Scott Stephenson, CEO and Founder of Deepgram, would like to respectfully offer the following thoughts on this news:
“OpenAI’s new model shows progress, but the benchmarks make it clear: latency, turn-taking, and lack of control remain its Achilles’ heel in real conversations,” said Scott Stephenson, CEO and Founder, Deepgram. “When you measure what makes conversations actually work — speed, politeness, and turn-taking — Deepgram still leads the pack. The benchmarks confirm what users feel: conversations with Deepgram just flow more naturally.”
Stephenson continued, “Why does this matter? In real-world deployments, people don’t judge a voice agent by its feature set — they judge it by how the conversation feels. Latency and turn-taking aren’t technical footnotes; they’re the difference between a helpful interaction and a frustrating one. That’s why benchmarks that measure conversational flow, not just functionality, are the true indicator of readiness for production.”
Benchmarks That Back It Up
#1 across all tests: Deepgram ranked highest under every VAQI weighting — equal, politeness-heavy, and latency-heavy.
Politer conversations: Fewest interruptions, meaning agents don’t talk over users.
Faster responses: Sub-second average latency (0.85s) vs. OpenAI’s 2.55s.
Smarter timing: Strong turn-taking with a competitive miss rate (0.427).
Consistent edge: Even when benchmarks shifted priorities, results held — Deepgram stayed on top.
Posted in Commentary with tags Hisense on September 2, 2025 by itnerd
Hisense is redefining the home cinema experience with the official launch of its TriChroma Laser TV L9Q, an innovated product that combines cinematic intensity, rich audio and an award-winning design to transform any space into an unforgettable personal theatre.
The L9Q features 5,000 ANSI lumens of brightness and contrast ratio of 5000:1, producing incredibly vivid images with deep blacks and sharp detail — even in well-lit rooms. With precision light control and vibrant colour accuracy from triple laser technology, every frame comes alive with cinematic intensity.
The L9Q is built for versatility, supporting screen sizes from 80 inches up to a massive 200 inches to turn any room into an epic home theatre. To ensure a perfect picture no matter the light, Hisense pairs the L9Q with an Ambient Light Rejection (ALR) screen, which is available in a variety of sizes and delivers brilliant images even in bright rooms.
Thanks to its IMAX Enhanced and Dolby Vision certifications, the L9Q presents visuals with extraordinary clarity and scale. This technology guarantees stunning, studio-grade immersion, whether enjoying a blockbuster film or binge-watching a favourite series. For audio that truly matches the stunning visuals, the L9Q features the exclusive Opéra de Paris | Devialet edition. Its 6.2.2 channel speaker system with Dolby Atmos creates rich, multidimensional sound that fills the entire room, making it perfect for both movies and music.
With an Ultra Short Throw (0.18 Throw Ratio) design, the L9Q can project a ginormous screen from just inches away, simplifying setup to fit into any space. Beyond its technical performance, the L9Q is a work of art. Its design, inspired by the circular structure of the Royal Opera House and the acoustic chambers of Roman theaters, reflects Hisense’s philosophy of “Timeless Artistry Meets Dynamic Technology.” This iconic design has been recognized with both the 2024 Red Dot and 2025 iF Design Awards, elevating the home cinema experience in both form and function.
According to Omdia Q1 2025 data, Hisense ranked No.1 globally in Laser TV volume share (69.6 per cent), extending its leadership for six consecutive years. As Hisense’s most advanced laser model, the L9Q reflects Hisense’s vision to help users Own the Moment — transforming everyday experiences, from family movie nights to solo concerts and game-day thrills, into unforgettable memories.
The L9Q will be available soon in Canada at authorized retailers.
Today, Comparitech researchers released a study looking at the state of global ransomware attacks in August 2025.
Ransomware attacks continued to climb again in August, rising from 473 in July to 506 last month. August also saw a first-of-a-kind attack on the State of Nevada. While hundreds of US government organizations have suffered ransomware attacks, this is the first-ever statewide attack.
Rebecca Moody, Head of Data Research at Comparitech, commented:
“If we needed a reminder of how dominant a threat ransomware is, August’s statistics provide it. Not only did we see a steady increase in attacks but we also witnessed a first-of-its-kind attack on the State of Nevada. The latter in particular highlights how no one, not even a multi-billion-dollar government organization, is immune to these types of attacks. And, even though numerous countries and governments are looking to ban public entities from making ransom payments, this is doing little to deter hackers.”
“Why? It’s likely due to a number of reasons. Firstly, these attacks are often random, e.g. because the hackers start exploiting a known vulnerability or a staff member happens to click on or download something they shouldn’t. Second, even if the hackers don’t receive the ransom, they’re most certainly going to gain notoriety when they make their claim on the State of Nevada. So, when another entity finds itself facing an attack from the same organization, they’ll instantly recognize the group’s name and may be more inclined to pay up before the attack escalates any further. Finally, it’s more than likely that the hackers will have stolen data in this attack on Nevada, so they’ll always have this to sell on the dark web if needed.”
“While banning public entities from making ransom payments may be a step toward reducing ransomware attacks, it isn’t the silver bullet. Rather, it should be part of a multi-pronged approach and one that makes sure the basics are covered. This includes patching any vulnerabilities as soon as they are flagged, making sure systems are regularly updated, carrying out frequent backups, investing in employee training, and having a step-by-step plan in place should an attack occur.”
Posted in Commentary with tags KnowBe4 on September 2, 2025 by itnerd
By Anna Collard, SVP of Content Strategy and Evangelist at KnowBe4
Inda Sahota, Group Cybersecurity Office (GCSO) Cyber Culture & Training – Fresenius
Recently, I had the pleasure of speaking with Inda Sahota, the dynamic and deeply empathetic force behind cybersecurity awareness at Fresenius Group. What struck me most wasn’t just her deep understanding of human-centric security, it was how naturally she bridges the gap between personal values and professional practice.
Inda brings her whole self into her work: her empathy, intuition, and a grounding in values passed down from her parents, progressive thinkers and first generation Punjabi Indian immigrants to the UK. They instilled in her and her sisters a quiet but powerful sense of agency. When cultural voices around them suggested that girls were somehow less capable than boys, her father would respond with a deceptively simple challenge:“But you can eat, can’t you?”
His way of creating initial confusion sparked critical thinking, and a gentle dismantling of limiting beliefs that, if left unchecked, could have developed into lifelong insecurities.
Our conversation got us thinking about the intersection of critical thinking, values-based education, self-efficacy, and digital mindfulness, especially in a world where we are exposed to online manipulation on a daily basis.
From Awareness to Agency
In security awareness design, we often focus on rules: don’t click this, don’t trust that, don’t reuse your password. But what if we focused instead on values? On presence. And on the cultivation of agency and critical thinking, the kind that Inda’s father nurtured in her from a young age? Psychologist Albert Bandura’s concept of self-efficacy, the belief in one’s capacity to act in the face of challenges, is central here (Bandura, xx). Research shows that self-efficacy is a strong predictor of behaviour change, and it has been linked to improving cybersecurity awareness attitude, knowledge, and behaviour (Arachchilage & Love, 2014; Zainal et al., 2021).
As Inda put it:“Resilience is like water. You need to be able to flow.”
In other words, we need to prepare, not just protect, our people. Whether we’re speaking to employees, children, or our broader communities, we need to teach them how to adapt fluidly, not just obey. How to stay present, not just paranoid. “This is about more than cybersecurity,” Inda notes. “It’s about helping people reclaim their agency in a world designed to exploit their attention and emotions. This fluid resilience allows individuals to:
Recognise when they’re being emotionally manipulated
Pause before responding to urgent digital demands
Stay centred when algorithms try to steal their attention
Respond with intention, rather than react impulsively
Presence vs. Performance: The Cost of Multitasking
One of the biggest threats to cybersecurity by the way, isn’t malware. It’s human error, often linked to distraction, overwhelm and media multitasking. And attention is one of our most compromised assets. Studies show that frequent multitasking reduces cognitive control, impairs memory, and increases difficulty in impulse control (Ophir, 2009; Baumgartner, 2014). And people who engage in high media multitasking engage in riskier cybersecurity behaviours compared to the low multitaskers (Hadlington & Murphy, 2018).
This fragmentation of attention doesn’t just make us less productive, it makes us more vulnerable. Scammers, phishers, and social engineers exploit us best when we’re rushed, distracted, over-stimulated or overwhelmed without realising. As a result, mindfulness becomes a cybersecurity imperative, not just a wellness buzzword.
Habits that Shape the Mind
Digital hygiene, like brushing your teeth, only becomes effective when it’s habitual. But forming habits, particularly in high-distraction environments, requires deliberate design. If we want people to pause before clicking a link or question a seemingly friendly DM, we need to design cues and rewards that reinforce critical thinking. This is where digital mindfulness practices can play a critical role in training the brain.
What Inda’s father modeled for her was a form of cognitive scaffolding. He didn’t control her environment or scare her into obedience. Instead, he provided intuitive frameworks for situational self-awareness, such as: “Have eyes at the back of your head.”
This is a powerful metaphor for living with conscious awareness and for being both vigilant and empowered. And those are precisely the qualities we need to foster in our digital citizens. So how can we apply this to our digital spaces?
Here are 5 practical ways to build digital resilience starting today.
1. Question, Don’t Lecture
Instead of explaining all the dangers of the internet, ask questions that help think critically:
“What do you notice about how you feel after scrolling for an hour?”
“What is the intent behind this narrative, article or social media post?”
“What emotions are triggered by the narrative?”
2. Build Self-Efficacy Through Practice
Research by Dr. BJ Fogg at Stanford’s Behavior Design Lab shows that lasting behavioral change happens through tiny habits that feel easy to do. In the digital realm, this might mean:
Pause for three seconds before clicking on links
Creating simple rituals around device usage – i.e. no screens at meals, or in bedroom
Play critical thinking games, illusions and logic riddles
Phishing tests and “spot the phish” or “spot the deepfake” games
The key is making these practices feel natural rather than imposed. Creating safe opportunities for people to practice digital decision-making and learning from mistakes also helps building self-efficacy.
3. Model Mindful Technology Use
We learn more from what we observe than what we’re told. You can model mindful technology use by:
Putting devices away during conversations
Thinking out loud when you encounter suspicious emails
Demonstrating how you fact-check information before sharing
Try the 5-minute rule. Tell yourself: “If I still need to check this in 5 minutes, I will.” This pattern interrupt helps break unhealthy autopilot impulses.
4. Develop Emotional Regulation Skills
Social media platforms and cybercriminals alike exploit our emotional responses to drive behaviour. They create artificial urgency, leverage fear of missing out, and use variable reward schedules that mirror addictive behaviours. Training should show how to recognise when one is being emotionally manipulated by technology. Simple practices like taking three deep breaths before responding can activate the prefrontal cortex and reduce reactive behaviour.
5, Create an emotionally safe environment
People need to feel psychologically safe to slow down. Create environments where questioning is welcomed, where “Let me verify this first” is praised, not criticised. When it’s okay to ask “Does this seem right to you?” without fear of looking incompetent, people actually become more vigilant, not less.
Bringing the Being into the Human
One of Inda’s most poetic expressions stayed with me: “We need to bring the human back into the being, and the being into the human.”
What if we saw our intuition and self-awareness as cybersecurity superpowers? What if we cultivated presence alongside password hygiene? We might just build a digital culture where security isn’t only about understanding the risks, but about knowing ourselves.
The ability to communicate and collaborate across languages and cultures is no longer a luxury but a necessity. A new generation of intelligent AI translation tools is breaking down old barriers, turning complex, fragmented global workflows into seamless engines for connection and growth. By making life easier and boosting productivity, these tools are not just changing how businesses operate—they are helping people connect on a global scale more effectively than ever before.
This article explores how these AI tools are delivering tangible benefits, highlights key applications across different professional fields, and looks at the emerging trends that will shape our future.
A New Era of Productivity and Connection
The most immediate impact of AI translation tools on global teams is a dramatic boost in speed and efficiency, which translates into more time for meaningful work. For any organization operating globally, this shift is transformative.
Reduced Cycle Time: Tasks that previously took weeks, such as localizing a global marketing campaign or updating training materials in multiple languages, can now be completed in days or even hours. This acceleration allows ideas to reach a global audience while they are still relevant and impactful.
Fewer Handoffs: Intelligent automation eliminates many of the repetitive, manual tasks that slow teams down—like copying content between systems, reformatting files, and managing endless review cycles. This frees up human talent to focus on strategy, creativity, and building relationships.
Lower Operational Costs: By streamlining processes and reducing reliance on manual, third-party interventions for every task, organizations can operate more efficiently. The 2025 Stanford AI Index Report highlights these gains, with companies reporting significant cost savings in service operations (49%), supply chain management (43%), and software engineering (41%).
Ultimately, this efficiency is about more than just the bottom line. It’s about removing friction. When teams can scale their work globally without the usual roadblocks of delays and cultural silos, they can spend more energy on innovation and connection. As Ivan Smolnikov, CEO of Smartcat, puts it, “every enterprise dreams of scaling globally without the usual roadblocks—global campaign delays, cultural silos, and bottlenecks in transferring expertise.”
The best systems achieve this through a powerful partnership between AI and human expertise. By letting AI handle the initial heavy lifting and empowering human reviewers to refine and approve the work, this “human-in-the-loop” approach delivers both speed and quality, achieving better outcomes than either AI or humans could alone.
A key innovation driving this partnership is adaptive AI translation. Unlike static models, adaptive AI tools learn in real-time from every correction made by a human reviewer. When a linguist adjusts a phrase for tone or accuracy, for example, the system instantly absorbs that feedback, ensuring it won’t make the same mistake again in future content. This creates a powerful feedback loop where the AI becomes a continuously improving student of your brand’s unique voice.
Uniting a Global Voice: Consistency Across Cultures
Maintaining a clear and consistent message across dozens of languages is a monumental challenge. AI tools are proving instrumental in solving this, ensuring that an organization’s core identity and values resonate universally.
These AI translation systems act as guardians of brand consistency, applying the same rules, terminology, and brand guidelines to all content, regardless of the target language. Through features like centralized glossary management, they ensure that critical terms, from product names to compliance language, are used correctly everywhere. More advanced AI translation tools take this a step further by incorporating AI-driven quality estimation. This feature automatically scans translated text, flagging sentences that may sound awkward or deviate from the source’s meaning. This allows human reviewers to focus their attention on the small fraction of content that needs refinement instead of manually checking every word, dramatically accelerating the quality assurance process. This prevents the kind of linguistic inconsistencies that can confuse customers, dilute a brand’s message, and erode trust.
This synergy between AI-driven consistency and human oversight empowers global teams to maintain high standards of quality while moving at an unprecedented pace, fostering a unified brand presence that speaks authentically to every market.
Practical Applications: AI Tools at Work
Across industries, intelligent AI software is already delivering powerful results. These real-world applications show how these tools are making work easier and more impactful today.
1. Marketing and Global Communications Marketing teams use AI translation tools to create, localize, and launch global campaigns with remarkable speed. In fact, nearly 70% of companies are already using cloud-based generative AI, according to Statista research on AI tool usage. Teams can now create content in multiple languages simultaneously, ensuring that a single, cohesive message is adapted for local nuances and ready for a global launch in days, not weeks. Nicole DiNicola, VP of Marketing of Smartcat, who has also led AI implementations for her teams at major brands, notes that, “marketing has always been about impact. AI agents give us a chance to have more of it, with less friction.”
Example applications: Global product launches, multilingual email campaigns, website localization, and social media content.
2. Learning & Development (L&D) AI translation now makes it possible for L&D professionals to create and update corporate training programs for a global workforce in a fraction of the time. This is especially vital for compliance training, where accuracy is non-negotiable. When regulations change, updates can be rolled out instantly across all language versions, keeping the entire organization aligned.
Example applications: Employee onboarding, compliance training, product knowledge courses, and skills development programs.
3. E-commerce and Retail Operations For e-commerce businesses, AI is accelerating the process of localizing vast product catalogs. By integrating directly with Product Information Management (PIM) systems, AI can ingest product data, apply translation and glossary rules, and sync updates automatically. This keeps online stores and retail partners perfectly aligned without cumbersome spreadsheets. Companies using this approach have reported dramatic reductions in turnaround time, with tasks that once took hours now completed in mere minutes.
The Road Ahead: What’s Next for AI Translation
The AI landscape is evolving at a breathtaking pace. Key trends point toward even deeper integration and broader adoption in the near future.
From Niche Tool to Workplace Staple: AI is rapidly moving from specialized departments to enterprise-wide implementation. Gartner forecasts that by 2028, approximately one-third of all enterprise software applications will have these intelligent capabilities embedded within them.
The Power of an Integrated Ecosystem: Standalone tools are giving way to integrated platforms where multiple AI agents work together to automate end-to-end processes. By connecting with CRM, CMS, and LMS tools through APIs, these ecosystems create seamless workflows across the entire organization.
Smarter, More Capable AI: The underlying large language models are becoming exponentially more powerful. Research shows that the length of complex tasks AI can handle is doubling roughly every seven months. When applied to business, this means AI can tackle specialized content like legal documents and technical manuals that once required extensive human effort.
A More Connected Future
AI translation technology has moved beyond hype to deliver tangible outcomes. Teams are more productive, brands are more consistent, and global operations are more efficient. But the real story is a human one. By automating the mundane and accelerating collaboration, these tools are empowering people to connect more effectively across linguistic and cultural divides. The future of work is emerging as a partnership between human talent and technology, building a world where great ideas are shared, understood, and embraced by everyone, everywhere.
Cybernews’ latest overview of the first half of 2025 reveals that the total number of breached accounts fell by 20 timescompared to the same period in 2024, according to Cybernews’ Personal Data Leak Checker tool.
However, while breaches have dropped significantly compared to last year, several countries remain highly vulnerable as millions of accounts remain exposed. The top three countries with the highest number of breaches in 2025 are the U.S., with 2.5 million breached accounts, France, with 1.8 million breached accounts, and India, with 1.2 million breached accounts.
This report offers key insights into global data breach trends in the first half of 2025, highlighting that unsafe online practices such as subscriptions to suspicious websites and weak password reuse remain the leading causes of breaches.
Key findings of this research:
The first six months of 2025 recorded 15.8 million breached accounts globally.
The top five countries most breached worldwide in the first six months of 2025 are the U.S., France, India, Russia, and Venezuela.
January has the highest number of breached accounts in 2025 so far.
Countries experienced a spike in breach numbers in March, making it the second most breached month in the first six months of 2025.
The second quarter of 2025 saw a 77% dip in breached accounts.
Despite a 20-fold drop compared to 2024, breach density per capita shows the U.S. remains the most affected nation, with 8 in every 1,000 internet users impacted.
Deepgram’s Unfiltered Views on The Announcement From OpenAI
Posted in Commentary with tags Deepgram on September 2, 2025 by itnerdOpenAI just made an announcement titled, “Introducing gpt-realtime and Realtime API updates for production voice agents” found here: https://openai.com/index/introducing-gpt-realtime/
Scott Stephenson, CEO and Founder of Deepgram, would like to respectfully offer the following thoughts on this news:
“OpenAI’s new model shows progress, but the benchmarks make it clear: latency, turn-taking, and lack of control remain its Achilles’ heel in real conversations,” said Scott Stephenson, CEO and Founder, Deepgram. “When you measure what makes conversations actually work — speed, politeness, and turn-taking — Deepgram still leads the pack. The benchmarks confirm what users feel: conversations with Deepgram just flow more naturally.”
Stephenson continued, “Why does this matter? In real-world deployments, people don’t judge a voice agent by its feature set — they judge it by how the conversation feels. Latency and turn-taking aren’t technical footnotes; they’re the difference between a helpful interaction and a frustrating one. That’s why benchmarks that measure conversational flow, not just functionality, are the true indicator of readiness for production.”
Benchmarks That Back It Up
Source: VAQI Benchmark, August 2025
Deepgram published a blog today with further details: https://deepgram.com/learn/vaqi-openai-gpt-realtime-test-with-sensitivity-analysis
Leave a comment »