Archive for January 12, 2026

Instagram responds to allegations that it got pwned

Posted in Commentary with tags on January 12, 2026 by itnerd

Last Friday, security software vendor Malwarebytes claimed “Cybercriminals stole the sensitive information of 17.5 million Instagram accounts,” while Instagram has publicly denied that it suffered a data breach, even though a large dataset allegedly containing information from around 17 million accounts was being discussed online. 

This alleged Instagram data was released on numerous hacking forums for free, with the post claiming it was gathered through an unconfirmed 2024 Instagram API leak.

The dataset contains the following counts of unique values:

  • ID:17,015,503
  • Username: 16,553,662
  • Email:6,233,162
  • Phone number:3,494,383
  • Name: 12,418,006
  • Address: 1,335,727

The social media company acknowledged a technical issue that allowed an external party to trigger mass password reset emails for many users but emphasized that its systems were not breached and user accounts remain secure. Instagram says it has already fixed the bug that enabled the unauthorized reset requests. 

Instagram has urged users to ignore unsolicited password reset emails unless they personally initiated the request and to take standard security precautions such as enabling two-factor authentication. 

Steven Swift, Managing Director, Suzu Labs had this to say:

   “There are two separate issues with the Instagram incident. One being that it was possible to initiate a password reset for other users (this one is reported as fixed) and separately, someone aggregated what appears to be old breach data into a new package. Neither of these are huge issues, though it will certainly make some users concerned.

   “It’s going to be concerning for users to see someone else attempting a password reset. Note that this issue was limited to initiating a password reset. There’s no indication that attackers were able to actually complete a password change. Making this more of an annoyance rather than a major security threat.

   “For the data leak itself, this is old data. The only thing new here is that someone aggregated a bunch of leak data together and is now bragging about it. One of the unfortunate realities about using services on the internet, is that personal data tends to leak out of most services, eventually.

   “Once the data is leaked, there’s no way to put it back. If it’s out, it’s out.

   “So, what can users do about it? For this incident, not much. It doesn’t appear passwords were exposed, and the leak data was old. However, some general recommendations still apply.

   “If you’re ever concerned after seeing suspicious activity on your account, any account, reset your password and double check that you have MFA in place. It’s generally better to be a bit cautious here. Use a password that you don’t use anywhere else. Ensure that its sufficiently long and/or complex. Save your passwords in a password manager.”

Michael Bell, Founder & CEO, Suzu Labs follows with this:

   “Two separate issues hit at once. The dataset appears to be from a 2024 scraping or API exposure, while the password reset bug is a separate technical issue. No passwords in the leak sounds reassuring, but it doesn’t take much to fill that gap. Those 6 million email addresses can be cross-referenced against infostealer logs and existing credential dumps to find matching passwords. Most people reuse credentials somewhere along the line. Instagram users should enable MFA and make sure they didn’t use the same password a bunch of other places.”

John Carberry, Solution Sleuth, Xcape, Inc. adds this:

   “The recent disclosure of 17.5 million Instagram user records highlights the ongoing tension between how companies define a “breach” and the actual risks faced by users. While Meta insists its central systems weren’t hacked, the appearance of this data on BreachForums demonstrates how a 2024 API “scraping” vulnerability can be just as harmful as a direct attack.

   “This incident underscores the blurring lines between a confirmed breach and large-scale data exposure, both of which erode user trust. Even if Instagram’s main systems weren’t breached, a vulnerability allowing mass password reset abuse can still lead to account takeovers and widespread social engineering.

   “The presence of millions of email addresses and phone numbers in these datasets raises serious concerns about data aggregation from previous leaks, scraping activities, or API misuse.

   “From a user’s perspective, the technical difference between a system breach and a massive API scrape is meaningless when their inbox is flooded with convincing reset links. Transparency regarding data origin is crucial, especially when free data releases facilitate abuse. This situation also emphasizes how reset mechanisms can be exploited if not carefully rate-limited and monitored.

   “When platforms downplay failures, attackers fill the gap, and users pay the price.”

The cynic in me says that Meta who owns Instagram doesn’t want to admit that they got pwned in some way. I guess we’ll have to see when people start getting pwned in various ways to prove or disprove if this is factual.

CRN Recognizes Hammerspace for AI Training and Inferencing Performance on 2026 Cloud 100 List

Posted in Commentary with tags on January 12, 2026 by itnerd

Hammerspace today announced it has been named to the 2026 CRN® Cloud 100 list by CRN®, a brand of The Channel Company. The annual list includes the most innovative channel-focused cloud technology companies transforming how enterprises deploy and scale cloud infrastructure. 

Hammerspace was recognized as it shapes the market for how organizations run AI training and inference with the cloud. Its data platform delivers Tier 0 storage performance to speed AI results, then automatically transitions data to cost-efficient object storage once demand subsides.

Hammerspace software was purpose-built to operate across on-premises, cloud and hybrid environments, allowing enterprises to move data to compute wherever GPUs are available.

This architecture makes Hammerspace ideal for organizations that need to:

  • Maximize GPU efficiency during AI training or inferencing
  • Avoid permanent costs for large high-performance cloud storage pools
  • Maintain open, standards-based architectures

Tier 0: Maximum Performance, Without Permanent Cost

Hammerspace’s Tier 0 delivers direct, NVMe-class storage performance to GPUs, eliminating the I/O bottlenecks that commonly stall GPU pipelines in the cloud. Unlike traditional cloud storage models that force customers to pay premium prices for ongoing high-performance storage, Hammerspace enables a dynamic, workload-aware approach. This allows organizations to:
 

  • Run AI workloads in the cloud on GPU clusters with Tier 0 data storage performance
  • Sustain full GPU utilization with parallel, high-throughput data storage access
  • Automatically orchestrate data movement of job outputs to object storage
  • Improve cloud economics without sacrificing performance

The result is faster AI, higher GPU cluster efficiency, and dramatically lower cloud storage costs.

How Hammerspace’s Data Platform Works

Image: The Hammerspace Data Platform provides a single, unified namespace that spans existing on-premises storage and cloud resources, giving users and applications a single, secure way to see and access data across storage types, clouds and multiple sites.
 

1. AI Workload Starts – Tier 0 Becomes Engaged: Data is delivered directly to cloud GPUs using Tier 0 NVMe-class performance, eliminating I/O bottlenecks and keeping GPUs fully utilized.

2. Workload Completes, Hammerspace Orchestrates Data Movement: Hammerspace’s Data Platform automatically transitions outputs to object storage, where cost-efficient, scalable capacity makes economic sense.

3. Unified Namespace = No Silos, No Rewrites: Applications see a single global namespace across on-premises and cloud environments, which means no application changes, no manual data movement, no vendor lock-in.

4. Repeat On-Demand: When demand spikes again, data is instantly staged back to Tier 0 for performance — without permanent high-performance cloud storage infrastructure costs.

CRN’s Cloud 100 companies demonstrate dedication to supporting channel partners and advancing innovation in cloud-based products and services. The list is the trusted resource for solution providers exploring cloud technology vendors that are well-positioned to help them build cloud portfolios that drive their success.

In 2025, Hammerspace launched several new campaigns and resources to help its partner community drive success, including extensive cloud marketplace and enablement resources, new distribution models, an expanded partner portal and continued growth of its global team. In addition, the international “Hammerspace Partner Roadshow 2025: AI Anywhere” equipped Hammerspace partners with critical insight, tools, and connections to accelerate their AI businesses.

Reminder: Google shutting down the Dark Web Monitoring Tool this week

Posted in Commentary with tags on January 12, 2026 by itnerd

This week, Google will start shutting down its dark web monitoring tool — the Dark Web Report — which was designed to scan the dark web for users’ exposed personal information. Users who want to stay “in the loop” should seek other tools.

Shutdown timeline

  • January 15, 2026: The scans for new dark web breaches stop.
  • February 16, 2026: The dark web report is no longer available, all data related to the report will be deleted.

Google previously stated its intention to focus on tools that provide customers with clearer, more actionable steps to protect their online information. However, no concrete announcements regarding new cybersecurity tools have been made by the company to date.

Karolis Arbaciauskas, head of product at the cybersecurity company NordPass, comments:

“It’s a useful tool. But I guess it’s time for something new, especially since other similar tools already offer prescriptive advice and practical recommendations for users whose data is found on the dark web. Google often replaces its products and features with new ones instead of updating them. Users should look for reliable tools that are dedicated to this task and are constantly supported and updated.

“Proactively monitoring the dark web for your credentials is a critical security habit. Fortunately, Google’s tool was never the only option. Security-conscious users who wish to continue scanning the dark web can utilize the tools offered by modern password managers.

“These integrated tools have evolved significantly in recent years. Advanced password managers now feature built-in scanners that operate 24/7, continuously monitoring the dark web and instantly alerting users if credentials or credit card data are detected. This enables individuals to take swift action before threats escalate.

“In case of a breach, the key is to act quickly. If you get an alert about your data being exposed, take immediate steps: change all affected passwords, cancel compromised credit cards, and review your account activity for anything suspicious.”

I’m currently looking around for a tool or tools to replace this. If I come across any, I will let you know. But if you have any suggestions, please leave a comment and let us all know.

Parallel Works Deploys Mission-Critical Weather Forecasting to the Cloud

Posted in Commentary with tags on January 12, 2026 by itnerd

 Parallel Works, provider of the ACTIVATE™ control plane for hybrid multi-cloud computing resources, today announced that its collaboration with General Dynamics Information Technology (GDIT) on a solution to expand a large federal agency’s complex, mission-critical weather forecasting HPC workloads to the cloud has been awarded the Editors’ Choice Award for Best Use of High-Performance Computing (HPC) in the Cloud, as part of the 22nd edition of the HPCwire Readers’ and Editors’ Choice Awards, presented at SC25 in St. Louis, Missouri.

GDIT partnered with Parallel Works and utilized the ACTIVATE platform, to support a large federal agency, and move complex, mission-critical HPC workloads from traditional on-premises systems. The project maintained the reliability and performance standards essential for life-saving weather forecasting, demonstrating that even the most demanding mission-critical HPC operations can successfully transition from traditional HPC systems to cloud infrastructure.

The coveted annual HPCwire Readers’ Choice Awards are determined through a nomination and voting process with the global HPCwire community, as well as selections from the HPCwire editors. The awards are an annual feature of the publication and constitute prestigious recognition from the HPC community. They are revealed each year to kick off the annual supercomputing conference, which showcases high-performance computing, networking, storage, and data analysis. Winners were revealed at the SC25 HPCwire booth, as well as on the HPCwire website.

Resources

LambdaTest Rebrands to TestMu AI

Posted in Commentary with tags on January 12, 2026 by itnerd

LambdaTesttoday announced its rebrand to TestMu AI, marking a bold step forward in its evolution from a cloud testing platform to the world’s first full-stack Agentic AI Quality Engineering platform.

Founded in 2018, LambdaTest rapidly became one of the most trusted names in cloud-based test orchestration and execution. They built a scalable, high-performance test cloud that removed flakiness, improved developer feedback loops, and drastically accelerated release velocity.

In 2022, the company began a deep transformation, moving deep into agentic AI across its products and workflows to empower autonomous, intelligent quality engineering at scale. Today, TestMu.ai’s AI Agentic unified platform powers end-to-end quality engineering for over 2.8 million developers and testers worldwide, helping teams ship faster and build with confidence.

The transformation addresses a critical inflection point in software development: as AI generates code at unprecedented rates, traditional testing creates a bottleneck. Quality Engineering teams need intelligent systems that can reason about change, observe failures, and adapt continuously.

As TestMu AI, the organization is expanding its platform to support the next generation of software builders, including ‘vibe coders.’ Their AI agents allow developers to ‘vibe test’ and move at the speed of thought, and ensure that vibe-coded apps are not only of high quality but are also reliable when they come in front of customers

TestMu AI has re-architected its platform be AI Native, deploying autonomous AI Agents to plan, author, orchestrate, and analyze software quality with minimal manual intervention. The platform now delivers:

  • Autonomous AI Agents for Testing: Plan, author, and evolve end-to-end tests using company-wide context or simple natural language prompts. Users can now test every layer of the database, API, UI, performance, and more.
  • Agentic AI Test Cloud: A scalable and unified test execution cloud to run any type of test at any scale, including visual regression, accessibility, API, and performance testing, web and mobile, to custom enterprise environments.

This shift reflects the company’s average 110% year-on-year growth over the last 2 years. To date, the platform has executed billions of tests for 18,000+ enterprise customers, including Microsoft, OpenAI, NVIDIA, Vimeo, Dunlem across 90+ countries.

The name ‘TestMu’ was adopted directly from its community. Since 2022, the TestMu Conference has served as the industry’s primary forum for advancements in AI and quality engineering.

By adopting this name, the company is signaling that the community is the core of the organization, infusing the spirit of that open, collaborative conference into the platform itself.

The company was recently recognized in the 2025 Gartner® Magic Quadrant™ for AI-Augmented Software Testing Tools report and in The Forrester Wave™: Autonomous Testing Platforms 2025 report.

Looking ahead, TestMu AI’s roadmap includes fully autonomous AI agents, agent-to-agent testing, evaluation of AI systems by AI systems, and deep integration with codebases and developer workflows, positioning quality engineering as a continuously learning, self-governing layer of modern software development.