Ab Emam

IDF is expanding its language intelligence programs, citing the Gaza war as proof that human expertise is irreplaceable; soldiers are being trained in Arabic dialects, Farsi and cultural nuance to meet evolving intelligence challenges *** The war in Gaza has demonstrated clearly how wrong it was to believe that technology can replace human beings when dealing with language. In the coming years, the IDF intends to enlist more people who deal with language, said Lt. Col. H, who commands the military's signal intelligence field and heads the Intelligence Corps' school of languages. Hundreds of people have gone through the military intelligence language schools since its establishment, being taught foreign languages, often from scratch, including different dialects of Arabic spoken in Syria, Lebanon, Iraq, the Palestinian communities and among the Bedouins. A recent addition is the Yemeni Arabic, used by the Iran-backed Houthis. Farsi has also been taught to soldiers serving in the IDF and it is among the longest-lasting language courses since it was added 25 years ago. "We understood back then that we would need soldiers who could speak Farsi," H. says. "We understand that we must grow our numbers and have added people who had no background in the language and taught them to speak it perfectly. Israelis study 25-week-long courses before they are inducted into the service and come out specializing in reading and the spoken language.

89

4 Comments

🤖 We're outsourcing our choices faster than we're questioning them. I had the privilege of joining brilliant minds like Merve Hickok (Center for AI and Digital Policy; University of Michigan), Ece Kamar (Microsoft AI Frontiers Lab), Eelco Herder (Utrecht University), Liz Miller (Constellation Research, Inc.), and my good friend FAISAL HOQUE in a thoughtful new piece brilliantly pulled together by KJ Bannan for Communications of the ACM about the AI-enabled agentic Internet. Here's the paradoxical insight I offered up for this piece and now can't stop thinking about: "The easier technology makes things, the harder we must think about its effects." Right now, AI agents are booking our flights, curating our news, paying our bills. Convenient? Absolutely. Neutral? Hardly. These systems aren't just completing tasks — they're quietly authoring the story of who we become. When AI agents handle our routine choices, they do more than save time: they shape our behaviors, preferences, and ultimately, our autonomy. The Now-Next Continuum (https://lnkd.in/ergiTwSu) is crystal clear here. Today's "helpful assistant" becomes tomorrow's invisible architect of human preference. The question isn't whether this shift benefits us — it's whether we remain conscious participants in our own lives. Consider the harms of inaction: We drift into algorithmic dependency without guardrails, transparency, or kill switches. But the harms of thoughtless action? We build systems that amplify bias at network scale, where one flawed agent talking to another creates exponential distortion (as Merve Hickok astutely noted). The solution isn't to resist the agentic future — it's to design it with intention. Ece Kamar's insight rings true: we need human-agent collaboration that keeps people in control. Liz Miller's call for enterprise accountability? Essential. We must demand transparency from developers and platforms alike. Here's what gives me hope: We still get to choose. Every procurement decision, every implementation, every moment we pause to ask "What matters here?" shapes whether AI serves human flourishing or simply human convenience. The agentic Internet is already reshaping reality. The only question is whether we'll be conscious co-creators or passive consumers of that change. Your turn: What choices are you making — or avoiding — that could determine how AI shapes your life? Read the full exploration: https://lnkd.in/ekBv8qtV #AI #TechHumanist #AgenticAI #HumanFlourishing #DigitalTransformation #TechnologyStrategy #AIEthics

23

7 Comments

Most enterprise AI agents are "slop." That was the blunt opener delivered at our recent VentureBeat AI Impact salon by Amjad Masad, CEO of Replit. When the CEO of the fastest-growing SaaS company (#1 on Ramp’s December rankings for new customers and percentage change in new customer count: https://lnkd.in/gqEj6zxm) told us that 99% of agents outside of coding and customer support are currently "toy examples," we listened. We hosted an intimate, packed room of decision-makers from Google, Walmart, Uber, Wells Fargo, JPMorganChase, Cisco, Oracle, PayPal and others, to debate the real state of agentic AI. The conversation ran straight through dinner until 10:30 PM, and the consensus was that "intelligence" isn't a bottleneck. The infrastructure is. The systems around the model aren't ready. Here are three themes that emerged from the debate: 1. The "Squishy" Problem (Amjad Masad, Replit): Amjad argued that agents fail in the enterprise because most workflows lack a "Unit Test." You can verify code (it compiles or it doesn't), but you can't verify a marketing strategy. Until we solve the "messy data" problem and build better verification loops (like the "Computer Use" hack Replit built internally), most agents will remain in the "slop" zone. 2. Probabilistic systems inside deterministic orgs: (Michael Clark, Google Cloud): Mike noted that tech is moving exponentially, but organizational readiness is moving logarithmically. We are trying to deploy probabilistic agents into deterministic governance structures designed for "IBM typewriters." The Scoop: Mike joined us on the evening Google launched its Interactions API – a massive shift providing "statefulness" to agents. This is a fix for the "amnesia" problem that plagues cost and performance today. 3. The Physics of Memory (Shimon Ben-David, WEKA.io): While everyone worries about compute, Shimon pointed out the hidden wall: memory. As agents become stateful (see Google’s update), the "Model Zoo" and related to token usage explodes. Shimon shared data showing how offloading KV Cache to high-performance storage can make 100 GPUs perform like 420. That is a 4x efficiency gain that changes the unit economics of autonomy. The takeaway: 2026 is going to feel like a plumbing year – the stateful memory, the governance rails, and the data pipelines – to let them survive production. A huge thank you to Weka for powering this conversation, and to Taryn Plumb for capturing good details in the VentureBeat article here: https://lnkd.in/gfNe82kk #AI #EnterpriseAI #Agents #Infrastructure #Replit #GoogleCloud #VentureBeat cc: Val Bercovici

148

17 Comments

I’m honored that my conversation with Bilawal Sidhu at TED2025, The AI Revolution is Underhyped, was the most popular TED conversation of the year. I think the reason it resonated is simple: people want to know what AI means for their future, and it matters deeply that they do. The opportunities for AI are extraordinary, including breakthroughs in biology and material science, new approaches to education, and global productivity gains at a scale we’ve never seen. The risks are equally real, from great-power competition to the potential misuse of powerful open-source systems. We are entering a period where the slope of progress will matter more than the starting point, shaping economies, geopolitics, and daily life. My advice hasn’t changed since our talk in April: use these systems, understand them, and build with them. AI’s potential is far greater than most realize. The gap between adopters and non-adopters is widening quickly, and those who engage now will be far better positioned for what comes next. I’m thankful to Bilawal for a thoughtful, incisive conversation on where AI is headed, from autonomous agents and scientific discovery to open-source debates and the deeper questions of non-human intelligence. The race for AI dominance isn’t just about innovation; it’s about power, sovereignty, security, and ultimately, human dignity. How we navigate this moment will define the next century and the kind of world we build, one that either empowers humanity or erodes it. If you haven’t seen the conversation yet, I encourage you to watch it and keep learning as much as you can about AI. AI may be the most transformative development in centuries, and it’s happening in our lifetime. As I said in April, let’s not screw it up. Full conversation here: https://lnkd.in/ewsg8AET?

584

66 Comments

One of my favorite panel discussions at the recently concluded OSJ 2025 was on “Navigating the Future of the OSV Sector.” The session brought together some truly stellar speakers — Fazel Fazelbhoy, Alex Brykalov, and Samer Qiblawi — each offering sharp, experience-backed perspectives on where the OSV industry is headed. It was the first time I heard Samer Qiblawi from Marcap on stage, and I was genuinely impressed by his clarity and conviction on how digitalization is no longer an enabler, but a decisive driver of the future OSV landscape. He articulated how digital platforms like Shiparc.AI and VoyageIQ (VIQ) assist in data-driven decision-making, and how real-time operational visibility are transforming: • Fleet utilization and charter performance • Fuel efficiency and emissions management • Predictive maintenance and asset reliability • Safety, compliance, and crew effectiveness What resonated strongly was the emphasis that digitalisation is not about technology alone — it’s about changing mindsets, workflows, and culture across owners, operators, and charterers. As the OSV sector navigates energy transition pressures, tightening margins, and increasing operational complexity, discussions like these reinforce a simple truth: Those who embrace digitalisation strategically will lead — not follow — the next phase of the industry’s evolution. Well done to the OSJ team for curating such a thought-provoking panel. #digitilization #Maritime #osj2025

41

Parliament Speaker Niko Peleshi met today with Albanian-American entrepreneur Donald Leka. In a social media post, Peleshi said Leka is changing the way Artificial Intelligence works. He explained that Leka’s patented technology uses his dynamic reasoning model (LDRM) and predictive intelligence to forecast complex events by understanding the real world. This approach opens new opportunities for innovation and global strategic impact. Donald Leka is the founder, chairman, and CEO of the Jumptuit Group. He also created Glide OS, a leading cloud-computing platform and award-winning industry solution. Read More: https://lnkd.in/dnNHVvGi “We believe that Albania can become a hub of AI-powered innovation in Europe and build a diverse economic future that draws upon the talents of its youth and rewards its people. We are determined to help Albania write this critical and impactful chapter of its history.” - Donald Leka

57

3 Comments

I spent the last few days at the AI+ Expo in DC as I’m trying to learn more about the change that’s coming our way. The conference gathered experts from across government, military, industry, and academia. The most common refrains I heard were: “we don’t know yet” and “it depends.” It’s a reminder that we’re still early in where this is going. A few reflections: ➡️ We’re going to need to train 22-year-olds entering the workforce to do the kind of work we currently expect of 30-year-olds. AI is collapsing timelines — in how fast we can learn and give people responsibility. Managers and people entering the workforce (and those in it already) will need to be flexible as things change in the next few decade. ➡️ There will be job displacement — that’s pretty clear at this point. The real question is: whose responsibility is it to ensure workers are upskilled? Employers? Governments? Tech companies? As a society we’ve not seen such rapid change before and we need to figure it out or *a lot* of people are going to suffer. ➡️ AI success isn’t magic. It comes down to three things: compute power, talent, and data quality. The organizations that invest in all three are the ones that will thrive in the next thirty years. On a positive note: if there was one area that sparked universal excitement, it was healthcare. From diagnosis to drug discovery to tracking fitness, the potential for AI to save lives and improve outcomes is enormous. There’s so many unknowns, but also a real opportunity to shape how AI works for people. It’s both exciting and terrifying — one of *many* contradictions AI surfaces.

14

6 Comments

AI didn’t replace our product. It expanded what it could do. At Jotform, forms are still core—but AI Agents help turn them into conversations and outcomes. I explain how that shift happened and why it matters on What The Teck Show. https://lnkd.in/dcQwET2u

39

We’ve been teaching the wrong math for decades. 📉 Not only is it irrelevant to most careers — it may actually be making us less prepared for the future. In this episode of The CTO Show With Mehmet Gonullu, I sat down with Ted Dintersmith — bestselling author, former VC, and education reformer — to talk about: ✅ Why most high school math has no real-world application ✅ The math skills that actually drive innovation, startups, and decision-making ✅ How AI changes what it means to be “future-ready” ✅ Why creativity and risk assessment matter more than rote memorization Ted’s journey from venture capital to transforming education is a masterclass in spotting signals before the market sees them — and applying them to prepare the next generation. 🎧 Available now on all podcast platforms + YouTube (links in comments) #AI #Education #Startups #Leadership #FutureOfWork #VentureCapital #Math #Podcast

7

2 Comments