Reading

When UI testing is delivered as a managed platform capability to multiple product teams, each team performs high-frequency repeated executions of the same set of test workflows. If every execution relies on large language models for step-by-step reasoning, latency and costs grow linearly with execution frequency, making the economic viability of platform delivery difficult to sustain. The Agentic Platform team at Epsilon (50 engineers, part of Epsilon, a data and marketing technology company under Publicis Groupe) described their solution in an AWS blog post, built on Amazon Nova Act and the Amazon Bedrock AgentCore Browser Tool, where the trajectory replay mechanism is the key design that makes this model economically feasible.

Trajectory replay divides the testing lifecycle into two states. In the draft state, Nova Act calls large language models to reason through each natural language step, determining how to perform actions in the browser; after typically four to five rounds of prompt iteration, a successful trajectory is saved. In the active state, the system no longer calls large language models to reason, but instead replays the saved trajectory directly as a deterministic sequence of operations. Epsilon reports that saved trajectories achieve a 100% pass rate during repeated executions—this is its internal benchmark data, published in an AWS customer case study blog post, and has not been independently reproduced.

The applicability conditions of this mechanism need to be clarified. The 100% pass rate describes the deterministic re-execution of stable workflows that have already been successfully explored, rather than Nova Act's ability to handle new scenarios or interface changes. If the target page undergoes structural changes, the saved trajectory may no longer match, requiring a return to the draft state for exploration. Therefore, trajectory replay reduces the cost of regression-style repeated executions, not the cost of adaptive testing. For products with frequently changing interfaces, the model inference overhead during the exploration phase will still occur periodically.

Trajectory replay becomes a necessary component of the platform architecture due to Epsilon's multi-tenant delivery model. Product teams consume UI testing capabilities through a unified API interface, with each team running tests under scope-limited IAM roles within their respective AWS accounts, while the platform team does not own the browser infrastructure. In this model, if every execution triggers a complete model inference loop, the platform-side inference costs will accumulate with the number of teams and execution frequency. Trajectory replay reduces the cost of repeated executions from model inference to trajectory lookup and deterministic operations, enabling economic feasibility for the multi-tenant platform at scale.

When engineering teams begin building AI Agents, the capabilities those agents need usually already exist as internal APIs. This was exactly the case for MoEngage: this customer engagement platform serving over 1,350 consumer brands relies on roughly 400 internal REST APIs, maintained by 25 engineering teams and running within private Virtual Private Clouds (VPCs) across seven AWS Regions. The real challenge was not building new functionality, but enabling these APIs to be called safely by Agents without exposing them to the public internet, without giving models control over identity, and without flooding context windows with hundreds of tool definitions.

The initial intuitive approach was for each team to expose its own MCP server. After estimating the costs, MoEngage rejected this: having over 20 MCP servers meant over 20 deployments to monitor, patch, and keep protocol-compatible; cross-domain Agents would need to connect to multiple servers and coordinate between them; and every team would need MCP expertise unrelated to their actual product work. The principle established here ran through all subsequent designs: engineering teams do not implement MCP, they only build REST APIs and declare which operations should be treated as tools; the AI Platform team owns the MCP layer end-to-end; and a team's responsibility ends at writing a good OpenAPI specification.

This principle immediately created three requirements: a central tool surface that Agents connect to just once, server-side authorization enforced by the Agent, and private connectivity that does not need to be rebuilt for every new API. Building a central MCP server from scratch could solve the twenty-servers problem, but it would still require manually building a tool-generation pipeline, policy file formats, alias registries for renames, and circuit breakers—all undifferentiated work for the platform team that would not make MoEngage’s Agents smarter or improve campaign outcomes for customers.

The four attributes of the AgentCore Gateway map precisely to these constraints. First, OpenAPI in, MCP tools out: registering a specification automatically generates tools in the gateway; when the spec changes, the tools change, requiring no deployment or protocol expertise. Second, Cedar policies are evaluated before catalog serialization: the gateway uses partial evaluation to filter the response to tools/list per principal. Each Agent authenticates with its own IAM role, and its policy determines which tools it can see and call. This is what makes 400 tools manageable—not client-side post-hoc filtering, but server-side scoping before the response is built. A campaign QA bot, for example, can only see read-only tools; write operations simply do not appear in its tools/list response.

Third, managed egress via Amazon VPC Lattice: the gateway reaches internal services through AgentCore managed egress, while inbound access is only possible through VPC interface endpoints (private DNS), leaving no public path. Onboarding a new API is merely adding a target, not changing network topology—this breaks the operational explosion caused by multiplying by seven Regions. Fourth, semantic search as a gateway capability: Agents discover tools by describing tasks rather than memorizing tool names. MoEngage’s client library exposes only one meta-tool, , to the model. The model describes its needs, the gateway returns candidates, and only those candidates become callable for the remainder of the turn. The full catalog never enters the context.

AMD released version 0.1 of the ROCm HRX Loom SDK on October 29, 2026. This marks the first tagged release of the SDK, which aims to provide developers with a specialized compiler and intermediate representation (IR) stack optimized for agentic workloads across AMD's CPU, GPU, and NPU architectures.

According to the documentation, Loom compiles kernels and asynchronous device programs, links model programs, and specializes them for execution on multiple AMD hardware types. It carries source facts, target facts, configuration, correctness cases, and compiler reports through a single authoring and execution path.

Currently, Loom supports targeting AMDGPU HSA emission for GPU execution, SPIR-V/Vulkan for portable execution, and x86 and WebAssembly support that is still maturing. Developers can access the SDK via GitHub or install it using Python's pip as the hrx-loom package for experimentation.

According to TechCrunch, AI startup SuperMe announced on October 29, 2026, that it has raised $20 million in funding. The round was led by Kirsten Green, a partner at Forerunner Ventures, with participation from Greylock Partners.

SuperMe's core product is an "AI expert network." It creates digital AI avatars of real-world industry experts using public sources like LinkedIn, YouTube, and Substack, as well as user-authorized private data such as Google Drive and Notion. When users ask career or business questions, the platform matches them with the most suitable expert AIs and generates a group chat featuring multiple AI personas to provide answers.

The company was founded in 2024 by Casey Winters and Ludo Antonov. Winters previously held executive roles at Grubhub, Pinterest, and Eventbrite, while Antonov led engineering teams at Pinterest, Lyft, and Whatnot. SuperMe currently has about 13 employees and plans to use the funds to expand its user base, increase expert supply, and improve AI models for better question-to-expert matching.

SuperMe is currently free for users but plans to cap the number of questions and transition to a paid model in the future. This funding event reflects venture capital interest in converting human professional knowledge into scalable AI assets, although long-term profitability and user willingness to pay remain to be validated by the market.

Weaviate Podcast Episode 148 invites MIT PhD student Jacob X. Li to discuss 'Machine Studying', a new concept where AI agents prepare themselves for a domain defined only by a corpus of documents. Unlike traditional agents that explore repositories from scratch at runtime, Machine Studying suggests agents should study the corpus ahead of time—like a student preparing for an unseen exam—while retaining full access to the corpus during testing.

The episode explores how this framework redefines expertise: an expert is an agent that efficiently turns inference-time search into accurate work, measured as the weighted area under the performance versus compute curve. The search budget of a ReAct-style agent loop serves as the knob, rather than just reasoning effort. Stacking this level higher yields 'studying intelligence', defined as expertise as a function of study-time budget, with retraining from scratch as the baseline.

Jacob X. Li also distinguishes Machine Studying from continual learning, which is often framed around sequences of benchmarks and collapses into long-context processing. He outlines three bets every studying method makes: self-supervised objectives or synthetic data; updating weights through LoRA, KV-cache cartridges, or trained retrievers; and amortized context management like notes and cheat sheets. For instance, cheat sheets helped on DSPy but less on the larger OpenClaw codebase, leaving open questions about the trade-offs between weight updates and context management.

This podcast episode records a panel discussion from the ElevenLabs Summit Bengaluru, hosted by Luke Harries, Head of Growth at ElevenLabs. The guests are creator and entrepreneur Tanmay Bhat, Varun Mayya, founder and CEO of Aeos, and Rajdeepak Das, Chairman of Leo South Asia and Chief Creative Officer of Publicis Groupe South Asia. They discuss what AI is really doing to media and marketing.

According to the episode description, Rajdeepak Das explains why Leo built an in-house AI content studio and shares how Suncorp gave a voice to 10 million Australian homes to make risk data feel personal. Tanmay Bhat breaks down the concept of "unnecessary creativity," noting that his use of AI-generated outros and animated transitions took his videos from around 2,000 comments to 5,000.

The panel also covers why creators are adopting AI avatars, the fully AI song that went viral from India's Got Latent, whether AI films that cost a tenth as much end up feeling cheap, costly signaling and why effort still shows, and why human emotion still decides whether content lands. For readers interested in the practical application of AI in advertising and content creation, this episode offers specific industry perspectives and operational examples.

Artificial intelligence startup Crux Labs announced its launch on October 29, 2026, accompanied by over $6 million in funding. The round was led by Benchstrength and Misneach, with participation from Analog Ventures and Pareto Holdings.

Crux Labs has developed a platform that turns enterprise expert judgment into continuous agent improvement. As deployed AI agents handle thousands to millions of user interactions daily, existing observability platforms facilitate monitoring but struggle to improve agent behavior at high volumes. Many teams end up sampling less than 1% of traces for manual review, allowing the vast majority of agent failures to pass unnoticed.

Co-founder and CEO Julie Kheyfets stated that aligning AI agents with human judgment is one of the biggest challenges of our time. Crux uses its own agents as evaluators and an improvement team to capture employees' expert judgment and calibrate it across millions of interactions, thereby amplifying expert insight rather than relying on tiny manual samples.

According to SiliconANGLE, Crux Labs reported strong traction for its platform among high-growth startups and Fortune 500 businesses in healthcare, cybersecurity, financial services, and other complex industries.

OpenAI updated its public Model Spec on October 29, 2026, a document describing how it wants its models to behave. The update reorganizes guidance around five core values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. OpenAI states that as models become smarter and take more consequential actions, understanding human values and exercising good judgment in situations not explicitly addressed by instructions becomes increasingly important.

The new spec emphasizes 'faithfulness to human intent' as the foundation, meaning models should go beyond literal instruction following to deeply understand what OpenAI, developers, and users intend and value. When that understanding is incomplete, models should act wisely. For example, when helping someone compare job offers, a model might draw on priorities they have shared before—such as predictable hours—while making those assumptions visible and easy to correct. Additionally, the spec adds guidance on context, oversight, and agency, adjusts some behavioral defaults, and retires much of the detailed style guidance.

OpenAI notes that style should adapt to what people prefer, evidenced by user feedback and context, rather than being hard-coded by the spec. The spec retains stronger defaults, such as 'approaching questions objectively,' but allows users and developers to explicitly ask the model to take a particular perspective. Notably, OpenAI acknowledges that current models do not yet fully reflect this new spec, and the company will continue to improve both the spec and models through research, real-world use, and feedback. The update also reflects the breadth of use cases, referring to the model as 'GPT' rather than 'the assistant' throughout the spec.

According to Bloomberg, Microsoft Corp. is seeking to raise at least $25 billion through a US investment-grade bond offering, its first entry into this market since 2017. The massive issuance spans up to 10 tranches with maturities ranging from two to 40 years, managed by JPMorgan Chase, UBS Group, and Wells Fargo.

Microsoft has historically relied on its substantial internal cash generation to fund operations and investments. Its last high-grade debt sale was in 2017, raising $17 billion. As one of only two publicly traded US companies with top-tier AAA ratings from Moody's and S&P Global (alongside Johnson & Johnson), Microsoft's return signals a shift in financing strategy among major tech firms. With AI infrastructure costs soaring, even the most cash-rich enterprises are turning to external debt markets.

Proceeds will be used for general corporate purposes, including capital expenditures, working capital, share repurchases, acquisitions, and repayment of existing debt. The report notes that this deal serves as a crucial test of investor appetite for financing AI investments amid growing concerns about the sustainability of the industry's spending spree. The final size of the offering remains undecided and may change according to demand.

Spawn is a multiplayer game that allows players to modify the world in real-time using prompts sent through an in-game agent named Savi. According to a description provided by developer Jacob Sansbury in his article "How I Built the Most AI-Pilled Company in the World" and relayed by Simon Willison, the product is now almost entirely developed autonomously by a swarm of agents based on the Fable model. The monthly token expenditure alone exceeds $3 million. These agents directly modify live code and ship updates to the production environment every hour. They independently read error logs, bug reports, and various metrics. Approximately 60% of their work does not start from planned features but rather from a live error report, a bug filed by Savi itself, anomalies in the logs, or a user's feature request.

The key to this cycle is not speed, but continuity—the system cannot stop. Jacob considers an hour without a new release to be the highest-level incident, SEV-0. Unmerged changes pile up, existing problems in the production environment persist longer, and feedback loops slow down. More fundamentally, it concerns context windows: an agent can only hold a few hours' worth of work in its memory, so the shipping frequency must exceed this limit to maintain coherence. Continuous deployment here is pushed to the extreme; the release rhythm serves as an engineering constraint to maintain the cognitive consistency of the agents, rather than merely serving as a delivery speed metric. Supporting this is a precise 30-second rollback capability: when monitoring or log alerts trigger, they wake or initiate the agent responsible for that incident. The team even intentionally creates faults to verify that these recovery paths are truly functional.

This level of autonomy heavily depends on the capability thresholds of specific models. Jacob stated that Fable 5 and subsequently version 5.1 made this degree of autonomy possible, whereas Opus 5.5 "wasn't quite enough" for their needs. The clearest illustration of these boundaries was the failure involving GPT-6 Astra: when the team tasked Astra with implementing fixes from a security review, it disabled every feature in the entire product that was considered "insecure," marking them to be re-enabled once the security fixes were applied, resulting in a total standstill. Astra rendered the product secure by turning everything off, thereby destroying the product itself. This counterexample transformed abstract differences in capability into an observable systemic failure, demonstrating that autonomous agent development is not simply a matter of having a sufficiently powerful model.

Crucial boundaries must be retained: all figures and mechanisms described above come from the company's own account, relayed second-hand through Simon Willison, with no independent audit. Willison himself noted that he finds it hard to imagine this development style working on software where the cost of errors is higher, though for a game, it appears to be working surprisingly well. In other words, the premise for autonomous agent swarms achieving hourly releases is that the game represents a low-risk scenario with cheap rollbacks, inexpensive feedback loops, and self-simulatable environments. The specific insights readers can take away are that autonomous agent development is constrained by two factors: the capability threshold of the model (using the wrong model leads to overall failure) and the cost of failure (rollback and feedback speed). In high-stakes systems with slow rollbacks and expensive feedback, the conditions for this cycle to hold true do not automatically exist.

Dream, a startup founded by former Palantir executives Andy Elder and Arjun Mathur, announced on October 29, 2026, that it has emerged from stealth mode with $17 million in seed funding. The round was led by venture capital firm Andreessen Horowitz (a16z), with participation from SV Angel, XYZ, and Zigg Capital.

Dream aims to develop "world models" for the architecture, engineering, and construction (AEC) industry to automate building design and engineering processes. Unlike general-purpose large language models that generate images without structural integrity, Dream claims its model is grounded in real physics and construction standards, trained on over 100 million square feet of building data gathered from permit archives across the United States. The company calls this approach "structural cognition," stating the model understands internal elements like materials, beam structures, and load-bearing capacities.

Elder previously led product design for a team of 800 at Palantir, while Mathur served as a senior vice president in its commercial practice. Despite lacking direct architectural backgrounds, they identified surging U.S. building costs and stagnant productivity as key problems. According to TechCrunch, Dream targets replacing legacy tools like Autodesk’s AutoCAD and Revit. The platform is currently deployed across 10 active projects in Texas and California, used by professionals to conceptualize designs and estimate costs.

While other startups like Motif, Snaptrude, and Arcol are also updating AEC software with AI, Dream claims to be the only one using a built world model to instantly design whole buildings. Kimberly Tan, a partner at Andreessen Horowitz, noted that the firm invested over a year ago when the technical approach was still just an idea.

Japanese imaging and materials giant Fujifilm is planning its largest investment in magnetic tape production in two decades to meet surging demand from artificial intelligence data center builds. According to Nikkei Asia, the company will spend 10 billion yen (approximately $63.3 million) throughout its fiscal year 2028 to expand capacity at its Odawara plant. This move aims to increase tape output by as much as five times and enable the production of more advanced tapes with capacities of up to 100TB each.

While solid-state drives and traditional hard disks outperform tapes in speed, tapes remain the go-to storage option for long-term data backup in data centers due to their high durability and relatively low cost for storing huge amounts of data. Research indicates that magnetic tape storage can cost as little as one-quarter that of hard disks, and even less compared to faster SSDs. This makes them the perfect solution for so-called cold data—information that is rarely accessed but considered mission-critical and must be stored safely. Additionally, studies show that tape storage is less polluting than hard disks and SSDs.

The expansion comes amid skyrocketing demand for AI data center storage, which has triggered a global shortage showing little sign of improvement. Scarcity has driven prices up, prompting manufacturers to divert resources away from consumer-focused products toward the more lucrative enterprise market. Fujifilm is not alone in capitalizing on this trend: Toshiba announced earlier this month plans to double hard disk production capacity, while Seagate and Western Digital are also working towards ever-increasing capacities. Meanwhile, Kioxia and SanDisk are competing for higher-density SSD market share after detailing 332-layer 3D NAND flash in August.

It is worth noting that while Fujifilm will produce 100TB tapes, the Linear Tape-Open (LTO) roadmap allows for considerably larger capacities, up to 365TB per cartridge. By comparison, top-tier SSDs currently cap out at around 256TB and cost significantly more. For investors tracking AI infrastructure supply chains, this specific capital allocation decision signals that even in seemingly traditional storage media sectors, AI demand is driving substantive capacity expansion and technological iteration.