Reading

Coding agents can already produce anything from cat memes to B2B SaaS applications. At the 2026 AI Engineer conference, Ido Salomon, founder of AgentCraft, creator of MCP-UI, and co-maintainer of MCP Apps, posed a counterintuitive question: If agents are so capable, why doesn't everyone have an army of them? His diagnosis is that the bottleneck lies not in model capabilities, but in humans. Each additional agent launched adds demands for guidance, direction, and review, all falling on the same person's attention until the act of supervision itself becomes exhausting.

Salomon argues that the skills needed to solve this bottleneck already exist. He draws an analogy to real-time strategy games: in titles like Warcraft or The Sims, players control multiple units simultaneously not by checking every frame of action, but through spatialized state awareness, rapid jumping to items requiring attention, and the ability to delegate tasks to sub-units. AgentCraft brings this control language into coding agent workflows—agents appear as units on a map, the file system is projected as terrain, heat maps show which directories are active, and pressing the spacebar jumps to the most recent item needing human attention.

Building on this foundation, Salomon described three progressive layers: visibility lets users know who is doing what; autonomy reduces continuous supervision through task delegation (agents scan the codebase and suggest next steps, which users accept with a click) and orchestrators (users provide broad goals, which the orchestrator breaks down and executes in local containers); and review toolkits package completed changes into file lists plus visual evidence (screenshots or screen recordings), allowing users to quickly judge whether to accept them, or even run multiple instances in parallel to select the best implementation. Finally, shared rooms allow designers and engineers to advance different parts of the same project simultaneously without waiting for the other to stop.

However, Salomon himself placed limits on the applicability of these mechanisms. He estimated that roughly 90% of people need a lower entry point, explicitly stating that this figure is an approximation rather than a measured result; he showed that the experimental simplified mode, Loopers, is still in a very early stage, with its name not yet finalized; and he acknowledged that while the interface may be stunning to some, it could be difficult for others, thus treating customization as an unfinished exploration. Nevertheless, the entire argument rests on a product demonstration and personal usage experience, with no controlled experiments proving that the game metaphor actually reduces cognitive load rather than simply swapping one complexity for another. For developers already accustomed to terminals and IDEs, whether this spatialized interface is truly more efficient than lists and logs remains an open question.

The software factory WorkOS initially built placed code in a Cloudflare sandbox, connected it to model routing, fed in prompts, and generated pull requests—the results were hard to distinguish from engineers running Claude Code on their laptops. Product engineer Ryan Cooke described this disappointing finding at the 2026 AI Engineer World's Fair and proposed an argument: sandboxes, models, and repositories are necessary conditions, but not sufficient ones; a true software factory must encode the organization’s own engineering workflows into automation.

Cooke argues the industry is optimizing for the wrong variables. Ramp reports over 50% of merged PRs come from AI, Spotify reports more than 1,500 AI-generated PRs merged, and Stripe merges over 1,000 PRs per week—these numbers describe production rates, not value-delivery rates. When building the factory itself consumes engineering time or requires dedicated teams, judging success by PR count alone obscures a harder question: did features reach customers faster? His alternative measurement directions are feature delivery time, defect rate at ship, and issue resolution time.

WorkOS’s response is to break engineering workflows into executable automation units. They built two system layers: TARS is the user-facing interface, embedded in Slack, Linear, and GitHub, subscribing to project events via webhooks; Horizon is the infrastructure orchestration layer, managing sandbox pools and event loops. The key design lies in TARS not just generating code, but participating in the engineering workflow itself.

The specific mechanism works as follows: Linear tickets have dependencies, where one ticket blocks another. When TARS receives a ticket-completion webhook, it automatically picks up the next unlocked ticket to continue execution. Engineers can ask TARS to reevaluate the project plan between steps, uncovering omissions exposed during implementation. This autonomy comes from responding to project state, rather than treating each coding session as an isolated task.

At the planning level, WorkOS uses Hilltop documents—a product requirements document consolidating project purpose, customer needs, competitive analysis, early designs, and milestones. A PM agent reads the Hilltop document, breaking it down into tickets with dependency chains. Humans review and add context; after merging, Horizon automatically unlocks the next ticket. Cooke used his Vaults API project as an example: describing the goal in a few sentences in Slack, TARS created a Linear project, a Notion draft, and decision logs; he then added specifications, filled in ambiguities, and handed it back for execution. Cooke reported the team now does more project briefs and Hilltop work because the agent solves the blank-page problem; however, drafts often grossly overestimate project scope, requiring engineers to cut substantial portions.

Cooke calls the internal MCP gateway the "context engine," arguing it is the most worthwhile investment for any team starting a factory. It does more than connect tools; it encodes organizational knowledge into tool descriptions: telling agents which semantic tables exist in Snowflake, what they contain, and which tables to query for specific questions. WorkOS later opened this gateway to Slack queries for data and customer analysis, finding it equally applicable to non-coding scenarios. Cooke’s quote: "Agents can figure it out with minimal context, but it's slow; they spend tokens rediscovering conventions. We keep tightening that loop."

At the AI Engineer World's Fair in September 2026, Zixuan Li, head of Zhipu (Z.ai), faced a question he was repeatedly asked: Why release a model with near-frontier capabilities as open weights? He directly acknowledged the subtext of the question: Does releasing near-frontier models as open weights mean handing the market over to inference service providers?

Li broke down the open-weight decision into three aligned sets of needs. The first set is security and control. He stated that if enterprises or governments—particularly institutions in the Western world—want to use the model, Zhipu uploads the weights to Hugging Face so they can run it on their own servers. This changes who operates the model and gives customers more control over deployment. He explicitly described this as a deployment option, not a completed installation or a verified security outcome.

The second set is diversity, or domain-specific customization. Customers in the legal, financial, and security fields need different model behaviors, which requires access to weights for fine-tuning. He cited Harvey, a legal AI company, as an example; Harvey is already fine-tuning GLM-4.1 and may next consider fine-tuning GLM-5.2. However, Li framed Harvey’s shift to 5.2 as a 'possibility' rather than a completed adoption. He also mentioned speaking with many other companies that are considering fine-tuning GLM as a differentiation strategy.

The third set is co-designing and predicting the future. If customers or developers want to participate in shaping the next steps of the model, they need to understand its architecture and training recipe. Weight access supports the first two paths (local deployment and fine-tuning), while publicizing architecture and training information supports the third—a deeper form of collaboration. Li summarized this as 'making the right bets': by letting customers see technical details, Zhipu can better judge market demand and reduce directional errors.

In Li’s framework, these three sets of needs are not a one-way concession but a two-way match: customers want security, customization, and a sense of participation; Zhipu wants trust, ecosystem expansion, and more accurate directional judgment. He said, 'Our needs and our ecosystem fit together.'

Boundaries to preserve: This is the company head’s explanation of his own strategy, not an independently verified market result. There is no public financial data to determine whether open weights truly generate more revenue than closed-source alternatives. Harvey’s fine-tuning of GLM-4.1 is a case mentioned verbally by Li, with no contract amounts or adoption scale provided. On the Artificial Analysis Intelligence Index, GLM-5.2 scores 51, ranking behind Claude Opus 4.8 (56) and GPT-5.5 (55) but ahead of other open-weight models shown—these are comparisons presented by Zhipu itself, and the evaluation details are insufficient to infer performance in specific workflows.

For readers interested in AI industry investment opportunities, the practical value of this speech lies in how it shows a frontier lab redefining open weights from a technical choice to a business strategy. If this logic holds, the adoption rate of open models, the number of fine-tuning clients, and the scale of ecosystem contributors may reflect an AI company’s commercial health more accurately than single benchmark rankings. However, Li’s framework has not yet been independently tested.

MEMS technology company xMEMS Labs has announced that two models in its "μCooling" semiconductor cooling chip series (XMC-2400 and XMC-4800) entered mass production on October 20, 2026. The chip, approximately 1.1mm thick, uses piezoelectric MEMS technology to generate airflow through ultrasonic vibration of a silicon diaphragm, enabling active cooling without traditional rotating fan blades.

The technology primarily targets devices with extremely limited internal space, such as AI smart glasses and smartphones. At the CEATEC 2026 exhibition, Mark Wood, Vice President of xMEMS Japan, noted that surface temperatures exceeding 45°C on smart glass frames can compromise safety and comfort, while passive cooling often struggles to handle heat from high-performance processors. A live demonstration showed the frame side equipped with μCooling at 38°C, compared to 61°C on the uncooled side.

xMEMS currently offers three specifications: the smallest model, XMC-1200 (standard power consumption ~70mW), remains in the sampling stage, targeting optical systems and processors; the medium-sized XMC-2400 (max airflow 27cm³/s, power ~150mW) and large XMC-4800 (max airflow 47cm³/s, power ~240mW) have entered mass production. Wood stated that multiple companies are evaluating adoption, with end products featuring μCooling expected to launch in early 2027.

US chipmaker NVIDIA and social media giant Meta announced on October 27, 2026, that they will expand their partnership on Meta's Muse model family and AI agents. As a core component of this deepened strategic relationship, NVIDIA will invest $3 billion in an equity offering by Meta.

The collaboration aims to optimize computing infrastructure for AI agents amid growing consumer demand. Meta plans to expand its deployment of NVIDIA's next-generation hardware, including Vera Rubin GPUs and Vera CPUs for improved performance per watt, as well as Spectrum-X Ethernet networking and BlueField-4 units to support network efficiency. Additionally, the companies will explore further cooperation on security architectures, including NVIDIA's OpenShell reference architecture and Sentry systems for monitoring agent behavior.

Jensen Huang, founder and CEO of NVIDIA, stated that AI agents represent a new computing paradigm, shifting from operating applications to working with intelligent agents. Mark Zuckerberg, founder and CEO of Meta, said Muse is designed to bring personal superintelligence to billions of people, noting that this partnership will help deliver the best personal agents and frontier models. This move signals tighter infrastructure binding between the two tech giants in the era of AI agents, with NVIDIA directly supporting the compute expansion of one of its largest customers through capital injection.

Laurie Voss from Arize AI explores the transition from 'vibes' to production-grade AI agent evaluation in a workshop video published on the AI Engineer channel. The episode describes a specific case: a financial analysis agent built with the Claude Agent SDK generated 13 reports, but an LLM correctness judge rejected all of them. Voss analyzes that the issue stemmed from the agent using live web research while the judge graded based solely on its own knowledge, lacking the necessary research context.

To address this bias, Voss demonstrates how to supply the agent's collected source data to a faithfulness evaluator. This adjustment produced a more useful result split: six reports were judged faithful and seven unfaithful. Beyond correcting the judge, the workshop introduces instrumenting the agent using OpenTelemetry and OpenInference to send traces to Arize AX. By reading these actual traces, developers can identify specific issues such as failed file writes, missing report content, and excessive web searches.

Additionally, Voss shows how to build a custom actionability rubric with explicit criteria, tagged inputs, examples, and binary labels. Each evaluator checks only one dimension, allowing failures to precisely identify what needs attention. Human annotations serve as a comparison set for evaluating the judge, with deliberately random demo labels illustrating how to investigate disagreement. The entire process emphasizes the core practice of reading actual traces and defining success before automating grading.

In a talk published on October 6, 2026, at the AI Engineer conference, OpenAI engineers Charlie Guo and Gabriel Chua detailed the application of Codex in real-world software development. They showcased a core case: four minutes after receiving a capture of a Slack conversation, Codex produced a macOS live translation app. This process highlighted how context, plugins, and supervision work together, with the conversation supplying the idea, plugins providing development practices, and a live German translation testing the result.

Beyond rapid prototyping, the two engineers demonstrated more complex projects, including turning repository history into a shareable site about the Agents SDK, requesting an IKEA-style repository visualization from a phone, downloading dashboard data through computer use, and building a Google feedback form via a Chrome extension. They specifically distinguished between browser control for local testing and access to applications where a user is already signed in.

For longer-term projects, Guo and Chua introduced how to set verifiable completion criteria using Goals, while dividing work and exposing blockers through subagents, visible threads, audits, and progress dashboards. Remote threads handed browser testing to a local machine, and hooks added deterministic checks at specific workflow points. Automations connected these pieces, such as waiting for server provisioning, reviewing Slack feedback, or starting a separate worktree to address a bug.

The recurring question in the talk is how to turn a successful individual request into a repeatable process with useful checks and human input where needed. For engineers looking to optimize their AI-assisted programming workflows, this resource provides specific practical references on Codex integration, plugin usage, and automated supervision.

South Korean e-commerce solution provider Indent Corporation, which operates the review platform VReview and influencer marketing platform Spray IO, unveiled its new commerce AI agent 'Indent M' at a press conference on October 27. The tool is designed to automate overseas influencer marketing workflows from planning and creator selection to campaign execution, currently supporting localized campaigns in 72 countries.

Users can interact with the AI via natural language to provide product details, marketing goals, and desired creator types. Indent M then evaluates expected performance and costs based on past data to automatically select creators and run campaigns. This capability stems from the company's experience managing over 930 campaigns for more than 100 brands and facilitating over 80,000 pieces of collaborative content through Spray IO over the last three years.

The company plans to integrate Indent M, which handles marketing, with Indent O, a cross-border sales platform upgraded from its Booboo Shop service launched last November. Together, they aim to offer a full Direct-to-Consumer (D2C) pipeline for small and medium-sized Korean brands expanding globally. CEO Yoon Tae-seok stated that the goal is to lower barriers to global marketing by reducing the need for local manpower and infrastructure.

South Korean generative artificial intelligence company DeepBrain AI announced on October 27 that it has completed the development of 'Face ID Discrimination' technology, designed to prevent the unauthorized use of personal likenesses via deepfakes. This technology serves as a core component of the company's generative AI safety engine, 'AI Guard,' and automatically detects whether image or video generation requests contain facial features of pre-registered 'protected individuals.'

The system compares faces in the request against a database of protected persons, categorizing judgments into three levels: 'Safe,' 'Caution,' and 'Match.' Platforms adopting this technology can automatically block generation requests judged as a 'Match' and route 'Caution' level requests for manual review. DeepBrain AI stated that the design accounts for variables such as glasses, hats, hairstyles, and shooting angles, aiming to accurately identify protected individuals while minimizing false positives where similar-looking people are mistaken for the same person. The company claimed that the technology met task criteria in evaluations by an internationally accredited testing institution, though specific performance metrics were not disclosed.

For video content, the system extracts frames at set intervals for facial analysis, aggregating segments featuring the same person along with maximum similarity data for administrator review. Additionally, DeepBrain AI provides tools for building and managing the database, which automatically collect photos from multiple sources and filter out blurry images, namesakes, or duplicates. The company offers six API interfaces, enabling content platforms and public institutions to integrate this functionality into their existing services. This R&D project was funded by the South Korean Ministry of Culture, Sports and Tourism and the Korea Creative Content Agency, led by the Electronics and Telecommunications Research Institute (ETRI) with DeepBrain AI participating as a joint research institution.

In a workshop published on October 11, 2026, from the AI Engineer Conference, Michael Forrester of Accenture and Whitney Lee of Datadog presented a deliberately permissive 'burrito ordering assistant' AI agent. The demonstration aimed to make the concept of 'excessive agency' tangible for the audience: a friendly business chat interface, if lacking strict application boundaries, can become a route for attackers to access tools, files, and cluster operations.

The workshop progressed through three configurations. The first left the bot exposed, demonstrating how an external container image could be deployed to steal a 'secret recipe' from its Kubernetes environment. The second added infrastructure-level controls, including Kyverno rules restricting image sources and deployment requirements. However, the presenters showed that intended runtime protections did not consistently stop access to sensitive data, prompting a distinction between what worked and what still needed fixing.

The third environment focused on AI-specific guardrails, allowing the toggling of LLM Guard input and output filtering. Comparisons showed responses that exposed sensitive content versus those that redacted it or blocked requests before model processing. Additionally, the demo revealed a subtle trust boundary issue where a weather tool's description contained instructions to read internal configuration alongside legitimate tasks.

Forrester and Lee emphasized that model refusals change between attempts, meaning an application's security boundaries cannot depend on the model choosing to refuse. They used OpenTelemetry traces in Datadog to visualize prompts, tool calls, and results, making this behavior inspectable. The final discussion connected agent security to established CI/CD, identity, network, and artifact controls, asking when an AI workflow has clear boundaries, a worthwhile cost, and verifiable outputs.

Taiwan-based market research firm TrendForce announced a forecast on October 22, 2026, stating that global server CPU shipments are expected to increase by more than 40% year-on-year in 2027. This growth is driven by the expanding inference demand associated with the widespread adoption of AI agents. The firm noted that server CPU demand has been rising since the second half of 2025 as AI service providers accelerate their deployment of AI agents, projecting a 31% year-on-year shipment increase for 2026.

In terms of architecture, TrendForce forecasts that Arm-based CPUs will exceed a 20% market share in 2026 and reach approximately 30% in 2027. This shift is supported by the continued expansion of custom chips such as AWS's Graviton, NVIDIA's Grace, and Google's Axion, alongside the mass production of Microsoft's Cobalt 100 and Arm's own AGI processor. Conversely, Intel's market share is projected to decline from 50% in 2026 to 42% in 2027, hindered by shipping delays caused by new process node and packaging issues, as well as internal capacity constraints. AMD, while starting small-scale production of its next-generation Venice CPU in 2026, remains reliant on existing Genoa and Turin models; constrained by TSMC's tight capacity, its market share is expected to slightly expand to around 27%.

In the video "Let Your Agent Cook: Using Skills to Evaluate and Improve Your App" published on the AI Engineer channel on October 11, 2026, Ankur Duggal, an engineer at Arize AI, demonstrated a workflow for diagnosing and improving LLM applications using coding agents. The video features a financial agent that generates seemingly reasonable but actually context-lacking investment advice, with its own evaluation system flagging issues such as wrong summaries, repeated tool calls, and fabricated answers.

Duggal showed how to connect these failure phenomena to a coding agent via Arize's Command Line Interface (CLI) and skills. The coding agent was tasked with investigating traces to identify specific interactions between supervisor agents, specialist agents, model calls, and tools. This process transformed vague quality complaints into evidence about exact failed steps. For instance, in an earlier experiment targeting stock queries outside the United States, Duggal pointed out that changing the country context alone was insufficient; the underlying tool also had to be updated to resolve the issue.

The video also discussed leveraging dataset, evaluator, and experiment skills to make proposed fixes testable rather than performing a series of unverified edits. Duggal addressed questions regarding local experiment execution, judge model selection, evaluation costs, sampling strategies, and the risk that a coding agent might spend effort on linting instead of addressing actual defects. At the close of the video, the live repair was still ongoing.

For developers seeking to improve AI agent reliability, this resource offers a concrete operational perspective. Duggal recommended human review of proposed changes and tailoring evaluation systems to specific application failure modes, ensuring that performance regressions can be detected promptly as traffic patterns and base models evolve.