Blurblogs
Global
Popular
Log in
alvinashcraft's blurblog
Sr. Content Developer at Microsoft, working remotely in PA, TechBash conference organizer, former Microsoft MVP, Husband, Dad and Geek.
164667 stories
·
33 followers

Long-horizon agents and durable execution by Martin Fowler (martin@martinfowler.com)
Tuesday November 10th, 2026 at 10:22 AM

Martin Fowler
1 Share

Long-horizon agents can automate complex tasks that take days or weeks to complete and involve multiple exchanges with people and external systems. Many things can go wrong during this time, including process crashes, network failures, and unanticipated real-world behavior. Kiran Prakash explains durable execution: an emerging pattern that helps agents recover from these failures and move towards their goal. In this programming model, the runtime keeps a record of the result of each step that interacts with the outside world so that, when a failure occurs, the program runs again, completed steps are skipped, and the program's state is rebuilt from the record.

more…

Read the whole story
alvinashcraft
just a second ago
reply
Pennsylvania, USA
Share this story
Delete

Fragments: November 10 by Martin Fowler (martin@martinfowler.com)
Tuesday November 10th, 2026 at 10:22 AM

Martin Fowler
1 Share

It’s that time of the year when the November Thoughtworks Technology Radar appears. Again my colleagues have taken a sweep of the technologies that our teams have used in the field, and pulled together a summary of what’s popped up in the last six months. It’s been a busy half-year, one that again has been dominated by AI, and that’s reflected in the many AI blips on the radar.

The mass of blips makes it hard to develop a coherent story around trends, but Chris Ford has done a fine job of the writing the accompanying article on macro trends.

One team of passionate people experimenting with agents can make almost anything work. They share a context, a credit card and a mental model. They review each other’s work because they sit together, and they know what to check and how much latitude to give the coding agent.

Two hundred people, though, cannot work that way. The moment agentic AI becomes an operating model rather than a pilot, a second layer of problems appears: How is context shared between teams? What’s the token bill? How do we deal with increased review demands? What’s the agent allowed to reach, read and change? And who’s now responsible for the harness everyone depends on?

Experience so far suggests that the questions that enterprises face are about: shared context, token cost, cognitive load, security, and platform building. While these issues are distinct, they are not separate, so strategies to answer them must take all of them into account. I also echo Chris in reminding us that however impressive these tools are, we must remember that while we have to rethink systems and workflows to utilize these technologies, they should be built to serve humans - and we should all benefit from its bounty.

 ❄                ❄                ❄                ❄                ❄

Ben Thompson’s Stratechery newsletter reflects on the recent election’s impact on data centers.

I watch a lot of sports, and particularly this fall, watched a lot of baseball. My Milwaukee Brewers — excuse me, my World Champion Milwaukee Brewers — demanded my full attention, which also meant I saw a lot of campaign ads, and by far the most common topic, at least here in Wisconsin, was data centers. Moreover, it was never entirely clear which way it cut politically: both parties were anti-data center, accusing their opponent of being the opposite.

He points out that this opposition is recent, and tied to a shift in focus from the internet and streaming movies to AI:

you know, the technology that in the best case scenario is going to take everyone’s jobs, and in the worst case scenario end the human race.

There’s going to be a lot said about how Democratic officials will be at the heart of blocking data center building, but as his point above about campaign ads illustrates, the opposition to data centers is (at least at the moment) a bi-partisan opposition.

He follows up with “ten takes” from these consequences. The big beneficiaries of this opposition to building data centers will be those who own the data centers that have already been built. The opposition will reduce supply of tokens, meaning that token costs will go up, thus increasing the profits for those who already have their token generators plugged in and humming.

There’s also benefits to those who aren’t constrained by the need to build more big data centers. Other countries aren’t affected by a change in US policy, China is the most talked-about case here, but it’s not the only country that may benefit from a drop in US data center building. Apple may also benefit, if it means more AI work is pushed to edge devices that are their territory, in addition to having not sunk so many dollars into AI-specific investments.

To his analysis, I’ll add that this may also have a somewhat counter-intuitive effect on regulation, and corporate attitudes to it. The common expectation is that businesses are opposed to government regulation, since that interferes with their free pursuit of profit. That’s true to an extent, but it’s quite common to see established businesses lobby (quietly) in favor of regulation that acts as a barrier to new competitors. So I wouldn’t be at all surprised to see companies that have already built data centers to be pushing for regulations that impede or prevent new ones from being built.

 ❄                ❄                ❄                ❄                ❄

I’ve been a long-time fan of Hadley Wickham, for how he has tamed the rather messy world of R programming into a well-designed tidyverse. He is another thoughtful voice to listen to as we move into the world of agentic engineering, in this case by trying to figure out what the real cost of tokens is.

I hear a lot of suspicion that AI tokens are unsustainably priced and that frontier labs are heavily subsidising costs with their VC money. They want to get us hooked now, so they can pull the rug out from under us whenever they want to make a bunch of money. Is that true? My sense is that it’s mostly not. But can I back that up with some data?

His conclusion is that it doesn’t seem that token costs are significantly subsidized on per-token API basis, but they are with subscription plans. (He caveats his analysis, saying it doesn’t take into account recouping the costs of training the models.)

 ❄                ❄                ❄                ❄                ❄

There’s a creepy excitement at the thought that models are clever enough to edit the weights of models, an activity that Takehi et al refer to as Weight-Space Programming.

Just by looking at the raw weights, GPT-6 Astra can tell what a network was trained to do. And without any environment feedback, it has no trouble writing, from scratch, small game-playing networks that outplay networks trained with PPO for 100 million steps.

While we don’t know if we’ll ever be able to download skills directly into our brains (“I know kung fu” - or for those in my particular demographic: Joe 90), this does raise the possibility of downloading capabilities into models by manipulating their weights.

Simon Wilison riffed on this paper by examining whether an LLM can understand model behaviors from the weights alone, by examining four simple neural networks, each trained to play a specific game, and figuring out which game it was by just looking at the weights.

 ❄                ❄                ❄                ❄                ❄

Jessica Kerr (jessitron):

My days of teaching AI to do stuff are numbered.

Getting it to teach me stuff is more fun.

 ❄                ❄                ❄                ❄                ❄

Houck et al (that “al” includes some folks who I’ve learned to pay a lot of attention to), look at how we can assess the quality of context supplied to agents.

An agent may draw from source code, specifications, issue trackers, chat histories, architectural documents, tool output, and user instructions, then act on that combination with limited human supervision. Information that was written for different audiences, at different times, and for different purposes is brought together into a temporary operating environment for a particular task. Whether each source is individually useful does not reveal whether the resulting context is fit for the work the agent is being asked to do.

They’ve come up with a framework for describing the quality of this assembled context, which they refer to as CAFE(S). The acronym refers to five dimensions they look at:

  • Clarity: Can the agent interpret the request as intended?
  • Actionability: Can the agent proceed, and know when it is done?
  • Fidelity: Can the agent trust that this context is true?
  • Efficiency: Can the agent focus on what matters?
  • Security: Should the agent have this context at all?

They also list a set of half-a-dozen “context smells”: common cases where these break down (and a nod to Kent’s original metaphor)

I’m not sure about the name, (how exactly should I pronounce “CAFE(S)”?) and the article exhibits rather more LLM Voice that I prefer. But on the whole I think this is a good framing for thinking about an important part of feeding our silicon servants/overlords.

 ❄                ❄                ❄                ❄                ❄

Cindy Sridharan noted that the problems she ran into six months ago are no longer around:

What was missing in my setup 6 months ago was the verification process I now use.

70% of my daily token spend goes into verifying patches, and only ~20-30% into creating them.

That resonated with conversations I’ve had with folks using agents in enterprise settings. They are using far more tokens to check the working of models (through guardrails and evals) than they are to generate a response to a request. A single request to a model from a customer may trigger a dozen turns to other models to probe its input and verify its output.

Read the whole story
alvinashcraft
13 seconds ago
reply
Pennsylvania, USA
Share this story
Delete

Debug Unity Builds From Rider With the TeamCity CLI and MCP by Olga Bedrina
Tuesday November 10th, 2026 at 10:22 AM

TeamCity : Powerful CI/CD for DevOps-centric teams | The JetBrains Blog
1 Share

At GameDev Day 2026, TeamCity solutions engineer Artem Rokhin demonstrated how to investigate a CI issue from Rider using the TeamCity CLI and MCP integration.

For the demo, he used a Unity game that worked locally but was missing its background on staging. Though the build had passed, Artem used the CLI to inspect the logs and found an incorrect build parameter.

After committing his fix, he asked an AI agent connected through MCP to explain the issue and follow the new build.

Let’s go through that example, with the commands and setup you’ll need to try it with your own project.

The problem on staging

In Artem’s demo, a Unity game works locally but is missing its moon landscape on staging, despite a successful TeamCity build. He traces the deployment back to its build and checks the logs and configuration to see where they differ from those in his local setup.

What you need

To apply this workflow to your own project, you need:

  • A Unity project open in Rider.
  • A TeamCity pipeline that builds and deploys the project, with its YAML settings stored in the repository.
  • Access to that pipeline’s build details and logs.
  • The TeamCity CLI installed on your machine.
  • For the AI portion, an activated MCP-capable agent in Rider and access to TeamCity’s MCP endpoint. Artem uses Claude Agent.

The recording uses TeamCity Cloud. TeamCity On-Premises users can use 2026.2 as a baseline with generally available Pipelines. See the Pipelines documentation for setup and version details.

The commands below use example IDs. Replace MyGame_BuildWebGL with your WebGL job’s ID, 12345 with the build you are investigating, and 12346 with the new deployment run when you reach that step.

Connect the TeamCity CLI from Rider

Open Rider’s terminal and install the TeamCity CLI if you have not already done so. On macOS with Homebrew:

brew install jetbrains/utils/teamcity

On Windows with WinGet:

winget install JetBrains.TeamCityCLI

Other installation methods are available in the Getting Started guide.

Connect to your server and check the connection:

teamcity auth login --server https://example.teamcity.com
teamcity auth status

Replace the URL with your own TeamCity address. Follow the login flow, which may open your browser for authorization. The authentication documentation covers the available methods.

Once connected, return to Rider’s terminal to inspect the build.

Find the build that produced the game

List recent runs of the WebGL job:

teamcity run list --job MyGame_BuildWebGL --limit 5

Choose the run that produced the version on staging. Include successful runs in your search: In this example, the relevant build is green.

If several runs have completed recently, check the deployment’s source build and revision before investigating. The most recent WebGL build may be newer than the version you have open in the browser.

Inspect the selected run:

teamcity run view 12345

In the recording, Artem starts with the pipeline run, identifies its WebGL job, and then opens that job’s build details. Going directly to the WebGL job’s recent runs is a shorter route when you already know which job produced the game.

Inspect the logs and the configuration

Read the WebGL build log:

teamcity run log 12345

The default log viewer supports search. Press /, enter a term, and use n or N to move between matches. Full command details can be found in the guide to managing runs.

Artem has recently been working with build flavors, so he searches for flavor. The log shows that the build used gameBuildFlavor with the value diagnostic.

That provides a specific setting to investigate. Search the checked-out project for gameBuildFlavor, including both the pipeline configuration and the Unity build code that uses it. 

The value printed in the log tells you what the CI build actually received; the repository helps you trace where it came from and what it controls.

In this demo project, the expected flavor is game. The diagnostic value is associated with the version missing its background. These names are part of the example project’s build setup; your project may use a different parameter or no flavor switch at all.

When you follow a similar lead, check both the definition and the code that consumes the value. Finding an unexpected parameter is useful evidence. Understanding its effect on the build lets you judge whether changing it addresses the symptom.

Commit the configuration change

Artem finds the diagnostic value in the pipeline’s YAML configuration and changes it to game. Because the configuration lives alongside the game code, he can review and commit the change from Rider.

TeamCity is configured to watch the repository, so pushing that commit starts a new pipeline run. You do not need to restart the old build manually in this setup.

For your own pipeline, check its trigger settings and confirm that the new run includes the configuration commit. Keep the original run ID so you can compare the settings and results afterward.

While the new pipeline runs, Artem shows another way to investigate CI results from Rider: an AI agent connected to TeamCity through MCP.

Connect Rider to TeamCity MCP

MCP gives the agent access to TeamCity tools, including tools for retrieving build logs and querying build information. The server endpoint is:

https://example.teamcity.com/app/mcp

For this investigation, create a TeamCity access token limited to the relevant project with the Read-only preset. That is enough to inspect the CI results while you make the configuration change through Git. See TeamCity’s AI integration documentation for token permissions and connection options.

In Rider, go to Settings | Tools | AI Assistant | Model Context Protocol (MCP) and add an HTTP connection. A valid connection configuration should look like this:

{
  "mcpServers": {
    "TeamCity": {
      "url": "https://example.teamcity.com/app/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TEAMCITY_TOKEN"
      }
    }
  }
}

Replace the server address and token placeholder in your private local configuration. Keep the token out of files you commit or share. Apply the configuration and check that Rider can connect. Our AI Assistant MCP guide describes the connection settings.

If you use Claude Agent, also open Settings | Tools | AI Assistant | Agents and enable Pass custom MCP servers. Then select Claude Agent in the AI chat. This exposes the configured connection to the agent, as described in the Claude Agent documentation.

Ask the agent to investigate the symptom

Give the agent the symptom and identify the build. For a similar investigation, you could adapt this prompt:

Use TeamCity MCP to inspect WebGL build BUILD_ID in PROJECT_ID. The game has its background when built locally, but the version deployed to staging is missing it. The CI build is successful. Inspect the build details and logs, then compare relevant settings with the project files. Explain the likely cause and cite the build ID, log entries, and files that support your explanation. Do not change anything yet.

Replace the IDs before sending the prompt. Attach the relevant pipeline YAML and Unity build script if the agent needs help locating them. TeamCity MCP supplies the CI information; the coding agent uses its IDE and file tools to inspect the local project.

In the recording, Claude identifies the diagnostic build flavor and connects it to the unexpected game behavior. It also notices that the correction has already been committed. This part of the demo happens after Artem’s CLI investigation and manual fix, so the response can refer to that existing change.

Check the evidence in the answer. Does it refer to the run you asked about? Does the parameter value match the log? Does the explanation fit the code that consumes it? Those checks make it easier to assess the proposed cause before changing a different pipeline or project.

Follow the new deployment

Artem then asks the agent to check the new pipeline and return the staging URL when deployment finishes. The agent queries TeamCity during the active session and reports progress.

A follow-up prompt could be:

Check the pipeline run for the configuration fix in PROJECT_ID. Report whether the WebGL build and deployment have completed. When deployment succeeds, retrieve the staging URL from the deployment log. If the run is still in progress, report its current state.

For continuous terminal updates, you can also watch the deployment run with the CLI:

teamcity run watch 12346 --logs

Here, 12346 represents the new deployment build’s ID. Use the ID from your own pipeline. Once it completes, inspect its log if you need to retrieve the URL:

teamcity run log 12346

Verify the deployed game

Open the staging URL and repeat the interaction that exposed the problem. In Artem’s demo, the moon landscape is visible again, and the player can drive around it as expected.

This check confirms that the configuration change resolved the observed problem. The earlier run was already green, so another successful pipeline result alone would not establish that the background had returned.

For your own project, also confirm that you are viewing the deployment from the new run. Then compare the result with the local version that behaved correctly.

Make the next investigation easier

The example suggests a few useful improvements to the pipeline:

  • Log the selected build flavor and other settings that change the game output, so they are easy to find.
  • Validate the allowed flavors for staging builds in your project’s build script, and ensure that the run fails early when the value is unexpected.
  • Add a check for the behavior that went wrong here. Depending on the project, that might be a WebGL smoke test that loads the relevant scene or a visual check of the deployed game.

Start with the missing check that would have caught this particular problem. That gives you a concrete way to improve the pipeline, along with a repeatable method for investigating its results in Rider.

Read the whole story
alvinashcraft
19 seconds ago
reply
Pennsylvania, USA
Share this story
Delete

AI Slowdown Listicle
Tuesday November 10th, 2026 at 10:22 AM

Agile Otter Blog
1 Share

Please read this as a research list. Do not read it as “the ammunition list for naysayers,” because that’s not how it is intended.

It seems that agentic code is not perfect or “at the highest level of skill” by default, and so a prudent developer will spend significant time adopting or building harnesses, skills, and tools to improve ai.

It behoves us as practitioners to know what AI does well, what it does poorly, and how/when to use it to best effect. That means we study not only strengths, but also failure modes.

Enjoy.

It is my understanding (2026-11-10) that AI doesn’t mean all the work will be done perfectly and immediately, nor does it mean that programming skill does not matter.

On the contrary, it seems that for AI to work, we have to double down on the skills that XP, code craft, clean coding, automated testing, and such communities have recommended for decades.

I have found that there are advantages to basing your work on The 8 Code Virtues and a solid base of unit tests and mutation testing.

Engage with me via social to add your favourite articles to this list.

Read the whole story
alvinashcraft
25 seconds ago
reply
Pennsylvania, USA
Share this story
Delete

IBM and Microsoft announce a new AI agent platform designed for industry-specific transformation by sbaynes
Tuesday November 10th, 2026 at 10:22 AM

Source
1 Share

IBM and Microsoft announce a new AI agent platform designed for industry-specific transformation

|
6 minutes
  • IBM Consulting brings more than 6,000 prebuilt AI agents and assets, the experience building them, and industry expertise, while Microsoft provides AI, cloud and security technologies
  • New platform comes with forward-deployed engineering support from both Microsoft and IBM Consulting
  • SEI’s enterprise transformation illustrates the types of business needs that agentic AI can help address

ARMONK, N.Y., and REDMOND, Wash. — Nov. 10, 2026 — IBM (NYSE: IBM) and Microsoft Corp. (NASDAQ: MSFT) on Tuesday announced Imagine Enterprise, an agentic platform designed to help enterprises build, run and govern AI agents that help people execute industry-specific processes. IBM Consulting and Microsoft have combined their technologies and expertise with forward-deployed engineering support for qualifying engagements to help enterprises accelerate AI deployment.

Imagine Enterprise aims to help organizations — especially in regulated industries — combine human expertise with AI agents and connect them to unique enterprise data and daily processes. The platform is built to help organizations govern AI agents with controls, observability and AI-compliance support designed to help secure industry-specific AI deployment.

IBM Consulting brings AI transformation expertise and a fleet of more than 6,000 AI agents and assets built on IBM Consulting Advantage, the company’s AI-powered delivery platform. Microsoft provides the AI, data, cloud and security foundation that is designed to help the agents operate securely across the enterprise, while forward-deployed engineers from both companies can help customers move from experimentation to production.

Instead of organizations having to start from scratch, replace the platforms they already depend on or spend months assembling the pieces themselves, Imagine Enterprise is designed around three simple advantages:

  • Instant Expertise: More than 6,000 AI agents and assets
    Starting with prebuilt AI agents, assets and deep industry expertise from day one. Organizations can begin with capabilities already built for their industry and adapt them to the unique needs of their business.
  • Fits the Enterprise: A platform that adapts
    Imagine Enterprise is designed to connect with an organization’s existing data, processes, systems and technology environment. Enterprises can start where they are, build on what they already have and scale over time, with enterprise security, identity, AI governance and control built into the experience.
  • Intelligence Multiplier: Puts the knowledge of the business behind people and AI agents
    Imagine Enterprise brings IBM Consulting’s hands-on experience building and running more than 6,000 AI agents together with the Microsoft technologies to connect enterprise knowledge, data and context across the organization. People and agents can operate with richer enterprise knowledge and context behind them, helping put what they already know to work across people and agents.

Supporting these advantages, Microsoft and GitHub technologies including Microsoft Foundry, GitHub Copilot and Microsoft Copilot Studio help coordinate AI agents and enterprise systems, while Microsoft Fabric, Microsoft 365 infrastructure, Microsoft Entra, Microsoft Purview, Microsoft Defender and Microsoft Azure provide the data, identity, AI governance, security and cloud foundation. IBM Consulting Advantage’s Process Studio, Context Studio and Advantage Core help organizations design, deploy and govern tailored AI agents.

From emerging technology to business opportunity

SEI, a leading global provider of financial technology, operations and asset management services within the financial services industry, is looking to advance its enterprise transformation by exploring emerging and agentic technologies with IBM Consulting, with the support of Microsoft, and is leveraging other technologies from IBM Consulting and Microsoft to help improve how the company operates and enhance the experiences it delivers to employees and clients, supporting SEI’s commitment to sustainable growth.

“AI is evolving quickly, and our focus at SEI is on applying it with intention in ways that strengthen trust, enhance human decision-making and create meaningful value across our business,” said Sneha Shah, Chief AI Strategist and Head of SEI Next. “Our work with IBM Consulting and Microsoft is helping us explore how emerging capabilities could support critical processes and be responsibly scaled over time, with the goal to create better experiences for our employees and clients. By combining emerging technology with SEI’s deep industry and operational expertise, we can move faster, scale more effectively and create new opportunities for growth.”

“Clients are no longer looking for AI that simply helps people do the same work a little faster. They want to build an agentic business,” said Mohamad Ali, Senior Vice President and Head of IBM Consulting. “Microsoft and IBM are helping organizations bring human expertise and digital labor together in a way that can scale across the enterprise, to turn AI from a productivity aid into a real competitive advantage.”

“Enterprise AI will be won by organizations that can turn their data, industry expertise and AI into action,” said Rodrigo Kede Lima, President, Microsoft Frontier Company. “Together, Microsoft and IBM Consulting are bringing the best of our technology and expertise, to help customers move from AI ambition to business impact faster and with confidence.”

Imagine Enterprise’s industry-specific frameworks initially include financial services, insurance and retail, and are designed to establish a repeatable blueprint for potential additional industries such as manufacturing, energy, public sector, mobility and more in the future.

Imagine Enterprise is available to clients by engaging IBM Consulting to deploy and operate the platform on the client’s behalf or co-innovating with IBM Consulting on enterprise-scale implementations. Self-management options are also available.

To learn more, visit https://www.ibm.com/consulting/imagine-enterprise.

About IBM

IBM is a leading provider of global hybrid cloud, AI, and consulting expertise. We help clients in more than 175 countries capitalize on insights from their data, streamline business processes, reduce costs, and gain the competitive edge in their industries. Thousands of governments and corporate entities in critical infrastructure areas such as financial services, telecommunications, and healthcare rely on IBM’s hybrid cloud platform and Red Hat OpenShift to affect their digital transformations quickly, efficiently, and securely. IBM’s breakthrough innovations in AI, quantum computing, industry-specific cloud solutions, and consulting deliver open and flexible options to clients. All of this is backed by IBM’s long-standing commitment to trust, transparency, responsibility, inclusivity, and service.

About SEI

SEI (NASDAQ:SEIC) is a leading global provider of financial technology, operations, and asset management services within the financial services industry. SEI tailors its solutions and services to help clients more effectively deploy their capital—whether that’s money, time, or talent—so they can better serve their clients and achieve their growth objectives. As of Sept. 30, 2026, SEI manages, advises, or administers approximately $2 trillion in assets. For more information, visit seic.com.

About Microsoft

Microsoft enables digital transformation for the era of an intelligent cloud and an intelligent edge. Its mission is to empower every person and every organization on the planet to achieve more. Microsoft delivers cloud platforms, AI capabilities, productivity tools, and business applications that help organizations innovate, improve agility, and drive meaningful outcomes at scale.

Statements regarding IBM and Microsoft’s future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only.

For more information, press only:

Microsoft Media Relations, We. Communications for Microsoft, (425) 638-7777, rapidresponse@wecommunications.com

Joel Rushing, IBM, Joel.Rushing@ibm.com

Alicia Rudd, SEI, Arudd@seic.com

Note to editors:  For more information, news and perspectives from Microsoft, please visit Microsoft Source at https://news.microsoft.com/source. Web links, telephone numbers and titles were correct at time of publication but may have changed. For additional assistance, journalists and analysts may contact Microsoft’s Rapid Response Team or other appropriate contacts listed at https://news.microsoft.com/microsoft-public-relations-contacts.

Tags:

The post IBM and Microsoft announce a new AI agent platform designed for industry-specific transformation appeared first on Source.

Read the whole story
alvinashcraft
37 seconds ago
reply
Pennsylvania, USA
Share this story
Delete

Decision Models in Hugging Face
Tuesday November 10th, 2026 at 10:22 AM

Hugging Face - Blog
1 Share
Read the whole story
alvinashcraft
48 seconds ago
reply
Pennsylvania, USA
Share this story
Delete
Next Page of Stories