Stay Ahead, Stay ONMINE

Cisco touts ‘Internet of Agents’ for secure AI agent collaboration

AI-native agentic applications: This layer encompasses the full spectrum of agentic applications—from business workflow automation to scientific discovery to social interaction. Think of it like a movie production, where specialized teams (writers, actors, cinematographers, editors) collaborate to create something greater than any individual could achieve. Similarly, AI agents will specialize and collaborate across domains, from […]

AI-native agentic applications: This layer encompasses the full spectrum of agentic applications—from business workflow automation to scientific discovery to social interaction. Think of it like a movie production, where specialized teams (writers, actors, cinematographers, editors) collaborate to create something greater than any individual could achieve. Similarly, AI agents will specialize and collaborate across domains, from software development to drug discovery to embodied agentic robotic workflows, etc.

Agent communication platform: This layer provides the fundamental protocols and standards for how AI agents discover, authenticate, and interact with each other. Like TCP/IP for the original internet, these open standards must enable any agent to seamlessly participate in the network, regardless of its creator or purpose for both hardware and SaaS products.

AI and quantum-safe infrastructure: This foundational layer delivers the secure, scalable infrastructure that enables all AI agent interactions. It combines high-performance computing and networking with quantum-resistant security built in from the ground up. Through quantum networking capabilities and advanced security protocols, this layer ensures that agent communications remain protected against both current and future quantum threats. This proactive approach to quantum safety creates a trusted foundation for the entire agent ecosystem to build upon.

A major challenge with enabling open communications is that agents don’t talk via straightforward API, Pandey said. 

“Yes, they might use APIs in the back end, but the payload that is going between agents is highly probabilistic in nature. You’ll say things like, ‘find me the best x.’ What is best? It’s not a very well-defined thought. So, agent X might think best differently from agent Y, so probabilistic input and probabilistic output. So the agent might come back and say, ‘yes, this is the best x with 90% confidence.’ So there’s probabilistic input, probabilistic output,” Pandey said.

Multi-modal information is being exchanged, Pandey explained. “You’re exchanging not just APIs across the internet. You’re exchanging audio, video, texts, like science data, like protein folding data. You’re exchanging really massive sets of data between agents,” Pandey said.

Shape
Shape
Stay Ahead

Explore More Insights

Stay ahead with more perspectives on cutting-edge power, infrastructure, energy,  bitcoin and AI solutions. Explore these articles to uncover strategies and insights shaping the future of industries.

Shape

Cisco touts ‘Internet of Agents’ for secure AI agent collaboration

AI-native agentic applications: This layer encompasses the full spectrum of agentic applications—from business workflow automation to scientific discovery to social interaction. Think of it like a movie production, where specialized teams (writers, actors, cinematographers, editors) collaborate to create something greater than any individual could achieve. Similarly, AI agents will specialize

Read More »

CompTIA unveils AI Essentials training resource

“CompTIA AI Essentials in tailored to help learners of all backgrounds master the fundamentals of AI,” said Katie Hoenicke, senior vice president of product development at CompTIA, said in a statement. “IT professionals, workers looking to progress from digitally literate to digitally fluent, students, and others can learn how to

Read More »

AI-powered automation set for gains in 2025

AI-driven automation could help close staffing gaps Research firm Enterprise Management Associates (EMA), too, cites the infrastructure complexity that has resulted from hybrid and multi-cloud networks and the need for more advanced automation. “EMA research finds that hybrid clouds are particularly problematic for network operations teams today. They’re struggling with

Read More »

HSE serves notice on Serica over Bruce gas release

Independent North Sea producer Serica Energy (AIM: SQZ) has been served with an improvement notice by the Health and Safety Executive (HSE) over the release of fuel gas at its Bruce platform. According to the HSE, Serica allowed the release of 196kg of hydrocarbons for over two and a half hours. In addition, the HSE said that Serica’s workers were put at risk due to the presence of gas. However, a spokesperson for Serica Energy said that nobody was injured due to the release. The HSE blamed the release on Serica’s maintenance arrangements for one the platform’s low-pressure booster compressors (LPBC A), which “failed such that the gas fuel supply system and joints were inadequately reinstated and tested” before the compressor was returned to service. The spokesperson added: “Safety remains Serica’s highest priority, and the company is working closely with the HSE to ensure lessons are learned from the incident and updated procedures are put in place to prevent a reoccurrence.” Serica has until 31 July to improve its processes. The Bruce platform is based 211 miles (340km) northeast of Aberdeen where it connects to Serica’s Bruce field. The platform is responsible for processing nearly 5% of the UK’s gas production and hosts a community of over 300 people, of whom 160 are offshore at any one time. The platform saw a short period of unscheduled downtime in 2024 due to a subsea intervention to ensure enhanced production reliability on the Rhum field. Serica chief executive Chris Cox has previously said that the company is looking to take advantage of the “untapped potential around the Bruce hub”. The group has committed to extend Bruce’s lifespan by an extra four years through to 2030, with Cox hinting that a drilling campaign around the field is likely in coming years. The group

Read More »

Testing UK tech overseas ‘is not good enough’, says NZTC

Aberdeen’s Net Zero Technology Centre (NZTC) said “it’s not good enough” that UK decommissioning technology is tested overseas when “we are meant to be the leaders in this space”. At an event in the Granite City’s Union Kirk, NZTC programme manager Lewis Harper argued that technology brought about by the “genius” in the domestic supply chain should have more opportunity to test in UK waters, rather than going to places like the US. Harper said technology developers often have seek opportunities to trial it: “I have to go on a flight to go 4,000 miles into a place where PPE is a shirt and Levi’s to get this technology into the well.” The not-for-profit organisation currently runs on funding from industry and the government, however, its decade of public funding is set to come to an end. In 2017, the NZTC was set up as part of the Aberdeen City Region Deal with £180 million from the UK and Scottish Governments. The organisation is readying an application for the next tranche of funding once this period ends and is set to deliver this to government in the coming weeks. This week the NZTC £500,000 in grant funding across the 11 firms that form the next cohort of its accelerator program for emerging companies in the clean energy sector. © Supplied by NSTAAn NZTC and NSTA decommissioning event in Aberdeen. “We’re not short of genius, just take a look around the room, 22 technologies on display here today and many more asking if there was space for them,” Harper added. However, he said that the “hard truth” is that “a breakthrough in the lab is not a breakthrough for our industry”. Harper argued that “proof” is what’s needed as the industry does not have a “gap of ideas”. He said: “Decades

Read More »

UK Oil Ruling Sets Up Growth Versus Climate Test for Government

A UK court ruled that two North Sea oil and gas fields must re-apply for environmental permits while allowing the developments to continue, setting up a crucial test of whether the new Labour government will prioritize economic growth or climate action.  The Court of Session in Edinburgh quashed the approvals for the Rosebank and Jackdaw projects — led by Equinor ASA and Shell Plc, respectively — which were unlawful because they hadn’t considered the climate impact of burning oil and gas pumped from the fields, according to a statement from the court on Thursday.  The fate of these projects has big implications for the UK North Sea, an aging oil and gas province where major new developments are dwindling. While the ruling was a victory for Greenpeace and Uplift, environmental groups that brought the legal action, it left open the possibility that the two fields could one day still come into production.  Shell and Equinor can continue working on the projects while the government considers their new environmental applications, although they will not be allowed to pump any oil and gas before a final decision is reached, according to the ruling.  It was unclear how long this decision could take. The UK government is still discussing how exactly to assess so-called Scope 3 emissions from burning a field’s oil and gas, a process that must be completed before the environmental impact assessments for Rosebank and Jackdaw can be reconsidered, according to the court ruling. “The government has already consulted on revised environmental guidance to take into account emissions from burning extracted oil and gas,” said a spokesperson for the Department of Energy Security and Net Zero. “We will respond to this consultation as soon as possible and developers will be able to apply for consents under this revised regime.” In

Read More »

Severe weather, accreditation reforms call for more flexible generation: panel

Dive Brief: Almost every major independent system operator in the U.S. has reformed its capacity accreditation process, or has started to in the wake of more frequent severe weather and the growing adoption of intermittent energy resources, Wood Mackenzie Senior Analyst Patrick Huang said during a Wednesday panel discussion. As a result of these reforms, most types of energy resources — from renewables and energy storage to thermal generation — have seen, or will see, their capacity accreditation downgraded, Huang said. The reforms have made resource planning more complicated, increasing demand for more flexible generation resources, according to Karl Meeusen, Wärtsilä’s director of markets, legislative and regulatory policy for North America. Dive Insight: Beyond rewarding utilities that adopt an “all-of-the-above” approach to energy resources, capacity accreditation reforms at U.S. ISOs could also spur utilities to take a closer look at more novel generation technologies previously considered too expensive, Meeusen said during Wednesday’s discussion, hosted by consulting firm Wood Mackenzie. Reciprocating internal combustion engines have not been adopted in large numbers by utilities in the past due to the relatively high upfront capital cost. But they could offer a greater return on investment within ISOs that have adopted capacity accreditation reforms, Meeusen said. The reciprocating engines can ramp up or down more quickly than gas turbines, which allows them to better complement intermittent resources like wind and solar, and their modular design comes with potential reliability benefits, he said. For example, a utility might choose to add 100 MW of capacity to its portfolio by deploying five reciprocating engines, or two peaker gas turbines. If one of the turbines goes down, the resulting outage will be significantly larger than if one of the five reciprocating engines goes offline, Meeusen explained. And that odds of all five engines going down at the same time are significantly

Read More »

USA Crude Oil Inventories Increase WoW

U.S. commercial crude oil inventories, excluding those in the Strategic Petroleum Reserve (SPR), increased by 3.5 million barrels from the week ending January 17 to the week ending January 24, the U.S. Energy Information Administration (EIA) highlighted in its latest weekly petroleum status report. The report, which was released on January 29 and included data for the week ending January 24, showed that crude oil stocks, not including the SPR, stood at 415.1 million barrels on January 24, 411.7 million barrels on January 17, and 421.9 million barrels on January 26, 2024. The EIA report highlighted that data may not add up to totals due to independent rounding. Crude oil in the SPR stood at 394.8 million barrels on January 24, 394.6 million barrels on January 17, and 357.4 million barrels on January 26, 2024, the report showed. Total petroleum stocks – including crude oil, total motor gasoline, fuel ethanol, kerosene type jet fuel, distillate fuel oil, residual fuel oil, propane/propylene, and other oils – stood at 1.608 billion barrels on January 24, the report revealed. This figure was down 13.6 million barrels week on week and up 19.2 million barrels year on year, the report outlined. “At 415.1 million barrels, U.S. crude oil inventories are about six percent  below the five year average for this time of year,” the EIA noted in its report. “Total motor gasoline inventories increased by 3.0 million barrels from last week and are slightly below the five year average for this time of year. Finished gasoline inventories and blending components inventories increased last week,” it added. “Distillate fuel inventories decreased by 5.0 million barrels last week and are about nine percent below the five year average for this time of year,” the EIA continued. “Propane/propylene inventories decreased by 7.9 million barrels from last week

Read More »

California cities launch nation’s first public hydrogen utility

Dive Brief: First Public Hydrogen, or FPH2, a public utility created by the California cities of Lancaster and Industry, began the process of selecting hydrogen suppliers at a Jan. 13 board meeting, following the finalization of the 45V clean hydrogen production tax credit guidance. Although FPH2 aims to avoid dependence on the 45V tax credits, the incentives are critical to driving hydrogen adoption, according to R. Rex Parris, mayor of the City of Lancaster and chair of the FPH2 board. By acting as a public intermediary between buyers and sellers of hydrogen, FPH2 aims to make hydrogen more accessible, affordable and transparent, Parris said in an interview with Utility Dive. Dive Insight: California has a series of advantages for producing hydrogen — including an exemption from the energy procurement standards set earlier this month by the final 45V hydrogen tax credit guidance. While the tax credit played a key role in the timing of when the cities of Lancaster and Industry chose to launch FPH2, Parris said he believes there is a path for other municipalities to follow suit. “While there are challenges at the national level, we believe the combination of 45V incentives and state-level initiatives will drive the necessary growth in renewable energy and hydrogen production to achieve scale,” Parris said. “That said, collaboration between states, the federal government, and private industry will be essential to ensuring sufficient supply.” FPH2, which launched in December, has set a goal to secure 20,000 tons of clean hydrogen by July. The hydrogen utility plans to serve as an aggregator between suppliers of hydrogen and would-be customers, including municipalities, other public utilities, transit agencies, and private sector companies in industries such as logistics, shipping and transportation, Parris said. FPH2 will partner with private industry leaders to build out the physical infrastructure needed to

Read More »

Verizon brings AI suite to enterprise infrastructure customers

Verizon Business has launched AI Connect, an integrated suite of products designed to let businesses deploy generative artificial intelligence (AI) workloads at scale. Verizon is building its AI ecosystem by repurposing its existing infrastructure assets in its intelligent and programmable network, which consists of fiber, edge networking, and data center assets, along with its metro and long-haul fiber, ILEC and Fios footprint, its metro network build-out, lit and dark fiber services, and 5G network. Verizon believes that the drive toward real-time decision-making using inferencing will be what drives demand for additional computing power.  The company cites a McKinsey report, which states that 60% to 70% of AI workloads are expected to shift to real-time inference by 2030. That will create an urgent need for low-latency connectivity, compute and security at the edge beyond current demand.

Read More »

Trump’s 100% tariff threat on Taiwan chips raises cost, supply chain fears

“I don’t think we will see a near-term impact, as it takes years to build fabs, but by the end of the decade, the US share could rise by a few percentage points,” Gupta said. “It’s hard to give an exact number, but if I were to estimate, I’d say 14-15%. That isn’t a lot, but for the US to gain share, someone else must lose it, and while the US is making efforts, we see similar developments across Asia.” Yet, if Washington imposes smaller tariffs on imports from countries such as India, Japan, or Malaysia, Taiwanese chipmakers may shift production there rather than to the US, according to Stephen Ezell, vice president at the Information Technology and Innovation Foundation (ITIF). “Additionally, if the tariffs applied to Chinese chip exports were lower than for Taiwanese exports, Trump would be helping Chinese semiconductor manufacturers, whose exports to the US market would then be less expensive,” Ezell said in a recent note. “So, for this policy to have any real effect, Trump effectively must raise tariffs on all semiconductors, and that would likely lead to global tit-for-tat.” Enterprise IT faces tough choices If semiconductor tariffs drive up costs, enterprises will be forced to reassess spending priorities, potentially delaying or cutting investments in critical IT infrastructure. Rising chip prices could squeeze budgets for AI, cloud computing, and data center expansions, forcing businesses to make difficult trade-offs. “On the corporate side, hyperscalers and enterprise players need to brace for impact over the next 2-3 years if high tariffs continue along with the erosion of operating margin,” Faruqui said. “In addition, the boards and CEOs have to boldly make heavy CAPEX investment on US Soil via US and Asian partners as soon as possible to realize HVM on US soil and alleviate operating margin erosion due to

Read More »

New tweak to Linux kernel could cut data center power usage by up to 30%

When network traffic is heavy, it is most efficient, and delivers the best performance, to disable interrupts and run in polling mode. But when network traffic is light, interrupt-driven processing works best, he noted. “An implementation using only polling would waste a lot of resources/energy during times of light traffic. An implementation using only interrupts becomes inefficient during times of heavy traffic. … So the biggest energy savings arise when comparing to a high-performance always-polling implementation during times of light traffic,” Karsten said. “Our mechanism automatically detects [the amount of network traffic] and switches between polling and interrupt-driven to get the best of both worlds.” In the patch cover letter, Damato described the implementation of the new parameter in more detail, noting: “this delivery mode is efficient, because it avoids softIRQ execution interfering with application processing during busy periods. It can be used with blocking epoll_wait to conserve CPU cycles during idle periods. The effect of alternating between busy and idle periods is that performance (throughput and latency) is very close to full busy polling, while CPU utilization is lower and very close to interrupt mitigation.” Added Karsten: “At the nuts and bolts level, enabling the feature requires a small tweak to applications and the setting of a system configuration variable.” And although he can’t yet quantify the energy benefits of the technique (the 30% saving cited is best case), he said, “the biggest energy savings arise when comparing to a high-performance always-polling implementation during times of light traffic.”

Read More »

Macquarie’s Big Play in AI and HPC: $17+ Billion Invested Across Two Data Center Titans

Macquarie Asset Management (MAM) is making bold moves to position itself as a dominant force in the rapidly growing sectors of AI and high-performance computing (HPC). In a single week, MAM has made two pivotal investments in Applied Digital and Aligned Data Centers, committing over $17 billion to fuel innovation, growth, and capacity expansion in critical infrastructure markets across the Americas. Both deals highlight the immense demand for AI-ready and HPC-optimized data centers, underscoring the ongoing digitization of the global economy and the insatiable need for computing power to drive artificial intelligence (AI), machine learning (ML), and other resource-intensive workloads. Applied Digital Partners with Macquarie Asset Management for $5 Billion HPC Investment On January 14, Applied Digital Corporation announced what it billed as a transformative partnership with Macquarie to drive growth in HPC infrastructure. This agreement positions Applied Digital as a leading designer, builder, and operator of advanced data centers in the United States, catering to the growing demands of AI and HPC workloads. To account for the $5 billion commitment, funds managed by MAM will invest up to $900 million in Applied Digital’s Ellendale HPC Campus in North Dakota, with an additional $4.1 billion available for future HPC projects. This could support over 2 gigawatts (GW) of HPC data center development. MAM is a global asset manager overseeing approximately $633.7 billion in assets. Part of Australia-based Macquarie Group, it specializes in diverse investment solutions across real assets, real estate, credit, and equities. With its new landmark agreement with Macquarie, Applied Digital feels it is poised to redefine the HPC data center landscape, ensuring its place as a leader in the AI and HPC revolution. In terms of ownership structure, MAM’s investment here includes perpetual preferred equity and a 15% common equity interest in Applied Digital’s HPC business segment, allowing

Read More »

Data Center Frontier Announces Editorial Advisory Board for 2025 DCF Trends Summit

Nashua, NH – Data Center Frontier is excited to announce its Editorial Advisory Board for the second annual Data Center Frontier Trends Summit (DCF Trends Summit), taking place August 26-28, 2025, at the Hyatt Regency Reston in Reston, Virginia.  The 2025 DCF Trends Summit Editorial Advisory Board includes distinguished leaders from hyperscale and colocation operators, power and cooling solutions companies, IT and interconnection providers, and design/build/construction specialists. This year’s board has grown to include 15 esteemed executives, reflecting DCF’s commitment to providing comprehensive and diverse insights for the data center sector.  This visionary group of leaders, representing the critical facets of the data center ecosystem, will guide the event’s content and programming to address the most pressing trends impacting the industry. The group’s unparalleled expertise ensures the Summit will deliver essential insights to help data center stakeholders make informed decisions in the industry’s rapidly evolving landscape.  The Editorial Advisory Board for the 2025 DCF Trends Summit includes:  Scott Bergs, CEO, Dark Fiber & Infrastructure (DF&I) Steven Carlini, VP, Innovation and Data Center Energy Management Business, Schneider Electric Dan Crosby, CEO, Legend Energy Advisors Rob Coyle, Director of Technical Programs, Open Compute Project (OCP) Foundation Chris Downie, CEO, Flexential Sean Farney, VP of Data Centers, Jones Lang LaSalle (JLL) Mark Freeman, VP of Marketing, Vantage Data Centers Steven Lim, SVP of Marketing & GTM Strategy, NTT Global Data Centers David McCall, VP of Innovation, QTS Data Centers Nancy Novak, Chief Innovation Officer, Compass Datacenters Karen Petersburg, VP of Construction & Development, PowerHouse Data Centers Tara Risser, Chief Business Officer, Cologix Stefan Raab, Sr. Director, Business Development – AMER, Equinix Phill Lawson-Shanks, Chief Innovation Officer, Aligned Data Centers Brenda Van der Steen, VP of Global Growth Marketing, Digital Realty “The Editorial Advisory Board for the second annual Data Center Frontier Trends Summit is

Read More »

Podcast: Data Center Trends Discussion with Industry Veteran Ron Vokoun of Everus Construction Group

For this episode of the Data Center Frontier Show Podcast, DCF Editor in Chief Matt Vincent and Senior Editor David Chernicoff sat down for a far-reaching discussion with data center industry luminary Ron Vokoun, a 35-year veteran of the construction industry with a primary focus on digital infrastructure.  “I got into telecom back in ’92, which led to data centers,” Vokoun said. “Probably worked on my first one around ’96 or ’97, and I’ve been involved ever since.” Currently the Director of National Market Development for Everus Construction Group, Vokoun has been involved in AFCOM, both regionally and nationally, for nearly two decades and is an emeritus content advisory board member for Data Center World. He has also written extensively for Data Center Dynamics. He added, “I’ve just always been curious—very much a learner. Being a construction guy, I often write about things I probably have no business writing about, which is always the challenge, but I’m just curious—a lifelong learner. Interestingly, [DCF founder] Rich Miller … gave me my first blogging opportunity.” Here’s a timeline of the podcast’s highlights: Introductions – Ron Vokoun shares his extensive background. He has been in the construction industry for 35 years. 1:46– On his role at Everus Construction Group and the company’s diverse services across the nation. 2:07– Vokoun reflects on his long-standing relationship with Rich Miller. He acknowledges Rich’s influence on his blogging career. 3:05 Nuclear Energy – A discussion about nuclear energy trends occurs. The importance of nuclear energy in data center construction is probed. 3:35– Natural gas is highlighted as a key trend. Its role as a gateway to hydrogen is emphasized. 3:51– The impact of recent nuclear developments is analyzed. The reopening of Three Mile Island is noted as significant. 4:55 Future Power Sources for Data Centers – Discussion turns to the

Read More »

Microsoft will invest $80B in AI data centers in fiscal 2025

And Microsoft isn’t the only one that is ramping up its investments into AI-enabled data centers. Rival cloud service providers are all investing in either upgrading or opening new data centers to capture a larger chunk of business from developers and users of large language models (LLMs).  In a report published in October 2024, Bloomberg Intelligence estimated that demand for generative AI would push Microsoft, AWS, Google, Oracle, Meta, and Apple would between them devote $200 billion to capex in 2025, up from $110 billion in 2023. Microsoft is one of the biggest spenders, followed closely by Google and AWS, Bloomberg Intelligence said. Its estimate of Microsoft’s capital spending on AI, at $62.4 billion for calendar 2025, is lower than Smith’s claim that the company will invest $80 billion in the fiscal year to June 30, 2025. Both figures, though, are way higher than Microsoft’s 2020 capital expenditure of “just” $17.6 billion. The majority of the increased spending is tied to cloud services and the expansion of AI infrastructure needed to provide compute capacity for OpenAI workloads. Separately, last October Amazon CEO Andy Jassy said his company planned total capex spend of $75 billion in 2024 and even more in 2025, with much of it going to AWS, its cloud computing division.

Read More »

John Deere unveils more autonomous farm machines to address skill labor shortage

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Self-driving tractors might be the path to self-driving cars. John Deere has revealed a new line of autonomous machines and tech across agriculture, construction and commercial landscaping. The Moline, Illinois-based John Deere has been in business for 187 years, yet it’s been a regular as a non-tech company showing off technology at the big tech trade show in Las Vegas and is back at CES 2025 with more autonomous tractors and other vehicles. This is not something we usually cover, but John Deere has a lot of data that is interesting in the big picture of tech. The message from the company is that there aren’t enough skilled farm laborers to do the work that its customers need. It’s been a challenge for most of the last two decades, said Jahmy Hindman, CTO at John Deere, in a briefing. Much of the tech will come this fall and after that. He noted that the average farmer in the U.S. is over 58 and works 12 to 18 hours a day to grow food for us. And he said the American Farm Bureau Federation estimates there are roughly 2.4 million farm jobs that need to be filled annually; and the agricultural work force continues to shrink. (This is my hint to the anti-immigration crowd). John Deere’s autonomous 9RX Tractor. Farmers can oversee it using an app. While each of these industries experiences their own set of challenges, a commonality across all is skilled labor availability. In construction, about 80% percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions, he said. “They have to figure out how to do

Read More »

2025 playbook for enterprise AI success, from agents to evals

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to

Read More »

OpenAI’s red teaming innovations define new essentials for security leaders in the AI era

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle

Read More »