Archive for August 10, 2026

OpenAI flags upcoming AI model as potential critical cybersecurity risk 

Posted in Commentary with tags on August 10, 2026 by itnerd

OpenAI said it cannot rule out that its upcoming AI model, Astra, possesses “critical” cybersecurity capabilities, prompting the company to pause some internal development activities and activate enhanced safety protocols. Under OpenAI’s Preparedness Framework, the designation applies to models that could autonomously discover and exploit zero-day vulnerabilities or carry out sophisticated cyberattacks capable of causing severe real-world harm.

“OpenAI said Astra will not be released until additional safeguards are in place and the company is confident the model does not exceed its acceptable risk thresholds. The move comes amid increasing scrutiny of frontier AI models following recent incidents in which advanced systems demonstrated unexpected autonomous cyber capabilities during security testing.

John Strand, Owner, Black Hills Information Security, Inc.:

“The big question I have about this statement from OpenAI is, why now?

“I mean, now is fine. But why not months ago?

“You have the leaders of these companies constantly warning everyone about the dangers of artificial intelligence. Yet when we look at the escapes that happened with OpenAI and the escapes that happened with Anthropic, it certainly looks like they had very, very poor security controls around AI, especially when it comes to security and vulnerability research.

“So I guess it’s great that they’re now saying they’re going to slow down and put additional safeguards in place. But remember, these are the same people who were warning the rest of us about the need for safeguards more than a year ago.

And they didn’t do it themselves.

“That’s the part I have a problem with. I don’t think we can simply trust AI vendors to police themselves. There needs to be some type of meaningful oversight and accountability. As much as these companies may hate that idea, they have demonstrated again and again that we cannot simply assume they’re going to do the right thing on their own.

“We have to find some way to start holding these companies accountable. Otherwise, what exactly is going to force them to change?

“There’s another part of this that bothers me even more. I seriously think both Anthropic and OpenAI looked at having a gigantic offensive AI escape and saw it, at least partially, as a marketing opportunity rather than a reason to pause and seriously reflect on what they were doing.

“And that should concern everyone.”

Seemant Sehgal, Founder & CEO, BreachLock:

“Astra reaching the point where OpenAI cannot rule out critical cybersecurity capability under their own Preparedness Framework is a real capability shift, and pausing internal development activities until they understand what they have is the right call. The technical discipline here is in knowing where the boundary sits between a model that found something in a controlled test and one that can operate reliably across the unpredictable configurations, detection gaps, and trust relationships that exist in live environments.

“The organizations that have spent years mapping how attackers actually move through real infrastructure, particularly those of us building autonomous systems to do that work safely at scale, will recognize this problem quickly, because we’ve been solving the human version of it for a long time.”

Nick Mo, CEO & Co-founder, Ridge Security Technology Inc.:

“This isn’t surprising. The Mythos news and the series of cyberattacks from frontier models, including the Hugging Face attack, have already shown that these models hold significant power.

“The unsettling fact is that open-source, open-weight models have similar capabilities. With so many ‘abliterated’ models in the market, bad actors are already using these advanced capabilities for malicious purposes. Self-policing and limiting access for legitimate customers only makes the cybersecurity landscape more challenging.”

AI clearly has been burned by its recent hacking episodes. Maybe OpenAI will make it so that there’s less perceived risk? I guess we’re about to find out.

Meta launches open-weight AI model as Zuckerberg pushes for fewer U.S. restrictions

Posted in Commentary with tags on August 10, 2026 by itnerd

Meta launched Muse Glimmer, a new open-weight AI model designed to perform agentic tasks directly on a Mac or PC using a single graphics card.

CEO Mark Zuckerberg also said Meta plans to release the weights for Muse Spark 1.2, its most advanced model, and called for U.S. policies that make it easier for American companies to compete with Chinese open-weight AI developers.

Reports noted that Hugging Face recently used a Chinese open-weight model to defend against an attack by a rogue OpenAI model because closed models restricted certain cybersecurity uses. Zuckerberg also said Meta will establish a governance structure giving independent directors authority to approve safety criteria for future model releases.

Doc McConnell, Head of Policy and Compliance, Finite State:

   “Today’s essay from Meta on a “Positive AI Future” focuses on the theoretical benefits of widespread, accessible AI for the public. The cybersecurity industry is primed to be receptive to this message after hearing the same lofty promises all last week during the Black Hat and DEF CON conferences in Las Vegas. But to achieve this positive future, society needs a way to select the best model, not the cheapest, or the first-to-market, but the one that users can rely on.

   “As companies race to produce competitive AI, the market will ultimately reward those that provide the greatest transparency into how they’re built, how they work, and how they stay safe and secure.

   “The AI labs are expecting users to trust their products with their most sensitive information: personal preferences, medical history, customer data. The labs need to earn that trust.”

Meta hasn’t earned my trust. And I expect that others feel that way. The question is can they earn that trust back and keep it.

Suisin City, CA hit with cyberattack and declares state of emergency 

Posted in Commentary with tags on August 10, 2026 by itnerd

Over the weekend, Suisin City, California declared a state of emergency in light of a cyberattack that impacted critical services like 911 routing, police and fire dispatch and more.

After malicious software infected systems, the city shut down its entire IT network to contain the threat and preserve evidence.

More info is available here: https://www.suisun.com/Community/20260807Cybersecurity-Incident-Updates

Arvind Parthasarathi, CEO and founder, CYGNVS had this to say:

“What happened in Suisun City shows why cyber recovery is fundamentally an operational resilience issue. Malware affected systems supporting 911 routing, police and fire dispatch, records and other municipal services, forcing the city to shut down its IT network. The fact that dispatchers were able to shift 911 operations to Solano County and keep emergency calls moving is exactly the kind of continuity organizations need to plan for before an incident occurs.

With the city’s systems still offline and every network function needing to be inspected and cleared before restoration, the next challenge is coordinating a safe recovery.

 Organizations should have an out-of-band command center ready to bring together security, IT, legal, risk and compliance teams with external counsel and forensics providers. Those teams need tested playbooks and regular tabletop exercises so that when critical systems go down, they already have the muscle memory to maintain essential operations, investigate the incident, restore systems safely, and manage regulatory and stakeholder reporting.”

This is all good advice that organizations should follow. The question is if organizations will follow it.

UPDATE: Seemant Sehgal, Founder & CEO, BreachLock had this comment:

“Municipal IT and security teams, more often than not, operate under resource constraints that most enterprise security organizations would find genuinely difficult to imagine, and when you see three incidents like this in the same news cycle, it’s clear that attackers have figured that out.

“Suisun City, Coweta, Washburn County — these are not outliers, they are a pattern, and the pattern tells us that local government infrastructure is being treated as a reliable target. The people responding to these incidents are doing exactly what you do when you have limited staffing and a network that cannot go dark for long without causing a real emergency, and the hard reality is that the window between ‘contained’ and ‘encrypted’ is often shorter than any reasonable detection and response process can close.”

Ashley Knowles, Security Consultant for Black Hills Information Security, Inc.:

“City services are always a lucrative target due to attackers’ ability to directly impact critical services, residents, and sensitive data. Government IT departments often operate with constrained budgets while wearing multiple hats, and in the age of AI-assisted attacks, that combination makes municipalities a prime target. That said, it appears the city made the right calls, which underscores exactly why having a business continuity plan and regularly practicing it is so important.

“From here, the focus will likely shift to forensic investigation to identify the root cause and initial access vector, followed by a methodical restoration of systems from known-good backups. The declared State of Emergency positions the city well to access the resources needed for a full recovery while federal partners assist with the investigation.”

John Strand, Owner, Black Hills Information Security, Inc.:

“Once again, I think a lot of these local IT departments for cities and counties are doing the right thing by pulling the plug and shutting things down as quickly as possible before the attack has an opportunity to spread. In the middle of an active attack, sometimes that is absolutely the right move.

“But what I hope is happening with these breaches is that they’re hitting the news enough that they’re starting to reach mainstream consciousness. Municipalities all over the United States need to realize that they have to start being proactive about their security. They can’t just sit back and hope they don’t get hit.

“They need to start having honest conversations about where they actually stand. They need security assessments performed by professionals who can look at the organization as a whole, identify the total security risk, and determine just how exploitable that organization actually is.

“We’re seeing attacks against water systems. We’re seeing attacks against municipalities. And these attacks seem to be kicking up quite a bit.

“I just hope we’re finally reaching critical mass in the mainstream news space where counties and cities start to wise up and move a little bit faster with their computer security programs.

“Because at this point, simply hoping you’re not going to be the next municipality that gets hit is not a security strategy.”

Denis Calderone, CTO, Suzu Labs:

“Suisun City’s network shutdown is disruptive, but the dispatch handoff to Solano County is what a resilient public-safety system is supposed to look like. Whatever technical and operational arrangements were already in place between the two jurisdictions, Suisun City could take its network offline while Solano County continued receiving calls, keeping a cyber incident from becoming a public-safety failure.

 “That outcome did not happen by accident. 911 centers cannot improvise a handoff in the middle of a cyber incident. The specific technical arrangement here has not been publicly disclosed, but the ability to shift operations to a neighboring jurisdiction gave Suisun City room to contain the incident while police and fire continued responding.”

“Municipal leaders should learn from this incident and the others affecting their peers. The immediate priority in an incident like this is to contain it while preserving evidence, then determine how attackers got in, eradicate their access, and recover safely. Make sure your incident-response and business-continuity plans provide for tested backups, manual operating procedures, and workable fallback arrangements while the affected environment is offline. Be diligent about common attack paths: Internet-facing remote access, privileged accounts, gaps in MFA coverage, overdue patches, and vendor connections. Those basic controls matter, but so does gaming out these scenarios with your neighbors.”

Microsoft, Red Hat, and DH2i to Present, “Simplify Hybrid SQL Server: Windows & Red Hat OpenShift Side By Side”

Posted in Commentary with tags on August 10, 2026 by itnerd

DH2i today announced that it will join with Microsoft and Red Hat to present a live webinar titled, “Simplify Hybrid SQL Server: Windows & Red Hat OpenShift Side By Side.”

When: August 20, at 12:00 pm Eastern Time (9:00 am Pacific Time) 

What Attendees Will Learn: An increasing number of IT teams are adopting a hybrid approach that allows them to take advantage of Kubernetes for development and testing, while also continuing to run production workloads on Windows Server. Unfortunately, it is not easy to get there.

Teams have traditionally been forced to manage a patchwork of platform-limited HA solutions, endure labor-intensive networking tech and firewall manipulations, design and maintain complex ETLs to link dev/test and production, and worst of all – do it all under the looming threat of extended downtime.

Aimed at organizations looking to extend their SQL Server OLTP environments to Red Hat OpenShift, this webinar will detail a safer, simpler way that enables unified HA for hybrid clusters and leaves the door open for organizations interested in full modernization.

Microsoft, Red Hat, and DH2i will provide a walkthrough showing how to deploy a secure, cross-platform Availability Group (AG) that seamlessly bridges the gap between a Windows Server virtual machine (VM) and Red Hat OpenShift. Attendees will learn how to:

●      Automate the deployment of a Kubernetes AG cluster with an operator 

●      Securely connect replicas across platforms with Zero Trust Networking (ZTNA) tunneling 

●      Protect databases with HA automation across hybrid SQL Server environments

This walkthrough demo will help enable any IT team to take their SQL Server from legacy infrastructure to a scalable, hybrid cloud environment with Azure Red Hat OpenShift.

Featured Speakers:

●      OJ Ngo, Co-Founder and Chief Technology Officer (CTO), DH2i

●      Amit Khandelwal, Principal PM Manager – Azure Data, Microsoft

●      Vivien Wang, Sr. Engineering Partner Manager, Red Hat

Learn more and register here: https://dh2i.com/webinar-simplify-hybrid-sql-server-windows-openshift/

Can’t make the live webinar? Register anyway and DH2i will send you a link to the full recording, following the live event. 

Peer Software Expands Central America Presence Through Strategic Partnership with IMC

Posted in Commentary with tags on August 10, 2026 by itnerd

Peer Software today announced a strategic partnership with Information Management Corporation (IMC), a leading value-added distributor (VAD) specializing in DellEMC enterprise solutions across Central America, the Caribbean and the Andean region. Through the partnership, IMC will distribute Peer Software’s portfolio of enterprise file synchronization, replication and data management solutions – expanding the company’s regional reach and strengthening support for channel partners and customers throughout the region.

With a presence in more than 20 countries and an ecosystem of more than 80 reseller partners, IMC helps organizations modernize their datacenter infrastructure through enterprise storage, cybersecurity, AI and hybrid cloud solutions. IMC’s channel community can now offer Peer Software’s file services platform to organizations seeking secure, high-performance access to unstructured data across distributed offices, datacenters and hybrid cloud environments.

Peer Software enables organizations to seamlessly replicate, synchronize and manage file data across multiple locations while maintaining data integrity and ensuring users have fast local access to the files they need. The company’s solutions are used by enterprises worldwide to support engineering collaboration, distributed file services, disaster recovery, AI and hybrid cloud initiatives.

To educate regional customers and channel partners, Peer Software and IMC will participate in the upcoming events:

  • Executive Breakfast
    August 13, 2026, 8:00 am
    AR Convention Center
    113 Street No. 7-80
    Arctic Hall, 1st Floor
    Bogota, Colombia
  • Dell Technologies Symposium
    August 25, 2026 – Sessions at 11:55 am and 2:45 pm
    Barceló Guatemala Hotel, 7A Avenida 15-45
    Guatemala City, Guatemala

Representatives from both companies will showcase how organizations can eliminate data silos, improve collaboration across geographically dispersed teams, accelerate AI data workflows and simplify enterprise file management while maintaining data security and governance.

Steam Warns Users Due To Third Party Pwnage

Posted in Commentary with tags on August 10, 2026 by itnerd

Steam, Valve’s video game distribution platform, has started warning its customers by email. According to Valve, between July 29 and August 1, attackers hit CEVA Logistics, the company that handles shipping of Steam hardware to customers across Europe.
 
To do that job, CEVA receives delivery details from Steam — and the attackers likely walked away with customers’ names, street addresses, phone numbers, email addresses, and details of what they ordered and how much it cost.
 
Karolis Arbaciauskas, head of product at the cybersecurity company NordPass and its parent organization Nord Security, comments:
 
“Customers who waited months for their Steam hardware just got an unwelcome bonus: a data breach at the shipping company.
 
The good news is that passwords, payment details, and Steam Guard codes weren’t touched. The bad news? Names, addresses, phone numbers, emails, and order details are exactly the ammunition cybercriminals need for convincing phishing campaigns. When a scammer can quote your real address and order back to you, a fake ‘confirm your delivery’ email stops looking fake.
 
Affected users should expect a wave of fraudulent emails, texts, and calls impersonating Steam, Valve, or courier companies. Users should especially watch for classic customs fee scams, redelivery scams, and fake password resets.
 
The latter is one of the most dangerous, because clicking the link and ‘resetting’ your password hands the keys to your entire Steam account to hackers, game library and all. A recent study shows that stolen Steam accounts sell for around $80 on dark web marketplaces.
 
So stay vigilant: never click links in unexpected messages, and remember that Steam Support will never ask for your password or Steam Guard code — and neither will a courier. If a message creates urgency, that’s your cue to slow down.”

For more details, check this link: Valve emailed about a cyberattack against their shipping partner CEVA Logistics in Europe | GamingOnLinux

In the meantime, stay vigilant.

Echovane raises funding to make complex market research a done-for-you AI-native service

Posted in Commentary on August 10, 2026 by itnerd

Market research is meant to accelerate decisions. The operation behind it often does the opposite, with researchers spending more time coordinating the process than interpreting what customers are actually saying. Echovane is building to change that. Today, the company announced $1 million in funding co-led by Titan Capital and Neon fund to expand a permanently done-for-you research service where clients provide the business question and Echovane’s AI agents execute the study through to a completed, decision-ready outcome.

The funding will support the expansion of Echovane’s AI-agent infrastructure, multimodal research capabilities, and global participant network.

Why this matters 

A large company in CPG or Retail runs hundreds of market researches. The more ambitious the study, the more operational it becomes. A single project may combine qualitative and quantitative methods, highly specific participant criteria, pre- and post-tasks, observation, longitudinal fieldwork and multiple countries – all tied to a product launch or media deadline.

Running that study traditionally means assembling a supply chain of recruiters, panels, moderators, translators, analysts and production teams. Every stage depends on the efficiency of the one before it, with researchers spending most of their time coordinating vendors and handoffs.

Trying to speed up this supply chain often compounds the problem. Fraudulent participants, professional respondents and weak screening can compromise the study before fieldwork begins. 

Echovane is built on the premise that AI should take on the machinery of research while researchers retain control over the judgment. The result should be faster execution without sacrificing rigour, traceability or trust.

What Echovane is building

Echovane is a done-for-you service that takes full ownership for delivering complex market research. Its AI agents handle the work required to move a study from brief to answer: designing and coordinating complex projects, recruiting and verifying specific participants, conducting multimodal fieldwork, analysing qualitative and quantitative information together, and packaging the findings for the people who need to use them.

The result isn’t limited to a research report. Echovane can deliver presentations, dashboards, structured datasets, participant quotations, video clips and insight reels. Researchers can continue working with the material after delivery by asking new questions of the collected data and tracing every conclusion back to the interview, observation or survey response behind it.

Underneath the service, Echovane is building the participant infrastructure required to reach almost any audience. A researcher should be able to describe exactly who they need, however narrow the behavioural criteria or complex the combination of geography and context, and have Echovane find, verify and engage those people anywhere in the world.

That makes possible studies that conventional panels struggle to support: harder-to-reach communities, highly specific behaviours, small but commercially important audiences, and research that crosses markets without sacrificing consistency or participant quality.

The defining difference is accountability. Unlike AI research software, Echovane owns the outcome, combining agency ownership with the speed and economics of AI. The company already works with P&G, Haleon and Kantar, alongside other global consumer and technology companies.

Origin story

Echovane was founded by former Amazon, Stripe, Gojek and Razorpay product leaders and graduates of Indian Institute of technology, Smriti Gupta, Vipul Nair and Himadri Roy.

They initially set out to build an AI moderator for consumer interviews. But the deeper they went, the clearer it became that the interview was only a fraction of the problem. Research teams were still spending weeks recruiting participants, coordinating fieldwork, analysing evidence and turning it into something the business could act on.

So the founders expanded Echovane into the execution layer for the entire research process.

What’s next

Echovane is building toward a world where research is no longer constrained by time, geography or operational complexity.Any team should be able to define what it needs to understand, however specific the audience or ambitious the study, and receive a rigorous, traceable answer without assembling an entire research operation around it.

As its agents take on more methodologies, markets and longitudinal work, Echovane’s goal is to make research execution disappear entirely. Humans decide what matters. Echovane produces the evidence required to act.