iterations.
Each loop iteration = another model call (cost + latency). If you expect 2–3 iterations on average, estimate cost = baseCost × expectedIterations + judgeCost (if used).
Use marker evaluators when iterations are cheap and you can control the prompt reliably; use judge evaluators when marker brittleness is likely.
For low-latency needs, constrain MaxIterations or move part of the evaluation logic to a fast heuristic.
Debugging, testing, and observability tips
Unit test evaluators with mocked AgentResponse objects. Keep evaluator logic pure when possible so it’s easy to test.
Mock provider clients for integration tests so you can simulate iteration sequences and failures deterministically.
Add correlation IDs across iterations for traceability.
Troubleshooting common failure modes
Marker never appended → verify prompt instruction clarity, or switch to judge evaluator.
Judge flip-flops (YES/NO across iterations) → add hysteresis: require two consecutive YES judgments, or use a heuristic plus judge.
Side effects duplicated → make tools idempotent or track operation IDs.
Loop stalls (no progress) → add progress metrics and a stuck detection rule (stop if no semantic change after N iterations).
Real-world scenario examples
Document extraction and QA: agent extracts structured data, runs validators, requests corrections until extraction passes validators or MaxIterations reached.
Customer support automation: agent composes reply drafts and refines them until a judge model or human reviewer marks them acceptable.
Conclusion and next steps Loop programming is a practical, powerful pattern for agentic AI: agent acts, inspects results, evaluates, and repeats — but only when bounded and observable. The Microsoft Agent Framework offers loop wrapper and evaluator patterns to implement this safely; adapt the example above to your MAF version and provider.
BlogWriter
My demo program, BlogWriter, uses a much more deterministic approach. While the author does pass its draft to the reviewer for evaluation, and they do create a small loop, a true looping approach would, perhaps, return better results. Something to try.
Jason and Myke have spent the last few days using the iPhone Duo, and they're ready to report back. And, oh yeah, there's also a new iPad mini and a draft of Apple's forthcoming announcements this week!
This episode of Upgrade is sponsored by:
Fitbod: Get stronger, faster with a fitness plan that fits you. Get 25% off your membership.
Sentry: Mobile crash reporting and app monitoring. New users get $100 in Sentry credits with code upgrade26.
Planet Money: Listen wherever you get your podcasts.
Decagon: The AI concierge for every customer. Get a personalized demo.
In 1984, Tomy sold a toy robot that promised a domestic future it couldn’t quite deliver. Forty years later, a maker in Miami gave it vision, a voice, and just enough intelligence to finally pull it off.
What’s the difference between a commercial and a literary plot? Discover how story goals, character journeys, and endings shape your novel.
There aren’t many authors (who actually sell books) who can get away without plotting. Bestselling authors, like Stephen King, who say they don’t plot, don’t realise they actually do it without realising it. They have an innate plotting talent. They just need less help than the rest of us.
I have noticed that even talented literary authors, who obviously don’t plot, have a shorter shelf life and fewer readers than writers who do plot.
My greatest wish as a reader is that more literary writers would learn to plot. Can you imagine the incredible books we would have to read? Some literary authors do this. Kate Atkinson writes beautifully, and her books have plots.
What’s The Difference Between A Commercial And A Literary Plot?
What is a literary plot?
Generally, literary plots are full of isolated characters without a clear goal.
They are slow and full of ambiguities.
The author concentrates on the inner journeys of the characters and their psychological setbacks. They are often passive.
The endings are often inconclusive and they mostly don’t end happily.
The bottom line?There are fewer book sales and the author makes less money, unless they win a literary prize.
What is a commercial plot?
Generally, commercial plots are driven by characters with a well-defined story goal.
There is lots of action and the author deals with the interior and exterior worlds of the characters.
There are physical and emotional setbacks.
The endings generally suit the genre of the story.
The bottom line?More of these books are sold and the author makes more money. There is next to no chance of winning a literary prize.
I think the way forward for modern writers is to marry the two.
If you write literary fiction, spend some time on plotting and add some pace.
If you write commercial fiction, make your characters more complex and add small twists to the endings of your stories.
Can You Combine Commercial And Literary Plotting?
You don’t have to choose between a commercial and a literary plot. You can use the strengths of both to tell a better story. Give your characters clear goals and obstacles to overcome, but also explore their inner lives, relationships, and motivations. Build a plot that keeps readers interested, and give them characters who stay with them long after they finish reading.
Examples Of Literary Authors Who Plot
In Life After Life by Kate Atkinson, Ursula tries to survive the dangers she encounters in her repeated lives. Eventually, she develops a more specific goal: to kill Hitler before he can unleash the Second World War.
In The Sense of an Ending by Julian Barnes, Tony wants to understand a mysterious bequest that forces him to revisit his past. As he investigates, he must confront memories and events that challenge his understanding of himself and his relationships.
In Never Let Me Go by Kazuo Ishiguro, Kathy and Tommy want to prove they are truly in love so they can obtain a deferral from the organ donation programme and have more time together.
The Last Word
The best of both worlds may be the way forward. Give your characters clear goals, build in setbacks, and create complex characters whose journeys keep readers turning the pages. Happy writing.
Log in
