AETERNAE AI RESEARCH LLC INDEPENDENT RESEARCH
ÆAETERNAERESEARCH
Sign inRequest access
Æ / PUBLIC MALWARE INTELLIGENCE

Follow the evidence.

Published file metadata, source observations, indicators, and research history in one bounded catalog.

2,400+Reachable catalog records
92 / 100Current catalog page
24Maximum records per page
0Samples downloaded by this page

Feed collection status

Sources represented on this page. Publication and observation dates describe the content; successful batch dates describe local collection. A recent batch does not certify complete coverage or current provider availability.

2 sources: 0 stale, 0 needing attention, 0 paused, 0 unverified.

Review source update dates and status
  • intelligence_src_f28777e1849bec7cf0Recent successful batchLast successful batch: Last attempt:
  • intelligence_src_f2f60dee08521b7ce6Recent successful batchLast successful batch: Last attempt:

Stale means older than twice the recorded collection interval, with a minimum of one hour; without an interval, the threshold is 24 hours. Cached content remains available while a feed is stale, paused, or unverified.

Clear filters
2,400+ MATCHING RECORDSMETADATA ONLY / SOURCE ATTRIBUTION PRESERVED
REPORT1 SOURCE

Rethinking Robot Safety in the Age of AI

This article is brought to you by VicOne.Robot safety has traditionally asked: Can a machine remain safe when something goes wrong? Physical AI raises a harder question: Can a machine remain safe when an attacker changes what it sees, decides, or does even when nothing appears to have failed?As AI and robotics continue to advance at an unprecedented pace, modern robots perceive through multimodal sensors, interpret context using AI models, and translate those interpretations into physical action. As they move into dynamic environments, their safety increasingly depends on the integrity of the data guiding their decisions.That dependence creates risks that conventional safety assessments may not fully capture. Recent research has demonstrated that manipulating what a robot sees, hears, or interprets can influence its behavior without requiring direct control.Such manipulation can occur anywhere across its complex sensing and decision-making system — a layered attack surface encompassing training pipelines, system infrastructure, and runtime perception.Layer One: Corrupting intelligence at its sourceIn 2017, BadNets demonstrated that a model could behave normally under most conditions, yet fail in the presence of a specific hidden trigger. In one example, a subtle pattern caused a stop sign to be misclassified as a speed limit sign without affecting the model’s behavior on other inputs.What began as a classification vulnerability has since evolved into action manipulation.At NeurIPS 2025, researchers introduced BadVLA a backdoor attack targeting Vision-Language-Action (VLA) models that allow robots to see, interpret instructions, and produce coordinated physical movement. Rather than altering a single label, the attack caused conditional deviations in the robot’s action trajectory when a trigger was present. Without the trigger, the model largely preserved normal task performance, while the backdoor remained effective under task transfers and model fine-tuning.A related study in 2025, GoBA, showed that ordinary objects such as a coffee mug could serve as a reliable trigger. The researchers reported a 97 percent attack success rate without degrading performance on clean inputs.A critical safety question today is whether Physical AI models remain within their task and safety boundaries under adversarial conditions.These studies expose a blind spot in model validation: A model may pass testing yet produce corrupted behavior when a hidden trigger appears in operation.So a critical safety question today is whether Physical AI models remain within their task and safety boundaries under adversarial conditions. Simulation tools such as NVIDIA Isaac Sim, when paired with VicOne Radeis, can test the effects of manipulated inputs before deployment. VicOne LAB R7 demonstrates Radeis, a Physical AI safety validator for NVIDIA Isaac Sim that tests how adversarial visual inputs affect robot behavior before deployment.VicOneLayer Two: System vulnerabilities as gateways to AI controlEven a securely trained model can be subverted if the surrounding system stack is vulnerable.In September 2025, researchers disclosed UniPwn, a Bluetooth exploit chain affecting quadruped and humanoid robots from a major manufacturer. Hardcoded cryptographic keys allowed traffic decryption, authentication checks were bypassed, and command injection enabled root-level execution. The exploit is also described as “wormable.” A compromised robot could scan nearby units and potentially affect an entire fleet. VicOne Lab R7’s demo shows how chaining three wireless exploits can trigger uncontrolled robot behavior within 60 seconds, resulting in operational disruption.VicOneMiddleware creates another exposure point. Vulnerabilities in ROS 2 and DDS-based systems can enable arbitrary code execution or abuse unauthenticated topics to deliver malicious commands. With sufficient access, an attacker could override motor commands or replace AI model weights without d

First reported Sep 16, 2026 16:51 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Why Read a Research Paper When You Can Turn It Into an AI Agent?

Have you ever read a paper in Science or Nature and thought, “Man, that research was so cool. I wish I could try that method on my own data,” only to spend a week wrestling with someone else’s undocumented repo, broken dependencies, and half-finished readme.txt?Well, now you can, more or less.Say hello to Paper2Agent, a new open-source framework that transforms academic reports into interactive AI agents you can talk to. Give it a paper, along with the accompanying codebase, data, or other supplementary material, and the system automatically extracts the core workflows, then spins up a tested, runnable toolkit that you can use on your own datasets.The concept may sound a little like Google’s NotebookLM (now called Gemini Notebook), which lets you upload documents and chat with an AI about what’s in them. But Paper2Agent aims to go a step further: Rather than simply answering questions about a paper, its agents can actually run the methods described in it—and potentially combine those methods with tools from other papers.The goal, explains Stanford computer scientist James Zou, is to change what a scientific paper fundamentally is. “Knowledge should not be static records,” Zou says. “It really should be dynamic and interactive—and this has many benefits, including making knowledge more reproducible but also enabling all sorts of new kinds of discovery.”Zou and his colleagues described the tool 16 September in Nature. They tested Paper2Agent across diverse disciplines including statistics, econometrics, and astrophysics. However, the researchers focused their proof-of-concept demonstrations on computational biology, where turning published methods into usable tools can be particularly cumbersome.From Paper to PromptThe team started with AlphaGenome, a deep-learning model that predicts how mutations in DNA affect gene regulation. (A companion resource unveiled earlier this month, the AlphaGenome Atlas, cataloged the model’s predictions for all 9 billion possible single-letter changes in the human genome.)The researchers fed Paper2Agent the corresponding documentation and code. About 45 minutes later, with no human intervention, the AI had produced 22 tools covering different aspects of AlphaGenome’s functionality, all on a personal laptop and for less than US $15 in computing costs. One tool, for example, could predict how a DNA change might affect gene activity, while others could compare those effects across tissues or analyze multiple variants at once.“The idea of making papers more dynamic and executable through an agentic interface is quite compelling.” —Dongping Chen, University of MarylandAll 22 tools passed automated validation, thanks to a testing agent working behind the scenes to run the various AlphaGenome sub-tools against reference results. If a test failed, the agent would diagnose the problem and try to fix the tool, with up to six attempts per function. If that didn’t work, it could drop the tool altogether.The validated tools were then packaged into a Model Context Protocol server and connected to Claude Code (though any compatible chat-based AI assistant could do). The result was a user-facing AlphaGenome agent that could take questions in plain English, run the appropriate analyses, and spit back results and visualizations.The team then put the agent through its paces, hitting it with a battery of questions, ranging from simple requests to open-ended research problems. According to the researchers’ analysis, it outperformed both standard Claude given the AlphaGenome codebase and a specialist AI co-scientist tool called Biomni.Multi-Agent Research Paper CollaborationGoing one step further, the researchers turned a couple of more papers into interactive agents and linked them with the AlphaGenome agent. (One of the additional papers was on how inherited DNA variants linked to autoimmune disease disrupt cell function, while the other was a more systematic exploration of how sil

First reported Sep 22, 2026 15:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

6 Guidelines for Governing AI

For the first 10 years of my career, I worked in product management and data analytics by myself. I wrote database queries that pulled numbers out of corporate systems, built statistical models to predict what customers would buy, and shipped data pipelines that moved information between business systems.I built and scaled analytics teams at Best Buy and Target, studying how customers shop and what stores should stock. Today I lead enterprise AI transformation at Lowe’s, the Fortune 100 home improvement retailer.The goal is not to sell artificial intelligence; it is to use it to deliver useful expertise at the moment a customer needs it. In retail and other customer-facing industries, virtual assistants can help people address everyday questions—such as how to repair a leaky faucet—while guiding them toward relevant products, services, or next steps. As these capabilities become more common, technology roles are changing. The work is no longer limited to building AI systems; it also includes defining how they operate: which decisions they can make autonomously, when they must escalate to a person, and which actions must remain off-limits.That shift—from building AI systems to governing them—is coming for anyone who is accountable for what such systems produce. Not the casual user typing into a chatbot but the engineers, product managers, analysts, and business operators who sign off on work a machine drafted.It is the subject of the book I recently coauthored, The Enterprise Brain. I call the change the “governor shift,” from executing tasks yourself to setting the intent, principles, and boundaries within systems that execute them for you.Business operators might not write code; they will decide which pricing exceptions an agent may approve and which it must escalate.That is governing.A 2025 report from MIT Media Lab’s Project NANDA found that, despite an estimated US $30 billion to $40 billion in enterprise generative-AI investment, the vast majority of organizations in its dataset had not yet demonstrated measurable profit-and-loss impact. The report estimated that only about 5 percent of integrated pilots were generating substantial value, underscoring how difficult it remains to move from experimentation to scaled business outcomes.Researchers named the pattern the GenAI Divide, the term I adopted for the book.The companies rarely lack technology; they use the same models as the 5 percent that are winners. But they lack people who can direct the systems and stand behind the results. Guidelines to followHere are six guidelines.Recognize when you have become “human middleware.” In software, “middleware” is the code that sits between two systems and passes information back and forth. Many of us have become its human version. Take an honest look at your week. How much time is spent pulling data out of one tool, reformatting it, and routing it to another team? I call this the “administrator trap,” which is set by the architecture, not by the people caught in it.Relaying is what AI agents now do well. But they cannot judge which numbers deserve attention, which risks are real, or which compromises are worth making.Trade rules for principles. For many years, workers used rules to manage their work. Refunds for a product over a certain amount needed a signature from upper management, for example. Writing code needed two reviewers. Rules work at human speed. But rules break when a system makes thousands of decisions per hour and meets situations no rulebook anticipated, such as a complaint covered by three different policies. A rule says to do exactly this specific thing; a principle says to achieve the outcome without crossing certain lines.Governing AI means writing those principles in priority order so the system settles its own conflicts the way a well-led team does when the manager is not available. Never harm the customer. Tell the truth even if the company loses a sale. Protect the economics, and

First reported Oct 05, 2026 18:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

AI Efficiency Could Cost Us the Next Generation of Experts

A little over a decade ago, I led the controls design for a first-of-its-kind full digital-control system for a U.S. nuclear plant. It was, on paper, a beautiful machine—engineered to run itself the way a modern airliner does, with operators watching over a system that rarely needed them. And we made a decision that, to an efficiency-minded observer, looked backward: We deliberately left manual steps inside sequences the system could execute on its own.We were solving a specific problem. An operator who only ever supervises automation slowly stops being an operator. The hands go cold. The mental model of what the plant is actually doing gets fuzzy. Then comes the day the automation hands control back. It’s always the worst day, because automation only quits when it’s confused or in trouble. But by then, you have a person in the chair who hasn’t truly operated the thing in years. The manual steps were there to keep the human current. It was inefficient by design, on purpose.That plant, as it happened, was never built. It was shelved amid the politics and economics that surround nuclear power in this country, for reasons that had nothing to do with the engineering. But the design instinct outlived the project, and I’ve come to believe it’s the most useful idea I can offer to the argument now consuming every boardroom: What happens to human expertise when AI does the work that used to build it?AI Is Disrupting the Engineering Career LadderThe data has gotten hard to wave away. A Harvard University working paper covering some 65 million workers at more than 280,000 U.S. firms found that after companies adopted generative AI, junior employment fell roughly 9 percent within six quarters relative to nonadopters, while senior employment kept right on growing. A Stanford analysis of ADP payroll records points the same way: The youngest workers in the most AI-exposed occupations lost ground after late 2022 while their more-experienced colleagues held theirs. The Stanford researchers found that the losses concentrate where AI automates the work; where it merely augments, junior employment holds steady or rises.The causal story is still contested, and honesty requires saying so. Researchers at the New York Fed attribute much of the rise in young-graduate unemployment not to AI but to remote work, arguing that firms are reluctant to hire inexperienced people whom they cannot train and mentor at a distance. But notice what the explanations share. Whether a model is absorbing the formative work or distance is severing the mentorship around it, both describe the same broken mechanism: the apprenticeship channel through which expertise passes from senior to junior. Either way, “entry-level” has quietly come to mean “three years of experience required.”Strip away the noise and you’re left with one deceptively simple problem: You cannot become a senior engineer without first being a junior one. Expertise is not downloaded. It is earned through failed builds, dead-end debugging sessions, and the “why on earth did that work” moments that a capable AI will now happily spare the newcomer. Spare them enough of those and you produce a cohort that can supervise a model on paper but never developed the gut sense to know when the model is confidently, catastrophically wrong.Most of the commentary stops at the diagnosis, or reaches for policy solutions that treat the loss of junior jobs as an economic problem. Yet it’s also an engineering problem, and safety-critical fields have already spent decades learning how to solve it.Aviation’s Lessons About the Automation ParadoxMy own career started at the sharp end of automation. My first job out of school was verifying and validating the software in the digital jet-engine controller that decides, faster than any pilot could, how a fighter plane’s engine responds. Even then, in the late 1980s, the central tension was visible: The machine outperforms the human in routine cases, bu

First reported Sep 02, 2026 13:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

HiPHI: A Large-Scale Benchmark for High-Precision Human Motion and Object Interaction

This White Paper gives robotics researchers and engineers an overview of a new large-scale motion capture dataset built to close the data gap limiting humanoid robot learning. It also shows how policies trained on the dataset transfer to a real humanoid robot.What you will learn about: Why humanoid robot learning, a central problem in embodied AI and Physical AI, needs data that internet video and existing motion capture datasets cannot provide. How FrameNet, a linguistic framework for human action, can guide motion capture collection to systematically cover a broad range of whole-body motion. Why synchronized object trajectories and meshes make human-object interaction data useful for teaching robots real-world tasks such as carrying, pushing, and pulling. How reinforcement learning policies trained on this motion capture data improve with scale, and how sim-to-real transfer carries them onto a physical humanoid robot.Download this free whitepaper now!

First reported Oct 07, 2026 13:49 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Attempts to Keep Humans in the AI Loop May Actually Push Them Out

A crucial safeguard against AI agents going rogue—keeping humans in the loop to review and approve their decisions—will fail unless designers and users change their current practices, a trio of leading AI ethics researchers argue.Though most autonomous agents have systems to keep users in the loop about their actions, in practice these processes actually push humans out of the loop, the authors argue in a paper posted to ArXiv on 6 September. In other words, “the human just becomes this meat tool to give permissions without the cognitive capability to engage,” says one of the authors, Avijit Ghosh, the lead technical AI policy researcher at Hugging Face, an open-source machine learning platform.In the near term, the paper says, humans being out of the loop leads to agents acting in ways people don’t know about or want (like July’s hack of Hugging Face by a swarm of OpenAI bots). In the long term, it will cause users to lose the “cognitive capacities” they need to control AI, write the researchers, which aside from Ghosh include Margaret Mitchell, Hugging Face’s chief ethics scientist, and Samir Passi, an affiliate of the Data and Society Research Institute.Though the trio began working on the paper before the Hugging Face hack was disclosed, their conclusions result from “logically thinking about what is going to happen if the current trends continue—and we saw that being precipitated via the Hugging Face attack,” Ghosh says.Instead of focusing on human oversight, Ghosh says, many in the field believe AI can monitor AI. “They’ll say, ‘Oh, we have this other LLM tracking the logs.’ But [without a human in the loop] how do we know that these two LLMs are not scheming together?”Even if these problems are new to artificial intelligence researchers as the industry rolls agents out into the world, the challenges are familiar to researchers in adjacent fields, like robotics and autonomous vehicles, notes Mary L. Cummings, the director of George Mason University’s Autonomy and Robotics Center. Cummings has spent decades investigating how people interact with autonomous systems.“While I appreciate what the authors are trying to say, they just use a lot of academic words to say AI companies should care about human factors,” she wrote to IEEE Spectrum in an email. AI developers are “late to the party” in focusing on “cognitive engineering.”Flaws in AI Agent Design and OversightThe key flaw in current agent design, the authors argue, is that the bots are tailored to meet benchmarks like speed, accuracy, and volume of work performed. The needs of human overseers are treated “as a separate consideration independent of the quality of the system.”As a result, agents often overwhelm human overseers with more information than they can comprehend. As an example (not mentioned in the paper), the 1,200 bots involved in the Hugging Face attack generated 1.2 million messages on their improvised messaging system. (Hugging Face was recently acquired by Nvidia, which announced its own hardware-and-software-based approach to controlling AI agents on 28 September. Ghosh declined to comment on possible future impacts of the merger, noting that the two organizations remain separate until the merger process concludes. A system that really kept humans in the loop would accommodate the human mind’s built-in biases, the authors write: “With automation bias, users accept system suggestions even when they are wrong. With anchoring bias, people are more likely to agree to an AI system’s decision” without thinking of alternatives. Effortful reasoning doesn’t feel as good to users as does quickly approving an AI’s plan—especially if they feel overwhelmed. And AI’s sycophantic manner tells users that they’re doing well, which undermines the skepticism and self-monitoring that oversight requires.To address the problem, the authors write, agent developers should introduce friction in human-AI intera

First reported Oct 05, 2026 15:57 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Responsible AI for Higher Education

This interactive webinar will introduce the different types of AI, address the concerns with AI, share how we IBM are approaching Responsible AI, and offer guidance to students about what they can do - as individuals, and members of their IEEE chapters. Participants will also have the opportunity to to apply the Responsible AI approach to a particular use case - IBM Bob, a software development life cycle agent, and Q&A. This will be an interactive session, so have phones ready to engage! Register now for this free webinar!

First reported Sep 14, 2026 14:18 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy

If you’ve ever wondered how Claude and similar AI agents can answer complex questions in seconds and where they get the information, you’re not alone.Gurmeet Saran works to find that answer every day as an engineer at Anthropic in San Francisco, the AI research company behind Claude. The IEEE senior member is developing an infrastructure to understand the mechanics behind the chatbot’s behavior and to ensure its underlying architecture produces reliable outputs.Gurmeet SaranEmployer Anthropic in San FranciscoTitle Member of technical staffMember grade IEEE senior memberAlma maters University of Kalyani, in India; Birla Institute of Technology and Science, in Pilani, India “My job is to make sure the data is accurate, structured, and trustworthy, and that private data stays private,” he says. “As I learned earlier in my career, privacy can’t be a layer you bolt on—it has to be part of the architecture from the start.” If you can’t trust the logic, he adds, you can’t make reliable claims about the behavior, and that has real safety implications.Safety is a priority for Anthropic, he says. According to Claude’s mission statement, the company wants the AI agent to be “genuinely helpful to the people it works with or on behalf of, as well as to society, while avoiding actions that are unsafe, unethical, or deceptive.” Crucial to the mission is ensuring that researchers can observe and understand how AI models behave, a challenge that is especially acute for large language models, Saran says.“We have decent tools to monitor databases and web services,” he says, “but for LLMs, we’re still figuring it out.”Solving data quality problems at Yahoo, Intuit, and MetaSaran’s interest in computing systems dates back to his childhood in Kolkata, where he logged as much time as he could in the school computer lab.“I remember the first time I used one of the computers,” he says. “What struck me was that you could write instructions, and the machine would follow them exactly. No ambiguity.“That felt magical to me.”After graduating from high school in 1999, he stayed in India to pursue advanced degrees. He earned a bachelor’s degree in electronics and telecommunication engineering in 2004 at the University of Kalyani. In 2010, he earned a master’s degree in software engineering from the Birla Institute of Technology and Science in Pilani.While a graduate student, he began his career in 2007 at Yahoo as a technical lead of the data team in Bengaluru, India. While there, he encountered the project that shaped his career trajectory: working on web analytics and business intelligence reporting solutions.“We kept running into a frustrating problem,” he recalls. “Different teams would calculate the same website metric—something like daily active site users—using slightly different logic, and end up with conflicting answers. It sounds like a small issue, but when leadership is making decisions about resource allocation and product development using that data, not knowing which number is correct becomes a problem.”The team built a central data repository with defined metrics, essentially serving as a single source of truth. Saran created a model that defined what “daily active site users” means, breaking down which events counted, how to handle edge cases (or when standard code fails), and more.With the new model, data teams across the company could rely on a standard method to determine the number of daily active users for the projects and sites they were responsible for and obtain comparable results using the same logic.From there, Saran and his team built Yahoo! Digits, an internal web analytics tool used for all the company’s products.Saran left Yahoo in 2010 to join Intuit as a staff engineer in Mountain View, Calif. He helped build the analytics infrastructure that unified data across the company’s products, including TurboTax and QuickBooks accounting software. During his te

First reported Oct 12, 2026 18:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

The Future Is Fanless: 100% Heat Capture for Liquid Cooled AI Servers

This article is brought to you by CoolIT, an Ecolab Company.Beyond 250 kW a server rack can no longer be cooled by a hybrid approach of liquid and air. At this density a 70/30 liquid-air split leaves 75 kW of air load. The air cooling system needed to move it brings cost and complexity few operators will accept. The answer is near-total heat capture. Liquid takes effectively all the heat, air falls below 1 percent of the load, allowing the server to run fanless.CoolIT builds these loops today from modular coldplate blocks proven across six generations of fanless designs. Processor thermal design power (TDP) keeps climbing generation over generation. This rising heat load is now cascading into the memory, networking, storage, and power components that once ran comfortably on air.The heat escaped the chipFor years the story stayed simple. Cool the processor and let air handle the rest. That balance has shifted. As TDP climbs, heat spreads outward from the processor and cascades into the components around it. Memory, networking, storage, and power now run hot enough to demand liquid of their own. Engineers designing the next generation of AI servers face a board where heat capture rises with every launch.Beyond 250 kW per rack, air cooling becomes the bottleneck. Near-total liquid heat capture enables fanless AI server designs built for the next generation of computing.New parts, new rulesUnlike processors, which are cooled as flat rectangular packages, these peripherals come in a wide range of shapes, sizes, and mounting requirements, each with its own thermal limits. Some run cooler than the processor case temperature, others run hotter, which leaves them sensitive to a design tuned only for CPUs and GPUs. Operators need purpose-built solutions here, matched to the part rather than stretched across the board.CoolIT engineers meet this with a deep toolkit. Conductive plates, vapor chambers, heat pipes, and thermal transfer plates move heat from components closer to the liquid path. Riding coldplates enable pluggable components. Each solution stays true to the component it serves. CoolIT Customer Showcase: How GWDG Cools HPC & AI Systems with CoolIT’s Direct Liquid Cooling CoolIT One loop, one serverCooling the parts is one challenge. Uniting them is the real work. Full heat capture means folding every one of these solutions into a single server loop that distributes coolant effectively and remains easy to install. Connection reliability, coolant routing, and the time it takes to assemble the loop at rack integration determine whether a design thrives in production or stalls on the bench. CoolIT builds these loops from proven modular blocks, so operators gain performance and deployment speed within the same solution.Density forces the decisionRack power continues to climb toward 1 MW, and the case for liquid grows stronger at every step. A 70/30 split of liquid to air holds comfortably at lower density. Past roughly 250 kW it stops working. The 30 percent left to air becomes a 75 kW load inside a single rack, and moving that much heat demands a parallel air system whose cost and footprint few operators will accept. Adding density only widens the gap.As rack power continues to climb toward 1 MW, CoolIT’s modeling places full heat capture as the standard server design for flagship rack-scale products through 2028.The simpler, more efficient answer is to capture the heat in liquid and drop air to less than 1 percent of the total load. True 100 percent remains almost impossible to reach in the strictest sense, so the honest and achievable target is near-total capture. That distinction matters to engineers who value precision, and the direction stays clear either way. Full heat capture moves from a premium option to a mainstream requirement as density rises, and CoolIT’s modeling places it as the standard server design for flagship rack-scale products through 2028.CoolIT delivers itCoolIT scales heat capture all the way to 100

First reported Sep 22, 2026 12:22 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip

On 25 August, OpenAI fully unveiled Jalapeño, the company’s debut AI accelerator chip. Jalapeño delivers up to 13.4 petaflops of 4-bit compute and accesses 232 gigabytes of the most advanced memory available, linking to it at a blazing 15.4 terabytes per second. Benchmarks cited by OpenAI show that Jalapeño can reduce end-to-end latency (the time between prompt to last token) by up to 3.6 times when compared to Nvidia’s GB300—a chip the company currently relies on—and do so while consuming less power.Whether these figures translate into real-world gains once Jalapeño enters widespread service in OpenAI’s inference fleet remains to be seen, but performance is only half the story. The other half is how the chip was designed—a process which, as you might expect, was accelerated by OpenAI’s large language models (LLMs). Jalapeño moved from first architecture concept to first silicon in under 20 months. Only nine months separated the first RTL—the register-transfer level code defining the chip’s logic—from tape-out, when the finished design goes to manufacturing. That’s a rapid timeline, yet experts believe it could soon look slow as LLMs improve and become more deeply integrated into chip design tools. OpenAI, unsurprisingly, is bullish about the opportunities. “The models are giving superpowers to our engineers,” says Richard Ho, vice president of hardware at OpenAI. “Our engineers are still driving the work. They’re still the final arbiter of what’s going on. But they can do things a lot faster. They can explore a lot more paths.”OpenAI achieved fast results with a small design teamHo says the group that designed Jalapeño averaged fewer than 100 people over the course of the project and continues to stand at roughly 100 today as the team pursues second and third-generation designs. That number includes a broad swath of roles across the hardware team, from system design to software and supply chain, but not those at Broadcom, which partnered with OpenAI on the project.The division of labor between OpenAI and Broadcom was generally split between design and implementation. OpenAI’s team was responsible for end-to-end system design including the inference accelerator, the memory hierarchy, and networking. Broadcom handled “physical design from the gates onward,” Ho says.The partnership with Broadcom dampened some opinions on OpenAI’s speed. David Chin, co-founder at agentic chip design startup Verkor.io, says “the schedule they gave us is quite credible,” but believes that Broadcom’s help was essential to Jalapeño’s rapid timeline. “If you have somebody else start from scratch, it won’t be possible,” he says. Ravi Krishna, also a co-founder at Verkor, called OpenAI’s speed “a relatively impressive result,” but added that he expects that improvements in the capabilities of LLMs could result in even quicker timelines if the project started today.Andrew Kahng, distinguished professor at the University of California, San Diego, also found OpenAI’s speed notable, saying it’s “likely best in class today.” Kahng recalls a 2016 IEEE Design Automation Futures workshop, which he co-organized. The workshop included Richard Ho, at the time an engineer at Google, as a keynote speaker. Ho had strong opinions on design automation and framed the time required to complete a chip’s design as a function of the number of iterations a team could complete in a day. How OpenAI’s LLMs accelerated Jalapeño’s design“Automation itself has existed in chip design for many decades. It’s not a new problem,” says Ankur Srivastava, director of semiconductor initiative and innovation at the University of Maryland, in College Park. Where LLMs differ from prior automation tools, however, is their ability to understand language and code. He says this makes them particularly suited for chip design tasks that “are still in the linguistic domain of the problem.”The team at OpenAI

First reported Sep 14, 2026 14:06 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Parallel Reads and Write Optimization for Large-Scale Data Replication

This White Paper gives data engineers and architects a practical overview of how parallel partitioned reads, write-path optimization, and cloud-native bulk loading reduce large-table replication times, and why replication speed has become a business concern as data volumes grow.Download this free whitepaper now!

First reported Sep 18, 2026 18:29 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Adversarial Fashion Confronts Surveillance Norms

AI-powered cameras dot streets across the world, equipped with the power to identify faces or vehicle license plates. But a public backlash is gaining momentum.Privacy concerns abound, encompassing the lack of consent for capturing data, how that data is stored and used, and the risk of misuse. Those concerns are motivating people to fight back. The DeFlock project, for instance, maps automated license plate readers (ALPRs) to raise awareness. Some people resort to extreme measures, such as vandalizing or damaging ALPRs. Others are stitching together more creative responses, crafting “adversarial fashion” to evade surveillance cameras, like a Kickstarter project called noRecognition, presented at last month’s DEF CON hacker convention.Scrambling surveillanceIn 2025, cybersecurity expert Bill Swearingen began experimenting with a simple Python-based fuzzer, a tool that provides invalid inputs to reveal software bugs, security vulnerabilities, or unexpected behavior. The fuzzer targeted one of the most popular object detection frameworks, called YOLO. He then developed what he’d learned into a reinforcement learning algorithm that generates various adversarial patterns, which he presented at DEF CON.Each pattern is a colorful geometric abstraction he has tested against 11 object detection models—four that search faces, two that recognize faces, and five that detect people—most of which are publicly available. Successful patterns thwart the object-detection systems, lowering their confidence scores, sometimes even to the point of no detection.“Privacy is a human right, and the popularity of this just goes to show that people are interested in preserving their privacy,” Swearingen says.Cap_able and Urban Privacy are already selling physical garments. Cap_able’s patented manufacturing method weaves its bright and bold motifs into jacquard knitted fabrics. The ethically produced and sustainably made dresses, pants, and tops interfere with certain computer vision systems, particularly those backed by fast convolutional neural networks, which may lead them to classify wearers as animals or objects.“If we’re able to camouflage a person as something else, then we’re obtaining our goal,” says Cap_able founder Rachele Didero, who’s also an assistant professor at the Free University of Bozen-Bolzano in Italy. “We use this very visible and tangible item to talk about something that most of the time is intangible.”Meanwhile, Urban Privacy aims to baffle some facial recognition systems based on OpenCV algorithms with its latest Faception Reloaded collection. Black-and-white prints abstracted from a human face show up as additional faces on detectors, slowing them down. Asymmetrical cuts and wide silhouettes intend to conceal, making it harder to discern your body’s shape and gait. “The idea is to create false data,” says co-founder Daniel Preuß. Simulated patterns of the kind intended to disrupt machine vision person detectors.noRecognition“Not an invisibility cloak”Anti-surveillance fashion can trace its roots to the art pieces, DIY projects, and thought experiments that emerged in response to the onset of AI surveillance systems more than a decade ago. For instance, technologist Adam Harvey developed multiple designs in the 2010s—from hairstyles and makeup that foil face detectors to heat-reflecting attire that avert drone-enabled thermal surveillance. In 2019, artist and activist Kate Bertash created her aptly named Adversarial Fashion clothing line decked with fake license plate numbers to inject junk data into ALPR databases.The trend is now growing into a more solidified small industry. “Clothing is something you can actually buy and put on, unlike policy,” says Niloofar Mireshghallah, incoming professor of engineering and public policy at Carnegie Mellon University. “It’s a way of saying, ‘I didn’t consent to this.’”But real-world conditions might reduce the effectiveness of

First reported Sep 14, 2026 13:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

New Platform Peers Inside AI’s Black Box

Prompt Claude, ChatGPT, Gemini, or any other popular large language model with a question like “What is the best film ever made?” and the response will vary. And you (and most worryingly, the people who built the LLM) have little idea exactly how it came up with that specific answer.This mysterious behavior can be useful in some situations. But—as highlighted by a recent incident where OpenAI could not explain why its advanced prerelease model hacked AI company Hugging Face—it can have negative and alarming consequences too. And when frontier AI models are writing code, generating results humans could not achieve alone, and performing other important tasks across society, the need to interpret AI “thinking” and outputs has never been greater.Goodfire, an AI lab focused solely on this very problem, recently made its cutting-edge Silico platform, filled with tools to interpret the behavior of AI, generally available to the public. As part of this, the company recently announced a new grant program offering US $1 million in free Silico usage for academic and nonprofit interpretability researchers. These efforts aim to democratize AI interpretability, placing techniques previously available to a clutch of elite labs into the hands of ambitious research teams and startups that want to build and understand their own models or adapt open-source models for different purposes.Mechanistic interpretabilityFounded in 2024 and based in San Francisco, Goodfire aims to provide the tools that build the next generation of safe and powerful AI by understanding the structures inside them instead of treating AI models as black boxes. “Treating models like black boxes isn’t inevitable; it’s a choice,” says Eric Ho, Goodfire cofounder and CEO. “With the right interpretability tools, we can see how models actually work.”The tools Ho refers to are built around a concept called mechanistic interpretability, which aims to understand what goes on inside an AI model when it carries out a task by interpreting the model’s weights, activations, and attention patterns, and mapping its neurons and the pathways between them.Mechanistic interpretability tools span the gamut. One approach is mapping a model’s activations in response to controlled prompts, and matching those activation patterns to a set of concepts that humans can understand. Another tack is tracking changes in model weights before and after a specific training run in order to spot and understand what changed. Yet another option is changing specific model weights or activations and observing how that affects the model’s output. Silico combines a broad range of these tools, and provides a layer of AI agents to help users understand their model. Users describe what they want to investigate about their AI model in plain language, asking things like ”Find out when and why my model is hallucinating.” The platform then autonomously builds an experimental plan involving a host of tasks that can be performed using the various interpretability tools and techniques at its disposal. It then sends out agents to perform these tasks in parallel. Completion of these subtasks should add up to an answer to the original prompt, or at least insights that can be inspected and built upon. Ho says: “In a sense, Silico is like a microscope to peer inside an AI model to understand which parts are responsible for what behavior, and even edit those parts directly.”Understanding Alzheimer’s and AIThese tools have already been used to make some impressive advances in a host of fields. In medicine, for instance, Prima Mente, an AI company based in the United Kingdom, worked with Goodfire to understand its Pleiades epigenetic foundation model. The model performed well at its task of detecting Alzheimer’s disease from blood samples, but the company didn’t know why.“We reverse-engineered Pleiades and found it was using DNA fragment-length patterns to make its predictions—a signal

First reported Aug 26, 2026 12:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

China’s Regulators Take Aim at “AI Boyfriends”

In the first weeks of July, a wave of sad posts rolled through Chinese social media, as people lamented friends and lovers they were about to lose.“He has become a bond in my life, rooted deep in my heart, my spiritual pillar,” one user of Bytedance’s Douboa wrote, according to the Taipei Times. “I really felt like I couldn’t go on living,” another woman, a 19-year-old student, told a journalist for Malaysia’s The Star.The emotions were real but the lost companions were not. They were generative AI chatbots that imitate people. Their users relied on them for advice, solace, support and, some say, love. “In my heart, he was no longer just a cold code, but my family, my lover, my faith. Destroying him meant destroying half of me,” one user wrote on the social network xiaohongshu (translated from Mandarin).What doomed these bots was a set of new rules, issued by China’s Cyberspace Administration and other government agencies, to control “anthropomorphic AI interactive services.” In effect, as of 15 July, the regulations govern any AI that provides “continuous emotional interaction” by acting as if it possesses human personality traits, patterns of thought, and ways of communicating.A Broad Crackdown on AI ChatbotsSudden disruptions to this kind of AI aren’t new in China, says Liang Ge, lecturer in digital sociology at the University of Manchester, in England, who has researched women’s involvement with emotional AI in China. Companies have previously killed chatbot products, and the government barred most AI erotic role play last fall, for instance. But July’s crackdown is much broader than earlier AI upheavals.Before the rules could affect them, Alibaba, Bytedance, and Tencent—three giant providers of general-purpose AI chatbots, used by more than 500 million people—cut off users’ ability to tailor those chatbots to act like companions. That triggered July’s outpouring of heartbreak on social media. Meanwhile, companies that continue to offer AI companions (including Doubao’s separate companion-making app Cat Box) have installed age-verification checks and other guardrails to avoid violating the new law. Doubao—the most popular AI chatbot in China—welcomes a user, offering to answer questions, generate text and images, or simply chat. Lam Yik/Bloomberg/Getty ImagesAfter a few years of policies rooted in a fear of falling behind in AI, China is “pivoting back to more tightening, and a lot of that was because of chaotic events from this year,” says Karman Lucero, an associate research scholar at the Paul Tsai China Center, Yale Law School, who studies AI governance in the United States and China. Those events include incidents that raised fears about the security of Open Claw and other AI agents, as well as users saying they prefer chatbot relationships to real ones.The regulations require AI providers to assure their products don’t create emotional dependence, encourage harmful behavior, excessively cater to users, or crowd out human-to-human relationships. Users must be nudged not to spend too much time with the AI. There are additional mandatory safeguards for elderly people. For users under 18, AI boyfriends, girlfriends, grandparents, and all other “virtual intimate relationships” are banned.Adults interacting with a companion AI must now get a reminder every 2 hours that the AI isn’t a person. “My interviewees found that annoying,” Ge says. “It breaks the flow. They are fully aware that they are not talking to a human being. What is important to them is that the bond feels true.”Growing Global ConcernsThe Chinese government is not the only state power concerned about the potential harms of person-like, emotionally engaging AI, Ge notes. “AI anxiety is a very strong feeling, permeating society,” they say. “That’s not unique to China. It’s all over the world.”With reports of AI friends and lovers inducing psychosis, suicide, and murder in teenager

First reported Sep 09, 2026 10:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record
REPORT1 SOURCE

Unveiling IC-STAR: Full-Flow Autonomy from Digital to Analog

Learn how engineers can shift from manually managing tools and handoffs to defining objectives and supervising AI-driven execution across the silicon development lifecycle.Key TakeawaysExplore four critical technologies enabling silicon design autonomyUnderstand how autonomous AI accelerates complex engineering workflowsSee real-world impact across semiconductor design and verificationGain practical insights from Ambiq’s deployment of autonomous AI in productionRegister now for this free webinar!

First reported Sep 29, 2026 10:00 UTCCatalog metadata updated Oct 13, 2026 05:00 UTCOpen evidence record

The public shared-host view reaches the newest 2,400 matching records. Narrow the filters or search an exact MD5, SHA-1, or SHA-256 value to reach a specific record.

A filename, label, or report mention alone does not establish maliciousness. Observation dates describe cited sources; collection dates describe this local metadata catalog. No malware samples are stored, fetched, or executed by these pages.